Stage-specific Recombination Factors Differentially Regulate Double-strand Break Repair Fidelity and Influence Genome Stability

Stage-specific Recombination Factors Differentially Regulate Double-strand Break Repair Fidelity and Influence Genome Stability | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Stage-specific Recombination Factors Differentially Regulate Double-strand Break Repair Fidelity and Influence Genome Stability Andres Clemente-Blanco, Lydia Iglesias, Adrián Campos, Celia Delgado, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8989286/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Homologous recombination (HR) repairs DNA double-strand breaks (DSBs) with high fidelity, yet how defects at distinct steps of the recombination pathway influence gene conversion (GC) outcomes and contribute to genome instability remains poorly defined. Here, we develop Di-GRAPH, a computational framework that integrates coverage profiling and discordant read mapping to systematically analyze mutational signatures, repair pathway choice, and GC architecture. At a genomic scale, Di-GRAPH detects, quantifies and classifies DNA damage-dependent gross chromosomal rearrangements across diverse genomic contexts. Applying Di-GRAPH to mutants defective in resection (exo1Δ), recombination intermediate processing (sgs1Δ, srs2Δ), and strand invasion (rad51Δ), we identify stage-specific GC patterns, mutation-specific repair signatures and distinct patterns of genome-wide rearrangements. Loss of Exo1 minimally affects GC formation but severely compromises genome stability in undamaged cells. Absence of Sgs1 and Srs2 leads to extended and discontinued GCs, accompanied by widespread ectopic recombination and retrotransposon mobilization following DSB repair. Rad51 deficiency results in severely impaired GC outcomes, strong reliance on non-homologous end joining, elevated break-proximal mutagenesis, and extensive genome-wide rearrangements in both DSB survivors and undamaged cells. Coverage and discordant-read signals were strongly correlated across most genomic features, indicating that recombination-associated translocations are frequently coupled to local DNA amplification. Comparative sequence identity analyses reveal distinct homology requirements, with ORFs showing the highest homology dependence, whereas repetitive elements recombine under more relaxed constraints. Together, this integrated analysis reveals that defects at discrete steps of the HR pathway impose distinct and predictable genome-wide instability signatures, providing a mechanistic framework for understanding how recombination fidelity is enforced across the DNA repair process. Biological sciences/Molecular biology/DNA damage and repair/Double-strand DNA breaks Biological sciences/Molecular biology/DNA damage and repair/Homologous recombination Double-strand break DNA repair homologous recombination genome-wide stability Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Faithful repair of DSBs is essential to maintain genome stability and prevent chromosomal rearrangements that could compromise cell survival. To counteract the detrimental effect of DSBs, cells employ evolutionarily conserved and highly coordinated repair pathways to restore these DNA lesions, preventing the transmission of unwanted mutation to the progeny. Non-homologous end joining (NHEJ) directly ligates broken DNA ends without requiring a homologous template, being often an error-prone mechanism that accumulates insertions, deletions or mutations at the break site. By contrast, HR constitutes one of the most faithful repair pathways, as it uses an intact homologous sequence as a template to recover the lost genetic information. HR is initiated by nucleolytic processing of DSB ends, which generates 3´ single strand DNA (ssDNA) through a process known as resection 1 . Resection occurs in two sequential phases. In the initial step, the nuclease activity of Mre11 within the MRX complex (Mre11-Rad50-Xrs2), together with Sae2, introduces an endonucleolytic nick at the boundaries of the DSB, and degrades DNA in the 3´-5´ direction toward the break, producing a short 3´ ssDNA overhang 2,3 . This intermediate is subsequently recognized by the exonucleases Exo1 and Dna2, that together with the STR complex (Sgs1-Top3-Rmi1) promote extensive 5´-3´ long-range resection, ultimately generating long ssDNA tracts 4,5 . Long-range resected tracts are coated with RPA and Rad51, which are necessary for ssDNA protection, homology search, strand displacement, and ultimately for promoting strand invasion, leading to the formation of the displacement loop (D-loop) 6-8 . After the missing genetic information has been copied, the D-loop is dismantled by the activity of helicases such as Sgs1 or Srs2 9-11 , and the newly synthesized DNA re-anneals to its original chromosomal context. The exposing 3´ ends then prime bidirectional DNA synthesis to complete repair. Because this recombination-mediated pathway involves unidirectional transfer of the sequence information from the donor to the recipient, it typically results in GC events. Although the core HR machinery has been well characterized over the past years, the specific contributions of HR factors to DNA repair, as well as their broader effects on genome-wide integrity, remains incompletely understood. This gap is particularly relevant because subtle perturbations in HR can generate complex genomic alterations, including gross chromosomal rearrangements, multi-strand invasions and chromothripsis, changes that are difficult to detect using classical genetic approaches 12 . Alongside, recent work has revealed that a single repaired DSB can induce long-lasting heritable alterations in chromatin topology and gene expression across large genomic regions, emphasizing that DNA repair outcomes extend beyond sequence restoration to persistent changes in genome organization and function 13 . Beyond sporadic or experimentally induced damage, DSBs also arise during meiosis within or near repetitive elements such as retrotransposons, increasing the risk of non-allelic recombination and genome rearrangements. High-resolution mapping of endogenous Ty elements in budding yeast has shown that these elements variably modulate local DSB formation, in some cases promoting nearby hotspots, highlighting the importance of the genomic context in recombination outcomes and genome stability 14 . In this context, genome-wide methods have recently transformed the ability to study DNA repair and genomic instability at unprecedented depth and scale. Techniques such as RADD-seq (Repair-Assisted Detection sequencing) allow mapping of DNA damage and repair sites across the entire genome rather than restricting analysis to a few pre-selected foci 15 . Moreover, high-throughput functional screening strategies such as Repair-seq and REPAIRome, which assay the outcomes of DSBs under hundreds of genetic perturbations, have exposed unexpected pathway dependencies 16,17 . Single-strand breaks (SSBs) have also been the focus in recent years, with the discovery of GLOE-seq providing a powerful tool to capture these DNA lesions and demonstrating its applicability to the mapping of Okazaki fragments and to the analysis of spontaneous distribution of nicks in yeast and human chromatin 18 . Discordant reads have also been used as an approach to detect various types of recombination events, both at naturally occurring and experimentally induced DSBs, providing a powerful tool to study genomic changes and DNA repair 14,19 . These genome-wide approaches provide a powerful, unbiased framework to detect, classify and interpret not only canonical repair products, such as GC products, crossovers or indels, but also complex structural rearrangements, as copy number changes or complex genome-wide reorganizations 20-22 , providing deeper insight into the mechanisms driving mutagenesis, disease, and genome evolution 23 . Here, we introduce Di-GRAPH ( D SB- i nduced G enome-wide R epair A nalysis and P rofiling of H omologous recombination), a computational framework that integrates whole-genome sequencing with coverage and discordant read analyses to quantitatively characterize the frequency, directionality and tract length of individual GC events associated to the repair of a single and site-specific HO break. Di-GRAPH also analyses the relative contribution of short- and long-range HR vs NHEJ in repair and their associated mutagenic frequencies. At the genomic scale, Di-GRAPH detects and annotates genome-wide gross chromosomal rearrangements and associates them with specific genomic categories (ORF, intergenic region, LTR, TEG, Ty, tRNA, rRNA, ncRNA, snRNA, snoRNA, ARS, centromere and subtelomeric region). By comparing pre-damage, repair and survivor populations, Di-GRAPH enables the classification of lethal vs non-lethal rearrangements and separates DNA damage-dependent effects from cell cycle-associated genomic changes. The application of Di-GRAPH to four key HR mutants representing distinct stages of the pathway (Exo1, Sgs1, Srs2 and Rad51), revealed distinct stage-specific repair signatures, both at the break site and genome-wide, highlighting essential roles of these factors in maintaining genome stability. Notably, most genomic features showed a strong correlation between coverage and discordant-read signals, indicating that recombination-driven translocations are commonly associated with local amplification, whereas sequence identity analyses revealed feature-specific homology requirements, with ORFs displaying the strongest homology dependence and repetitive elements recombining under reduced sequence constraints. Together, these findings demonstrate that HR factors acting at different stages of the pathway differentially safeguard genome integrity, both locally and genome-wide, and highlight Di-GRAPH as a powerful tool to define genome behavior in response to DNA damage. Results Survival after HO cleavage is uncoupled from DSB repair efficiency To examine how defects in specific steps of the HR pathway affect DSB repair and genome integrity, we focus on the role of the exonuclease Exo1 (long-range resection), the helicases Sgs1 and Srs2 (dissolution of recombination intermediates), and the recombinase Rad51 (homology search and strand invasion) in repairing an inducible HO break in the S. cerevisiae PMV background strain 24 . Briefly, this strain retains the endogenous wild-type MAT a locus on chromosome III but lacks the HMR / HML donor sequences, preventing intra-chromosomal repair after HO cleavage. Replacing the endogenous HO promoter with the inducible GAL1 promoter enables galactose-dependent induction of a defined DSB at the MAT a locus . As a donor for recombination, a 1.3 kb MAT a´ sequence is inserted at the ARG5,6 locus on chromosome V. This donor is identical to MAT a except for 23 evenly spaced polymorphisms (~60 nt). To prevent cleavage of the MAT a´ and re-cutting of gene conversion products, a C to G point mutation at the HO recognition site has been inserted (HO- inc ) (Fig. 1A). A single HO break was induced in wild-type, exo1Δ , sgs1Δ , srs2Δ , and rad51Δ cells by adding galactose to the media. Samples were taken after 12 h for wild-type and exo1Δ , and after 24 h for sgs1Δ , srs2Δ , and rad51Δ . Coverage recovery of a 2–4 kb DNA region right to the HO break varied substantially among wild-type and HR-deficient strains, reflecting distinct capacities of each mutant to repair the DSB (Supplementary Fig. 1A, left). Nevertheless, despite varying extents of repair efficiency, all strains exhibited reduced cell viability relative to their repair capacity (Supplementary Fig. 1A, right). This indicates that survival after HO cleavage is not solely determined by the repair efficiency and differs between DNA repair mutants, suggesting that these factors may act beyond the DSB to support cell viability. HO-selected survivors of wild-type, exo1 D , sgs1 D , srs2 D and rad51 D display distinctive GC patterns To investigate the origin of cell lethality after DSB repair, we induced an HO cleavage and collected samples immediately after DNA damage checkpoint release (non-selected survivors, NSS hereafter), and after several rounds of culture dilution (selected survivors, SS hereafter). As control for genomic changes unrelated to HO cleavage, parallel samples were collected from un-induced raffinose cultures (undamaged cells, UD hereafter) (Fig. 1B). We first quantified HO repair efficiency in each strain at the specified time points. Three of the MAT a´ polymorphisms generate Eco RI restriction sites (Supplementary Fig. 1B, left diagram), allowing the analysis of NHEJ vs HR and GC directionality by Southern blotting 24 . Hybridization with the left and right probes showed that GC in wild-type, exo1 D , sgs1 D and srs2 D was asymmetric to the left side of the HO site, incorporating the two Eco RI sites to this side of the HO- inc (Supplementary Fig. 1B, right panel). As expected, we detected a decrease in HR repair products in rad51 D cells, and a re-accumulation of the uncut band, indicative of DNA repair mainly by NHEJ in the survivor population (Supplementary Fig. 1B, right panel). Next, Di-GRAPH was used to calculate the coverage surrounding the HO site (800 nt up- and downstream) by aligning 75-mer sequencing reads to the reference genome (bowtie 25 in -m 1 mode to exclude reads mapping to duplicated regions). Coverage at each time point was normalized to their respective T0 values (Fig. 1C). Profiles showed asymmetric left GC with concomitant gain/loss of chromosome V/III, as previously reported 24 . Coverage patterns differed between strains, reflecting strain-specific GC. Differences were also observed between NSS and SS samples, mainly in sgs1 D , srs2 D and rad51 D cells, likely due to repair efficiency differences (Fig. 1C, extended low coverage at CIII in NSS, and Supplementary Fig. 1B). Extending this analysis to the entire chromosome III revealed low-coverage regions around the HO break in NSS samples of these mutants (Supplementary Fig. 1C). Notably, sgs1 D cells showed a coverage loss between two transposon elements (85-150 kb) in SS, indicative of stable gross chromosomal rearrangements in the absence of this helicase. No coverage changes were observed in uninduced raffinose samples, confirming that these effects are DNA damage-dependent (Fig. 1C and Supplementary Fig. 1C). To precisely quantify GC profiles, we split reads in 18-mer and aligned them using bowtie in -k 1 mode. In this condition, reads within inter-polymorphism regions mapped randomly to MAT a or MAT a´, generating a basal coverage line, whereas reads overlapping polymorphisms mapped specifically to chromosome III or V MAT regions (Supplementary Fig. 2A). Coverage graphs were then generated for each mutant and time point (Supplementary Fig. 2B). By measuring the size of MAT a/ MAT a´ peaks (distance between base and top/bottom), we quantified subtle changes in gain/loss at each polymorphism, reflecting small variations in GC, extension and directionality. NSS wild-type displayed declined MAT a with a symmetrical MAT a´ coverage enrichment extending up to 586 nt to the left of the break, with a preference for GC within the first 178 nt (Fig. 1D, left panel). No differences between NSS and SS were observed, indicating that GC tracts established during HO repair are stably inherited (Fig. 1D, right panel). Similar GC profile was observed in exo1 D cells, including comparable GC length and directionality in NSS and SS, indicating that Exo1 does not influence GC in the PMV strain (Fig. 1D). In contrast, sgs1 D , srs2 D and rad51 D exhibited poor MAT a´ gain and MAT a loss in NSS, consistent with low repair efficiency at this time point (Fig. 1D, left panel). Among SS, sgs1 D cells displayed extended MAT a´ tracts compared to wild-type, suggesting that Sgs1 limits GC length (Fig. 1D, right panel), whereas srs2 D cells showed no major defects in GC length, symmetry, or efficiency (Fig. 1D, right panel). rad51 D survivors showed equal distribution of MAT a/ MAT a´ polymorphisms, consistent with a complete loss of recombination between chromosomes III and V, except at the HO- inc , which retained an asymmetrical MAT a´ gain and MAT a loss pattern, indicating altered GC repair process (Fig. 1D, right panel). These GC profile variations were strictly HO-dependent, as un-induced raffinose cultures showed overlapping patterns with relative percentage near zero across all strains (Supplementary Fig. 3A). In summary, coverage profiling at the break site in the PMV background enables sensitive measurement of GC events, revealing stable strain-specific repair outcomes. Discordant reads at the MAT a/ MAT a´ loci as a method to detect small variations in the frequency, directionality and extent of individual GC events Discordant reads analysis has been established as a sensitive tool to determine small variations in the directionality and extent of GC products associated with the repair of an HO break by HR 19 . To gain further insight, we implemented this strategy in Di-GRAPH to assess GC efficiency in wild-type SS cells. Since the distribution of discordant reads delineates the two boundaries of GC events (Fig. 2A), we generated a densitometry map plotting the MAT a and MAT a´ coordinates for each discordant read pair, which allowed us to precisely infer the symmetry and extent of individual GC events and to identify the preferred regions where recombination boundaries occur (Fig. 2B). GC events accumulated at the right flank of HO site and at a distal region to the left of the break, marking the boundaries of the recombination events. We also observed an accumulation of discordant reads proximal to the left side of the HO site, likely resulting from short-range DNA repair outcomes (i.e., NHEJ or short-range HR). To facilitate the quantification of GC events distribution, we split reads in 18-mer, following the same approach described above. By quantifying the size of the MAT a/ MAT a´ reads peaks (Supplementary Fig. 3B), we can accurately measure small variations in discordancy at each MAT a´ polymorphism and thus detect subtle differences in the extension and directionality of GC events. Discordances at the right boundary of the HO break comprised the HO site (0 nt, MAT a´) and the first polymorphism to the right (+64 nt, MAT a) (Fig. 2C), indicating a drastic inhibition of GC to this side. Discordances to the left of the HO site ranged from -244 to the last polymorphism at -634 nt (Fig. 2C), reflecting variability in the length of GC products. To detect the presence of discontinuities in GC events away from the HO break, we analyzed the distribution of discordant MAT a/ MAT a´ and MAT a´/ MAT a reads to the left and right of the HO break (Fig. 2D). Discordant MAT a/ MAT a´ and MAT a´/ MAT a reads in the bottom-left and top-right quadrants represent, respectively, the left and right boundaries of GC events (Fig. 2D and 2E, blue labels). Conversely, discordant MAT a/ MAT a´ and MAT a´/ MAT a reads in the top-right and bottom-left quadrants represent, respectively, right and left discontinuous GC events (Fig. 2D and 2E, red labels). Repair occurred preferentially by continuous GC to the left, while only a small fraction repaired by continuous GC to the right (Fig. 2F). The frequency of discontinuities at the HO site was also asymmetric, occurring more often to the left than to the right. Overall, discordant read analysis enables precise mapping of GC tracts, revealing a strong leftward asymmetry at the HO cleavage site, with rare discontinuities also preferentially accumulating on this site of the break. Distinctive DNA repair mutants show unique discordant MAT a/ MAT a´ GC patterns To determine the influence of distinctive genes implicated at different levels of the HR repair pathway in GC efficiency, we applied the same approach described above to SS cells of exo1 D , sgs1 D , srs2 D and rad51 D mutants and assessed variation in frequency, length, symmetry and continuity of GC events. Loss of Exo1 displayed a profile similar to that of wild-type cells in the accumulation of discordant reads, covering nearly all polymorphisms in the left flank of the HO site, with a pronounced accumulation at the last two polymorphisms (Fig. 3A, left graph and Supplementary Fig. 3B). Consistently, the discordant read density map showed a profile comparable to that of wild-type cells, with increased density at proximal right and distal left of the HO site at the same coordinates (Fig. 3A, middle graph). No changes were observed in the frequency of the -64 polymorphism, indicating a similar proportion of cells undergoing short-range DNA repair. Likewise, the analysis of GC continuity revealed no differences relative to the wild-type strain (Fig. 3A, right graph and Fig. 3E). Both sgs1 D, and to a lesser extent, srs2 D mutants showed a substantial accumulation of discordant reads in the left distal -634 to -586 region (Fig. 3B and 3C, middle graphs and Supplementary Fig. 3B). This result was supported by the increase in discordant reads associated with the left GC boundary observed in the discordant reads density maps of both mutants (Fig. 3B and 3C, left graphs), indicating that Sgs1 and Srs2 are implicated in restraining GC length, as previously suggested for Sgs1 26 . The absence of Sgs1 or Srs2 did not alter the frequency of the -64 polymorphism, indicating no effect on the frequency of short-range DNA repair (Fig. 3B and 3C, middle graphs). Notably, both mutants exhibited an accumulation of discontinuous GC events, an effect particularly pronounced on the right side of the HO break (Fig. 3B and 3C right graphs and Fig. 3E), suggesting that these helicases are required to prevent HO-associated gaps in the recombinant products. Absence of Rad51 limited the appearance of discordant reads at distal regions of the HO break (Fig. 3D, left graph and Supplementary Fig. 3B), reflecting its inability to sustain long-range HR. However, we detected discordant reads spanning the HO site and its two flanking polymorphisms (-64 and +64), indicating that in rad51 D cells a fraction of cells can still incorporate the HO -inc from chromosome V to chromosome III. This result is further supported by the presence of MAT a´ coverage near the break site in rad51 D cells (Fig. 1D, bottom right graph). Accordingly, the density map showed an accumulation of discordant reads at the HO/HO -inc interface, with only a few events detected distal to the break (Fig. 3D, left graph). Discordancy symmetry was similar to that observed in wild-type cells, suggesting that the fraction of cells capable to sustain HR retains the ability to asymmetrically copy to the left side of the HO break (Fig. 3E). Finally, we observed a near absence of GC discontinuities (Fig. 3E), suggesting that these recombinant outcomes depend on Rad51. Altogether, these results show that discordant reads are sensitive markers to reveal how different DNA repair mutants alter the frequency, length, symmetry and continuity of GC during HR. HR repair mutants show unique HO associated mutagenic patterns upon DNA repair We have seen above that exo1 D , sgs1 D , srs2 D and rad51 D SS cells displayed different GC profiles upon HO repair. To assess the accuracy of this pathway and to determine the contribution of each DNA repair mutant to the associated mutagenic pattern, we devised an approximation within Di-GRAPH to link DNA repair by long/short-range HR or NHEJ to specific patterns of HO mutagenesis (Fig. 4A). We used the BWA aligner 27 , since its flexibility enables reliable mapping of insertions, deletions, and structural variants. 150-mer reads mapping at the HO break were filtered based on the presence of A/G 200689 at position -64 (first polymorphism left to the HO/HO- inc site) and C/T 200753 at position 0 (HO/HO- inc site) (Fig. 4A, top left diagram). Reads containing the combination T 200753 /G 200689 were associated with long-range HR (Fig. 4A, bottom left diagram), T 200753 /A 200689 with short-range HR (Fig. 4A, bottom middle diagram) and C 200753 /A 200689 with NHEJ (Fig. 4A, bottom right diagram). It is important to note that for all types of DNA repair, the combination T 200753 /G 200689 also represents reads derived from the donor MAT a´ sequence on chromosome V. We re-aligned the three subsets of filtered reads to chromosome III and quantified the percentage of mismatches for any given position within the HO recognition site (-11 and +9 relative to the HO site), both before HO induction and in SS cells. Representation of reads containing G 200689 and A 200689 at T0 for all strains revealed no associated mutations in any strain tested, indicating intact HO/HO -inc sequences prior HO induction. Importantly, the proportions of C 200753 and T 200753 reads dropped to 50%, accurately reflecting the equal representation of reads mapping to the MAT a and MAT a´ loci , respectively (Supplementary Fig. 3C). We did not detect long-range or short-range HR-associated mutations in the SS samples analyzed for any of the mutants assayed, indicating that the inherent HR defects linked to these mutants do not increase the HO mutation rate (Fig. 4B). Note that the high levels of base substitution at coordinate 0 nt reflect the C-to-T substitution intrinsic to HR repair, rather than a true mutation. We did not find A 200689 /C 200753 reads in wild-type and exo1 D cells, indicating a complete absence of NHEJ repair in these strains (Fig. 4F). By contrary, NHEJ-associated mutations increased in sgs1 D and srs2 D strains, and were markedly elevated in rad51 D mutants, with each strain displaying distinct mutagenic signatures (Fig. 4C, 4D and 4E). For instance, NHEJ-associated deletions in sgs1 D cells were accumulated between positions -1 nt and -3 nt, together with insertions at -2 nt, while srs2 D cells showed insertions at -3 nt (Fig. 4C and 4E). The rad51 D mutant exhibited extensive mutagenesis, including base substitutions and insertions from -1 to -3 nt, as well as widespread deletions across the entire DNA segment analyzed, most prominently between positions 0 and -3 nt (Fig. 4C and 4E). Regarding DNA repair pathway choice, exo1 D , sgs1 D and srs2 D mutants preferentially repaired the HO break via long-range HR, followed by short-range HR. By contrast, rad51 D mutants showed increased NHEJ and reduced long-range HR (Fig. 4F), ratifying the asymmetric MAT a´/ MAT a reads distribution obtained in figure 1D. Interestingly, short-range HR was unaffected in the absence of Rad51, suggesting that this protein is not required to sustain short GC tracts. Taken together, these results demonstrate that SS cells from different HR mutants display distinct HO-associated mutagenic signatures and DNA repair pathways choices. exo1 D , sgs1 D and srs2 D cells preferentially use long-range HR, whereas rad51 D shifts toward error-prone NHEJ without affecting short-range HR, highlighting the differential impact of HR factors on DNA repair accuracy. Genome-wide sequencing analysis reveals altered coverage of distinctive genomic features in response to an HO break We have shown above that HO repair levels in the PMV background do not correlate with cell viability (Supplementary Fig. 1A), suggesting that there might exist other sources of genomic instability that affect cell recovery after HO repair. To address this, we used the genome-wide sequencing data to look for genomic coverage changes between NSS and SS wild-type cells within the Di-GRAPH framework. The rationale behind this experimental approach is that lethal genomic rearrangements occurring during the repair of the HO break in NSS might be counter-selected when screening SS cells that have overcome the DNA lesion. We also include in the analysis an undamaged raffinose culture to identify DNA damage-independent genomic reorganizations. We normalized the mean coverage of each of the 14518 genetic elements in NSS, SS and UD cells to their respective uninduced T0 coverage values. Coverage ratios exceeding 1 indicate a gain, whereas ratios below 1 indicate a loss, relative to the uncut HO state. For downstream bioinformatic analysis and data visualization, all genomic elements were subsequently grouped into 13 genomic categories comprising ORF, intergenic region, LTR, TEG, Ty, tRNA, rRNA, ncRNA, snRNA, snoRNA, ARS, centromere and subtelomeric region (Supplementary Fig. 3D). Analysis of NSS cells showed that most of the genomic features within each category were very stable (as denoted for coverage values close to 1), except for the LTR ( L ong T erminal R epeats), TEG ( T ransposable E lement G enes) and Ty elements (Ty retrotransposons), where a drastic increment in coverage levels was detected (Fig. 5A, top graph). This coverage enrichment was counter-selected in SS cells (Fig. 5A, middle graph), suggesting that elevated levels of LTR, TEG and Ty during the damage response affects cellular viability. This graph resembles that obtained from UD cells (Fig. 5A, bottom graph), indicating a specific DNA damage-dependent effect on the genomic changes observed during the induction of the HO break. To precisely quantify DNA damage-specific genome-wide coverage changes, we plotted the ratio of NSS and SS samples relative to their corresponding UD sample. NSS displayed a 1.84% increase (C Wt > 1.2) and a 0.73% decrease (C Wt < 0.8) across all genomic elements analyzed (Supplementary Fig. 5A, top graphs), revealing an amplification of genome size, likely due to the hyper-recombinogenic state of DNA damaged cells. These values decreased to 1.23% and 0.3%, respectively, in SS cells (Supplementary Fig. 5A, bottom graphs), suggesting that most of these alterations are incompatible with cell viability. To determine whether these changes were randomly distributed across the categories analyzed, we examined both the distribution of raw values from these samples (Fig. 5B) and the percentage of genomic changes within each genomic category (Fig. 5C). Most of the coverage enrichments in NSS cells originated from increases in LTRs, TEGs, Ty elements, and tRNAs, likely reflecting retrotransposition activity associated with DSB repair 28-30 (Fig. 5B,C, top graphs). We also observed a slight increase in subtelomeric regions and a reduction at centromeres. All these coverage changes were restored in SS cells (Fig. 5B,C, middle graphs), indicating that lethality of NSS is primarily associated with alterations in mobile elements, centromeres and subtelomeric regions. Importantly, although reduced in SS cells, mobile elements still displayed a residual signal, suggesting that DNA damage-dependent retrotransposon mobility might be buffered by the cell. These changes were DNA damage-dependent, as only small variations were detected in UD cells (Fig. 5B, bottom graph), except for subtelomeric regions (Fig. 5C, bottom graph). Next, we asked whether the genomic changes observed in NSS cells were uniformly distributed across chromosomes. Consistent with the global coverage increase in coverage following HO induction (Supplementary Fig. 5A, top graph), most chromosomes exhibited a tendency toward increase in coverage levels, with some chromosomes more prone to enrichment than others (Fig. 5D, top graph). It is important to note that both MAT a and MAT a´ loci were excluded from the analysis to avoid chromosome III/V coverage artifacts resulting from HR repair. These results indicate that genome reorganization in response to a HO break is not equally distributed across all chromosomes. As before, although a residual tendency toward coverage enrichment was still detectable, most of these variations were reduced in SS and UD cells (Fig. 5D, middle and bottom graph). Overall, these data suggest that the genome exhibits a degree of plasticity in response to a DSB. Genomic changes vary across genomic categories and chromosomes, implying that certain regions are more susceptible to alteration, likely due to local genomic context or regulatory influences during DNA repair. exo1 D , sgs1 D , srs2 D , and rad51 D cells exhibit distinct genomic coverage profiles in response to DNA damage To investigate the genomic consequences of an HO-induced break in the absence of stage-specific factors of the HR pathway, we applied the same experimental approximation described above to the exo1 D , sgs1 D , srs2 D and rad51 D mutants. NSS exo1 D cells displayed distinct coverage changes compared with the wild-type (C exo1 D > 1.2 = 3; C exo1 D < 0.8 = 0.57), with a general tendency toward increased coverage levels (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs), particularly at mobile elements and centromeres (Fig. 6A and Supplementary Fig. 5B). Conversely, decreased coverage was observed at rDNA loci and, irregularly, at the subtelomeric regions (Fig. 6A and Supplementary Fig. 5B). Gain in chromosome coverage appeared randomly distributed, with a slight enrichment on chromosome I (Fig. 6B, top graphs). As in wild-type cells, SS exo1 D exhibited a pronounced reduction in coverage alterations (C exo1 D > 1.2 = 1.19; C exo1 D < 0.8 = 0.34) when compared with NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), across all chromosomes (Fig. 6B, middle graphs), and most notably at mobile elements (Fig. 6A and Supplementary Fig. 5B). This suggests that retrotransposition in the exo1 D background contributes to cell lethality following DNA repair. Notably, the reduction in rDNA coverage was maintained in SS cells (Fig. 6A and Supplementary Fig. 5B), indicating that this feature does not compromise cell viability in response to DNA damage. No major coverage alterations were detected in UD exo1 D cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), except for a minor, randomly distributed signal at subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs). These findings suggest that Exo1 is necessary to prevent mobile element transposition events and to maintain rDNA copy number following DNA damage. NSS sgs1 D cells exhibited coverage profiles that differed from those of the wild-type, reflecting highly dynamic coverage variations (C sgs1 D > 1.2 = 1.77; C sgs1 D < 0.8 = 1.59) (Supplementary Fig. 5A, top graphs). Increased coverage was detected at mobile elements (Fig. 6A and Supplementary Fig. 5B), randomly affecting all chromosomes (Fig. 6B, top graphs). Interestingly, SS sgs1 D cells still displayed pronounced coverage alterations (C sgs1 D > 1.2 = 1.7; C sgs1 D < 0.8 = 0.86) compared with NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), suggesting the accumulation of non-lethal genomic alterations. Mobile elements and subtelomeric regions displayed a markedly reduced coverage, whereas tRNA, ncRNA and rRNA showed pronounced increases (Fig. 6A and Supplementary Fig. 5B), both patterns occurring randomly distributed across all chromosomes (Fig. 6B, middle graphs), indicating widespread genomic imbalances. Unlike the wild-type, sgs1 D cells also showed pronounced coverage changes in UD conditions (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B). These variations spanned multiple genomic categories, with particularly strong effects at rDNA and subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs). Collectively, these results demonstrate that Sgs1 is critical for preserving genome integrity, even in the absence of DNA damage. NSS srs2 D cells displayed pronounced coverage changes compared with the wild-type (C srs2 D > 1.2 = 2.28; C srs2 D < 0.8 = 2.39), with both gains and losses in coverage (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs). Coverage increases were enriched at mobile elements and subtelomeric regions (Fig. 6A and Supplementary Fig. 5B), whereas losses were predominately observed at centromeres (Fig. 6A and Supplementary Fig. 5B). Chromosome coverage gains appeared largely random, with a slight enrichment on chromosome I (Fig. 6B, top graphs). SS srs2 D cells showed a marked reduction in coverage alterations (C srs2 D > 1.2 = 1.29; C srs2 D < 0.8 = 0.20) relative to NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), for all chromosomes (Fig. 6B, middle graphs) and most evident at mobile elements (Fig. 6A and Supplementary Fig. 5B), indicating that retrotransposition in the absence of Srs2 contributes to cell lethality following DNA repair. Similar to sgs1 D cells, SS srs2 D cells displayed an increase in rDNA coverage (Fig. 6A and Supplementary Fig. 5B), suggesting that although both helicases suppress rDNA expansion, this property does not impair cell viability under DNA damage. No major coverage alterations were detected in UD srs2 D cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), aside from a minor decrease at centromeres and a localized increase at the subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs), with the strongest effect on the recipient chromosome III (Fig. 6B, bottom panel). Taken together, these results demonstrate that Srs2 plays a critical role in preventing retrotransposition and maintaining centromere stability in response to DNA damage. NSS rad51 D cells displayed the most pronounced coverage changes of all strains analyzed (C rad51 D > 1.2 = 2.13; C rad51 D < 0.8 = 2.87), with both gains and losses occurring in a pattern similar to that observed in srs2 D cells (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs). Coverage gains were enriched at mobile elements and subtelomeric regions (Fig. 6A and Supplementary Fig. 5B), whereas losses were more prominent at snoRNAs, ARSs, centromeres and subtelomeric regions, and to a lesser extent to other genomic categories (Fig. 6A and Supplementary Fig. 5B). Chromosome coverage gains appeared largely random, with a slight enrichment on chromosome VI, while coverage losses were distributed uniformly across all chromosomes (Fig. 6B, top graphs). SS rad51 D cells showed a marked reduction in coverage alterations (C rad51 D > 1.2 = 0.92; C rad51 D < 0.8 = 0.43) relative to NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), extending across all chromosomes (Fig. 6B, middle graphs), while subtelomeric regions still exhibited slightly high coverage levels (Fig. 6A and Supplementary Fig. 5B). Notably, unlike the other strains analyzed, mobile elements in SS cells were completely downregulated, indicating a severe impact of retrotransposition on lethality in the absence of Rad51 (Fig. 6A and Supplementary Fig. 5B). No major coverage alterations were detected in UD rad51 D cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), aside from a minor decrease at centromeres and an increase at subtelomeric regions, again resembling the srs2 D mutant (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs), with no bias against any specific chromosome (Fig. 6B, bottom panel). Taken together, these results indicate that Rad51 deficiency produces a coverage profile similar to that of srs2 D cells, suggesting that both mutants have similar consequences for genome stability in response to DNA damage. Genome-wide analysis of discordant reads reveals global genome reorganizations in response to a DSB Coverage-based approaches are useful for identifying prominent biases in the duplication or depletions of genomic elements but cannot determine their relocation within the genome. Moreover, such changes must accumulate iteratively at the same loci for coverage gains or losses to be detectable and attributable to specific genomic elements. To address this limitation, we implemented Di-GRAPH with a module to apply the same discordant reads analysis 19 used above to map GC boundaries, this time with the aim of detecting global recombination events across the entire genome. Reads from T0, NSS, SS and UD wild-type cells were aligned using Bowtie2 avoiding multimapping, and discordant reads were filtered by requiring both mates of the pair to align to different chromosomes with no mismatches allowed. For internal normalization, we used a custom reference genome containing an intact ADE3 gene on chromosome VII and an artificial chromosome carrying the pGal1-10-HO-SSB1-vector sequence (chromosome Disc), identical to the construct integrated at the ADE3 gene at chromosome VII in the PMV background (Supplementary Fig. 6A). In addition, the endogenous pGAL1-10 and HO-SSB1 sequences were replaced with Ns to prevent removal of duplicated reads mapping to both loci in the reference genome. Under this set up, concordant reads at the ADE3-Vector and pGAL1-10-ADE3 interfaces appear as discordant reads between chromosome VII and the chromosome Disc (Supplementary Fig. 6A), thereby enabling the quantitative normalization of discordances across the genome (Disc-normalization). This normalization was applied to all subsequent analysis of discordant reads. Polymerization errors during library amplification or sequencing mistakes might generate erroneous reads that, if misalign relative to their pair, may produce false discordant reads (Fig. 7A). To minimize these artifacts, we implemented a bioinformatic approach in which, for a true discordant pair, both reads should not concordantly re-align under relaxed BLAST alignment conditions (Fig. 7A). In this scenario, concordant re-alignments increase the likelihood of identifying a false discordant pair (Fig. 7A). We filtered read pairs and classified them based on their ability to remain discordant (no alignment, hereafter NA), along with those containing ³3, 2 or 1 nucleotide variants that rendered them to a concordant state, in order to evaluate the reliability of these discordances (Fig. 7A). Finally, artifactual discordances arising from variations between the reference genome and the genetic PMV background were filtered by removing discordant read pairs detected at the T0 timepoint in any of the three experimental replicates. Each read of the pair from the NA group was plotted in a dot array comparing NSS and SS samples (Fig. 7B, left panel, left-bottom and right-top triangle respectively), mapping the two genomic features involved in each recombination event, with signal intensity reflecting the frequency of the rearrangement. The most abundant discordances appeared between two ORFs, followed by ORF-intergenic regions. Consistent with the increase in mobile elements and subtelomeric coverage observed in NSS cells (Fig. 5B and 5C), we detected higher frequency of discordancy in these regions (Fig. 7B, bottom right graph) compared to SS cells. Each genomic category showed a prevalence of discordances at ORFs and intergenic regions, followed by the rDNA and ARS sequences. Interestingly, rDNA discordances increased specifically in SS cells, indicating non-lethal translocation events in this genomic region in response to DNA damage (Fig. 7B, top right graph). Successive relaxation of the discordance confidence score slightly increased discordances at mobile elements (Ty, TEG ad LTR) and subtelomeric regions only in the NSS sample, again supporting the presence of lethal transposition events during the damage response (Supplementary Fig. 7A). Together, these results indicate that discordancy analysis is a reliable method to assess genome rearrangements during the damage response and suggest that ORF, intergenic regions, rDNA, transposons, ARS and subtelomeric translocations are the most frequent genomic rearrangements occurring during the induction of a DSB. Genome-wide analysis of discordant reads reveals distinctive patterns of genome reorganization in the absence of DNA repair factors To evaluate how different steps of the repair pathway influence genome integrity, we applied the experimental workflow described above to exo1 D , sgs1 D , srs2 D and rad51 D cells under NSS, SS and UD conditions (Fig. 8A and Supplementary Fig. 6B). To minimize false-positive discordance, we restricted the analysis to filtered pairs from the NA group. Additionally, to visualize the genomic localization of these events, we plotted genomic maps representing the discordance frequency of every genomic feature that was detected in our three replicates (Fig. 8B and Supplementary Fig. 6C, D). Supporting the tendency toward rDNA coverage decay, loss of Exo1 resulted in low discordances at this locus in NSS cells (Fig. 8A). We also detected a slight increase in discordance levels at ORF and intergenic regions, especially in UD cells, supporting a DNA damage-independent role for the nuclease in maintaining genome stability (Fig. 8A). This discordancy was evenly distributed along all genomic features comprising these structures (Supplementary Fig. 6D). Transposons showed a modest accumulation in SS samples when the discordance confidence score was relaxed (Supplementary Fig. 7B), indicating certain accumulation of Ty transposition following DNA repair. By contrast, the absence of Sgs1 leads to a drastic accumulation of ORF and ARS discordances in SS cells, accompanied by a rDNA reduction in NSS conditions (Fig. 8A,B and Supplementary Fig. 6C,D), indicating a specific role for Sgs1 in maintaining these genomic structures, particularly during the damage response. Discordances in sgs1 D cells were uniformly distributed across the entire genome, apart from a strong single hotspot located in an intergenic region that appeared under both NSS and SS conditions but was absent in the UD sample (Fig. 8B and Supplementary Fig. 6C,D). Supporting coverage data, sgs1 D cells also led to accumulation of Ty and TEG elements in NSS cells (Fig. 8A,B and Supplementary Fig. 6B), an effect that became dramatically enhanced under relaxed confidence score settings (Supplementary Fig. 7C, left panel). This elevated transposition activity is likely deleterious, as it was drastically reduced in SS cells (Fig 8A, Supplementary Fig. 6B,C and Supplementary Fig. 7C middle panel). Similarly to sgs1 D cells, a srs2 D mutant also exhibited high levels of transposition events under NSS conditions (Fig. 8A, Supplementary Fig. 6B). However, unlike sgs1 D , srs2 D cells retained elevated transposon-associated discordances in SS samples under relaxed confidence score conditions (Supplementary Fig. 7D), suggesting that this helicase is essential for preventing retrotransposition in response to a DSB. Consistent with the coverage records, loss of Rad51 resulted in increased Ty, TEG and LTR discordances mainly under NSS conditions (Fig 8A,B, Supplementary Fig. 6B and Supplementary Fig. 7E), indicating that transposition becomes lethal when HR is impaired. Importantly, some Rad51-dependent NSS discordances generated strong hotspots across distinct ORFs and intergenic regions, indicating fragile genome sites in the absence of this repair factor (Fig. 8B). rad51 D cells also displayed a high frequency of ORF-, ARS- and ncRNA-associated discordances in UD samples (Fig. 8A), highlighting a genome-stabilizing function of Rad51 even in the absence of DNA damage. Identifying strain-specific genome rearrangement networks via discordant reads analysis To detect genomic regions with a high prevalence of rearrangements, we filtered the reads used above to retain only those mapping to loci with discordant frequency higher than 6%. Each discordant pair from the assessed strains was assigned a distinct color, allowing the identification of strain-specific discordance patterns. We then connected the two reads of each pair to construct a discordance network of NSS, SS and UD (Fig. 8C, Supplementary Fig. 8A,B). To quantify the interactions observed in this network, we calculated the percentage for each discordant pair, generating an interaction map that illustrates the recombination probability between all pairs of genomic categories for each condition tested (Fig. 8D). Consistent with the high levels of retrotransposition observed earlier, several Ty and TEG elements occupied central positions in the NSS network, particularly in the sgs1 D and srs2 D mutants (Fig. 8C), being transposon-ORF the most frequent recombination events (Fig. 8D). This effect was particularly evident in the rad51 D strain, which also showed transposon-transposon recombination (Fig. 8D). Recombination at the rDNA cluster was also common, occurring in all strains except sgs1 D . The rDNA cluster appeared largely separated from the rest of the network, suggesting that rDNA rearrangements occur independently of the elevated transposition state characteristic of the NSS condition (Fig. 8C). The primary recombination partners for rDNA were ORFs and intergenic regions (Fig. 8D). ARS-associated discordances in sgs1 D and rad51 D predominantly involved ORF and LTRs (Fig. 8D). ORFs and intergenic regions were located at the periphery of the network (Fig. 8C), often in a strain-specific manner, typically recombining with other ORFs, and in the case of sgs1 D , srs2 D and rad51 D strains, also with transposons. Interestingly, exo1 D and sgs1 D cells showed a marked reduction in ORF-associated discordance (Fig. 8D), implicating these factors in the rearrangements of coding regions. In contrast to NSS conditions, SS cells exhibited minimal Ty-associated recombination, indicating that retrotransposition in response to DNA damage is largely lethal (Fig. 8D and Supplementary Fig. 8A). Notably, the rDNA cluster shifted to the center of the network in SS cells, a pattern observed across all strains (Supplementary Fig 8A). The stabilization of rDNA recombination in SS cells confirms that these events are not inherently lethal. The absence of Sgs1 led to a pronounced increase in ORF-ORF and ORF-intergenic recombination (Fig. 8D), indicating that this helicase in essential for preventing ectopic rearrangements at these genomic regions following a DSB. Similarly to SS conditions, UD cells also showed reduced transposition activity (Fig. 8D and Supplementary Fig. 8B), reinforcing DNA damage as the primary source of retrotransposition. Consistent with the elevated discordance frequencies observed in exo1 D cells in the absence of damage, this mutant dominated the UD network, confirming a DNA damage-independent role for Exo1 in genome maintenance. exo1 D cells displayed enhanced ORF-ORF and ORF-intergenic recombination, with similar, though less pronounced, patterns in rad51 D cells (Fig. 8D). Interestingly, srs2 D cells exhibited a marked reduction in ORF-associated discordances under UD conditions (Fig. 8D), suggesting that this helicase may be important for facilitating or regulating the ectopic recombination events that arise in the absence of DNA damage. Integration of coverage and discordant read data enables precise characterization of recombination identity As shown above, both coverage measurements and discordant-reads analysis can detect copy number variations of specific genomic features and track their translocation across the genome. To evaluate whether these approaches could be integrated to determine the nature of the recombination events, we grouped NSS, SS and UD data and collectively compared coverage values with discordance ratios (Fig. 8E). Coverage and discordance were generally well correlated, indicating that translocations commonly involve genomic amplification arising from recombination. Most correlated features corresponded to transposons, whose amplification during HO induction may enhance their translocation potential. Some ORFs and intergenic regions also showed strong correlations. Notably, the presence of paralogues for these coding regions in the genome contributes to their recombination bias. A subset of features showed high coverage but low discordance (Fig. 8E), consisting mainly of Ty elements whose elevated amplification does not necessarily result in integration, or whose rearrangements may be difficult to detect using our discordance-based assay. Conversely, features displaying low coverage but high discordance were largely represented by the non-transcribed rDNA region NTS1-2 (Fig. 8E). This pattern suggests that rDNA changes during the DNA damage response reflect translocation of rDNA units rather than expansion or contraction of the rDNA array. To evaluate whether the HO-induced global rearrangements observed across the analyzed genomic categories depended on homology-directed recombination (HDR), we examined the set of 110 highly discordant features used before (Fig. 8C and Supplementary Fig. 8A,B) and computed their random pairwise sequence identity relative to every other feature in the genome (Fig. 8F, blue line). We then calculated the sequence identity for each actual discordant pair, representing the true homology between the two features involved in the rearrangement (Fig. 8F, red line). The ratio between the observed and random identity values reflected the degree of HDR. Higher ratios correspond to rearrangements requiring greater sequence similarity, whereas lower ratios indicated recombination events that tolerate reduced homology. Strikingly, ORFs, which displayed the lowest random identity, showed the highest HDR ratios, suggesting that coding regions are highly selective with respect to sequence homology during recombination. In contrast, LTRs, intergenic regions and rDNA, all of which exhibited higher random identity values, displayed lower homology dependence, implying that recombination at these loci requires comparatively less sequence similarity. Retrotransposons showed a more variable pattern but generally trended toward higher HDR, consistent with the notion that most of their insertion events involve other retrotransposons and in agreement with the interaction network shown in figure 8C and 8D. In summary, these results demonstrate that integrating coverage profiles with discordant read analysis provides a powerful framework for characterizing the nature of recombination events at a genome-wide scale. This combined approach automated within Di-GRAPH does not only enables the detection and classification of rearrangements with high resolution but also reveals how distinct genomic features behave during the cellular response to DNA damage. By capturing both the structural outcomes and the underlying homology relationships, this method offers a comprehensive view of the recombination landscape and the dynamic contributions of different genomic elements during repair. Discussion Repair of DSBs by HR requires the coordinated execution of multiple steps, yet how defects in individual HR components shape genome integrity beyond the DNA break remains poorly understood. Using a defined HO-induced break in the S. cerevisiae PMV background, we developed Di-GRAPH, a user-friendly bioinformatic software that integrates DSB mutational signature analysis, repair pathway choice, coverage profiling, and discordant read mapping to quantitatively measure both local and genome-wide structural rearrangements. By applying Di-GRAPH to cells deficient in Exo1, Sgs1, Srs2 and Rad51, we determined the influence of each stage of the HR pathway at the cleavage site as well as genome-wide. High-resolution Di-GRAPH profiling of GC tracts demonstrates that these outcomes are not only strain-specific, but that the associated HO repair signatures are stabilized in the survivor population. Wild-type and exo1 D cells displayed nearly identical GC tract lengths and asymmetry, indicating that Exo1-mediated long range resection is largely dispensable for GC formation, in line with previous results 24 . In contrast, sgs1 D and to a lesser extent srs2 D mutants, accumulate extended GC tracts, consistent with their anti-recombinogenic activities in disassembling D-loop intermediates 9-11 . Notably, both mutants also displayed elevated discontinuities within GC tracts, particularly on the right flank of the break, suggesting that Sgs1 and Srs2 are required to prevent inappropriate template switching intermediates during HR. Nevertheless, since both helicases have been reported to bias DNA repair toward non-crossover outcomes 31 , it is plausible that these GC discontinuities represent crossover products that accumulate at the HO cleavage boundary during DNA repair. As expected, long-range GC events were drastically reduced in rad51 D cells, although these cells retained the characteristic left-side asymmetry, indicating that Rad51 is not required for generating GC asymmetry. Consistent with the established role of Rad51 in HR, NHEJ products accumulated in SS cells in this mutant. Remarkably, short-range GC at the HO site was unaffected by the absence of Rad51, demonstrating that this type of repair is Rad51-independent. In agreement with these findings, previous work has shown that the loss of Rad51 stimulates allelic GC and break-induced replication by approximately 30-fold in a Rad52-dependent manner 32 . Determining whether short-range GC depends on Rad52 remains an important question for future investigation. Another feature of Di-GRAPH is that it enables the quantification of repair pathway choice and the mutational patterns associated with long- and short-range HR, as well as with NHEJ. Wild-type and exo1 D cells repaired the HO break almost exclusively through HR, without detectable mutagenesis. In contrast, sgs1 D and srs2 D mutants displayed a slight increase in NHEJ usage accompanied by distinct mutational signatures (short deletions and insertions localized to the -1 to -3 region) suggesting that these helicases might also function as NHEJ inhibitors. Importantly, and in line with recent discoveries showing that binding of Rad51 to the break prevents single-strand annealing (SSA) or alternative end-joining (A-EJ) DSB repair 33 , rad51 D cells exhibited the most severe shift toward NHEJ and accumulated extensive base substitutions and indels along the HO cleavage site. These results highlight that loss of stage-specific HR factors induces distinct mutational patterns, emphasizing that genome integrity is determined not only by whether breaks are repaired, but also by how cells redistribute repair between alternative pathways when canonical HR steps are compromised. In addition to analyzing DSB repair at the break site, Di-GRAPH is integrated with a genome-wide coverage analysis that identifies DNA damage-dependent changes of coverage across distinctive genomic categories. This analysis revealed widespread alterations in rDNA, mobile elements, centromeres, tRNAs, ARSs, and subtelomeric regions following HO induction. NSS samples exhibited extensive gains in LTR, Ty elements and TEGs, consistent with the activation of retrotransposition during the DNA damage response 28 . The dramatic reduction of these rearrangements indicates that retrotransposition-induced lesions are frequently lethal and subjected to negative selection during DSB recovery 34 . Subtelomeric expansions and centromere losses, observed in NSS but absent in SS, further highlight the inherent fragility of repetitive or structurally constrained regions during HR-mediated repair. Nevertheless, each HR mutant exhibited a distinct genome-wide signature, indicating that these factors differentially contribute to genome stability at sites distant form the DSB. Loss of Exo1 showed pronounced increases in mobile elements and a reduction in rDNA coverage, suggesting that this nuclease both restrains retrotransposition and preserves rDNA structure following DNA damage. In agreement with the role of Sgs1 in preventing duplication-mediated gross chromosomal rearrangements 35 , loss of this helicase displayed persistent genome imbalance even in undamaged conditions. Interestingly, this role was particularly pronounced for the rDNA, which exhibited the most substantial alterations among all features analyzed in this mutant. In this context, it has been reported that rDNA is permissive to RAD51 accumulation in the absence of BLM (the human homolog of Sgs1), leading to rDNA micronucleation and potentially contributing to global genome instability 36 , a phenotype reminiscent of the coverage profile observed in the absence of Sgs1. The absence of Srs2 or Rad51 led to extensive coverage gains in mobile elements and losses in centromeric regions, displaying highly similar profiles, suggesting shared vulnerabilities in suppressing genome rearrangements in these mutants. Since Srs2 is required to remove Rad51 from ssDNA 37 , it is plausible that the absence of Srs2 exhausts the pool of free Rad51, providing a potential explanation for the similar phenotypes observed in both mutants in genome-wide analyses. Interestingly, the complete suppression of these elevated mobile element coverage signal in rad51 D SS cells indicates that such events are lethal to the cell. These data demonstrate that genome instability arising during DSB repair is not uniform but is instead strongly influenced by the specific HR step that is impaired, with mobile elements, subtelomeric regions, and centromeres emerging as major contributors to cell lethality. Applying discordant read mapping across the genome allowed direct detection of recombination events, overcoming limitations of coverage-based approaches, particularly their low sensitive and lack of spatial information. Several bioinformatic approaches have been developed to use discordant read as a source to detect genomic rearrangements 14,38 . Di-GRAPH incorporates a novel algorithm that not only identifies chromosomal rearrangements based on the paired-end reads discordance, but also evaluates the probability that these events represent true rearrangements rather than false-positive discordant reads generated during library preparation or DNA sequencing. Moreover, the use of a Disc-normalization allows precise quantification of these events, making this approach well suited for assessing differences in genome-wide rearrangements across distinct strains or experimental conditions. This analysis reveals widespread translocations involving ORFs, intergenic regions, rDNA, ARS sequences and mobile elements during the repair of a single HO break. The strong accumulation of discordant reads in mobile elements and subtelomeric regions in NSS, but not in SS cells, parallels the coverage analysis, validating this approach and suggesting that element mobility frequently results in duplications of the recombination regions involved. Importantly, rDNA rearrangements were enriched specifically in SS cells, suggesting that certain rDNA translocations are tolerated or even selectively retained following repair. In contrast, rearrangements involving mobile elements were largely eliminated in SS samples, consistent with their detrimental impact on cell survival following DNA repair 14 . Discordant read mapping in HR mutants further emphasized the diversity in genome reorganizations that arise when specific repair steps are compromised, reveling locus - and chromosome-specific patterns. Finally, our analysis of HDR reveals key insight into the mechanisms driving HO-induced genome rearrangements. By comparing observed versus random sequence identity of discordant pairs, we found that ORFs, despite their low baseline similarity, exhibited the highest HDR ratios, indicating that recombination in coding regions is strongly constrained by sequence homology. In contrast, LTRs, intergenic regions, and rDNA, which have higher inherent similarity, showed lower HDR dependency, suggesting that recombination at these loci tolerates reduced homology. Retrotransposons displayed a more variable pattern but generally relied on HDR, consistent with recombination among related elements, as supported by the discordant interaction networks. Thus, the interplay between local sequence similarity and the genomic context emerges as a central factor governing recombination outcomes. Overall, our findings show that the consequences of a single DSB extend well beyond the break site and are strongly and distinctively influenced by stage-specific HR factors. By integrating local repair signatures with genome-wide alterations we have revealed that global genome instability, rather than DSB repair efficiency alone, is an important determinant for cell survival following DNA damage. Finally, our results stablish Di-GRAPH as a powerful framework for dissecting the consequences of a DSB both at the break site and genome-wide, providing a versatile framework for studying genome maintenance in diverse biological contexts. Methods Yeast strains and growing conditions The strains used in this work are listed in Supplementary Table 1. Disruption of endogenous genes was achieved by gene targeting using PCR products as described previously 39,40 . Details of PMV construction have been described in Ramos et al., 2022 24 . PMV was HO-induced by adding 2% galactose to cells growing in YP with 2% raffinose. Samples were collected for DNA analysis before and at distinctive timings after adding galactose to the media. Chemiluminescent Southern blotting Cell cultures (10 ml) adjusted to an OD 600 of 0.4 were collected by centrifugation and subsequently washed with 1 ml of PBS. Following a second centrifugation step, pellets were flash frozen and stored at -80ºC. Cell lysis was carried out by incubating the pellets for 10 min with 40 units of lyticase in DNA extraction buffer (1% SDS, 100 mM NaCl, 50 mM Tris-HCl, 10 mM EDTA). Genomic DNA was isolated by extracting the lysates with phenol:chloroform:isoamylalcohol (25:24:1) for 10 min. After centrifugation, the aqueous phase was ethanol-precipitated and the DNA pellet was resuspended in water. Purified genomic DNA was digested with Eco RI, resolved on 1% agarose gels, and analyzed by Southern blotting. Hybridization probes were generated by labeling a PCR-amplified DNA fragment using a nucleotide mixture containing fluorescein-12-dUTP (Fisher Scientific, 10354280). The oligonucleotides used for probe synthesis are listed in Supplementary Table 2. Detection was performed with an anti-fluorescein antibody (F(ab´)2 Anti-FLUORESCEIN Alkaline Phosphatase Conjugated, Life Technology, 700-105-096) diluted 1:200,000. Membranes were developed using the CDP-Star chemiluminescent substrate (Amersham, RPN3682) and visualized by exposure to autoradiography films. Genomic sequencing Genomic DNA (1mgr) was used as input for sequencing library construction and subsequently sequenced on a DNBSEQ G400 platform in 150 paired-end mode (BGI) to obtain 2GB data per sample, accounting for 8M median reads per sample. In brief, DNA molecules were enzymatically fragmented and processed through standard end-repair and 3´-adenylation steps prior to adaptor ligation. Adaptor-modified fragments were amplified by PCR, after which the products were converted into single-stranded circular templates and subjected to rolling-circle amplification to generate DNA nanoballs (DNBs). Sequencing was carried out using the combinatorial probe-anchor synthesis (cPAS) method. Raw reads were filtered to remove adaptor-contaminated or low-quality sequences using SOAPnuke with the manufactured-recommended parameters (-n 0.001 -l 20 -q 0.4 --adaMis 3 --rmdup –minReadLen 150). Sequenced reads were aligned to the S. cerevisiae PMV reference genome (Supplementary Data, RG_PMV_v9.fasta) for downstream analysis. For visualization of the BedGraph 41 coverage profiles we used the Integrative Genomics Browser IGB (v9.0.2) 42 . Figures and analyses were done using the following R packages: dplyr (v1.1.4) 43 , tidyr (v1.3.1) 44 , stringr (v1.6.0) 45 , ggplot2 (v3.5.2) 46 , purrr (v1.0.4) 47 and ggbreak (v0.1.6) 48 . HTML report was generated using the following R packages: rmarkdown (v2.29) 49 and flexdashboard (v0.6.2) 50 . For each sample assessed, three independent experimental replicates were sequenced. Software development and data analysis We developed Di-GRAPH, a user-friendly software that automates i) the analysis of gene conversion efficiency, extent and symmetry at an HO-induced break; ii) the assessment of repair pathway choice an HO-associated mutagenic patterns; iii) genome-wide coverage profiling to evaluate genome instability in response to HO induction; and iv) genomic discordance analysis to identify and map genome-wide chromosomal rearrangements occurring during HO repair. Di-GRAPH is open source and publicly available at https://github.com/acb-lab/Di_GRAPH, together with instruction for its installation and usage. For this study, Di-GRAPH was run as follows: Di-GRAPH.sh -b $CONDA_PREFIX/files/BLAST -c $CONDA_PREFIX/files/Categories -g $CONDA_PREFIX/files/RG -i working_directory/ -r $CONDA_PREFIX/files/Report_files The following sections of the methods describe in detail the different steps and software that compose Di-GRAPH pipeline. Coverage and Polymorphism Analysis. The sequencing reads for each sample and experimental replicate, FASTQ files were concatenated and reads were sequentially trimmed at fixed-length fragments using cutadapt 51 (v5.1) to generate two datasets: 1) 75nt fragments for coverage profiling, and 2) 18nt fragments for polymorphism quantification. Quality filtering was performed using fastp (v0.25.0) 52 (Q30 threshold) and reads were then aligned to S. cerevisiae PMV reference genome (Supplementary Data, RG_PMV_v9.fasta) using Bowtie (v1.3.1) 25 (75nt: -m 1 -v 0; 18nt -k 1 -v 0). SAM files were converted to sorted BAM format with SAMtools (v1.22) 53 , and genome coverage profiles were computed using bamCoverage (RPGC normalization, Deeptools v3.5.6 54 ). BedGraph files were sorted using Bedtools, and per-base coverage values were extracted for chromosome III and V. Coverage profiles for whole-chromosome views were generated by binning data into 100 bp windows. Finally, combined datasets for each chromosome and condition were averaged, and coverage heatmaps and polymorphism plots were generated. WGS coverage profiling. Coverage-based enrichment across genomic features categories was quantified using Di-GRAPH. For each strain and time-point (T0, NSS, SS and UD), coverage profiles generated from 75nt mapping reads (see “Coverage and Polymorphism Analysis” section) were cross-examined against predefined genomic annotations. Genomic features sets were provided as category-specific BED/TSV files. For each experiment (E1-E3), per-chromosome coverage vectors were reconstructed by mapping read-depth values to the full coordinate span of each chromosome. Then, per-base coverage for each annotated feature region was calculated, aggregating values within each feature class. For each category and chromosome, the summed coverage of NSS, SS and UD samples was normalized to its corresponding T0 value to produce experiment-specific ratios. Ratios were then averaged across biological replicates, and standard deviations were calculated. Category enrichment was visualized as barplots with error bars and violin/boxplot distributions by using ggplot2 under R v4.4.3 environment. Coverage distribution along the entire genome was displayed using genome-wide barplots with positional segmentation, including annotated boundaries between major genomic categories. Threshold-based enrichment metrics were computed to quantify the proportion of sites exceeding ratio >1.2 or falling below <0.8 within each category, as well as their associated variability. Chromosome-specific enrichment analysis were further performed by sorting feature data by chromosome (I-XVI), and calculating the percentage of genomic features above or below coverage thresholds per chromosome. HO-associated mutagenic profile. Pair-ended 150nt sequencing reads were quality-filtered (Q³30) using fastp and subsequently aligned to the S. cerevisiae PMV chromosome III reference genome (Supplementary Data, RG_PMV_v9_CHRIII.fasta) using BWA-MEM (v0.7.19) 27 . Reads mapping specifically to the HO cleavage site (position 200753 on chromosome III) were extracted with SAMtools. Aligned reads were further processed to isolate those containing specific nucleotides at positions 200689 and 200753 (see figure 4A for details) using pysam (v0.23.1) 53 . Insertions, deletions and base substitutions were quantified with IGVtools (v2.17.3) 55 and BCFtools 53 , (v1.22), and variant frequencies were normalized to total read counts. Repair pathway contributions were inferred by calculating the proportion of reads corresponding to NHEJ (A 200689 /C 200753 ), short-range HR(A 200689 /T 200753 ) and long HR (G 200689 /T 200753 ). Data aggregation, quantification analysis, and visualization of the repair pathway distributions and mutagenic rates calculation and data visualization were performed using the R packages specified above. DNA sequence logos were performed using ggseqlogo R package (v0.2) 56 . Discordant-reads-based analysis of Gene Conversion. For each sample and experimental replicate, quality-controlled reads were sequentially trimmed into 75nt and 18nt fixed-length fragment using cutadapt to generate two datasets: 1) R1-R2 75nt paired fragments for gene conversion analysis, and 2) 18nt paired fragments within the same read (~ 60nt spaced) for polymorphism analysis. Quality filtering was performed using fastp (Q30 threshold) and datasets were aligned to S. cerevisiae reference genome twice: (i) using Bowtie (-m 1 -v 0), and (ii) using Bowtie2 57 under restrictive scoring conditions to emulate -m 1 -v 0 behavior (--no-1mm-upfront --score-min C,0,0 -N 0 --end-to-end --fr). Multi-mapping reads from Bowtie were identified and discarded from the inter-chromosomal discordant read pairs of Bowtie2 dataset. Reads spanning the MAT a/ MAT a´ loci were extracted based on coordinate filtering. SAM files were converted to sorted BAM format with SAMtools, and MAT a/ MAT a’ coverage profiles were computed using SAMtools and bamCoverage. BedGraph files were sorted using Bedtools, and per-base coverage values were extracted. The resulting chromosome-resolved, base-level coverage tables for each replicate were merged, positions were aligned relative to the HO cut site, and statistics were computed. Densitometry plots were performed using ggdensity R package (v1.0.0) 58 . Genome-wide discordance assessment. Global genome recombination analysis was performed using Di-GRAPH. 75nt inter-chromosomal discordant read pairs from the dataset described in the previous section were used. Per-base discordance coverage was intersected with each annotating genomic feature to define inter-chromosomal discordant pairs. To ensure discordance reliability, a BLAST alignment was performed to guarantee no concordant re-alignment (see figure 7A for details). Artifactual inter-chromosomal discordant pairs arising from variations between the reference genome and the PMV genetic background were removed. Disc-normalization was applied to calculate discordant frequencies at each genomic region. The number of the discordant reads between the ADE3 gene and the Vector-SSB1-HO-pGAL1-10 sequence was calculated for each timepoint and replicate. A normalization factor was computed using the discordant values obtained for the reference strain and each mutant tested. Frequencies were normalized using these correction factors allowing comparison between timepoints, replicates and strains. Radars plots were represented using fmsb R package (v0.7.6) 59 . Inter-chromosomal discordant network . To generate inter-chromosomal discordant networks, we set genomic features with high discordance frequency (³6%) as nodes and plotted all genomic features they showed a rearrangement with. Both loci were connected to generate the distance network using igraph R package (v2.1.4) 60 . To quantify the interactions observed in the network, we calculated the percentage of each genomic category combination for each timepoint (NSS, SS and UD). Coverage and discordant read data correlation. To evaluate the nature of recombination events, we compared coverage levels and discordancy frequency for each genomic feature with high discordance frequency (³6%). Genomic features showing both high coverage levels and discordancy frequency involve genomic amplification arising from recombination. Homology-directed recombination. To assess whether the global rearrangements observed following an HO induction were dependent on homology-directed recombination, we calculated the sequence identity of every genomic feature exhibiting high discordancy frequency (³6%) using EMBOSS (v6.6.0) 61 (needleall -gapopen 10 -gapextend 0.5). For each genomic feature, identity percentages were computed for two different genomic datasets: (i) all genomic features annotated for PMV strain, to calculate random identity; and (ii) genomic features from inter-chromosomal discordant pairs, to calculate discordant pairs identity. Statistics and reproducibility All data presented in this manuscript was generated from three independent biological replicates. For each analysis, mean values and standard deviations were calculated using the Di-GRAPH computational pipeline, which was applied consistently across all datasets to ensure statistical robustness and reproducibility. Data Availability Genomic datasets and are deposited in the Sequence Read Archive (SRA) database under accession code PRJNA1380437 (https://dataview.ncbi.nlm.nih.gov/object/PRJNA1380437?reviewer=dis753khgl1a9urvk7v1qjlrn6). Code Availability The Di-GRAPH code is available on GitHub, together with instructions for installation and usage, at https://github.com/acb-lab/Di_GRAPH. To ensure reproducibility, a test dataset has been publicly released to Zenodo (https://doi.org/10.5281/zenodo.18470248) and instructions for Di-GRAPH analysis are available at https://github.com/acb-lab/Di_GRAPH. Declarations Acknowledgments We thank J. Haber for kindly providing the background strains JKM139, from which the PMV strain was derived. We thank members of our laboratory for helpful discussions and comments on the manuscript. This work was supported by the projects PID2021-125290NB-I00 and PID2024-160235NB-I00, funded by the MCIN/AEI/10.13039/501100011033/ and by the “FEDER, Una manera de hacer Europa”, awarded to A. C-B. This work was also supported by the institutional grant “Deep-MaX-2024 (CSIC)”. L.I. and A.C. were recipients of a predoctoral fellowship from the “Junta de Castilla y León”. L.I. was the recipient of a “Jae-INTRO” grant from the CSIC. C.D. and L.P. were recipients of a predoctoral fellowship from the “Universidad de Salamanca”. C.D. was recipient of a predoctoral fellowship form the FPU “Formación de Profesorado Universitario” from the “Ministerio de Ciencia, Innovación y Universidades”. N.M.Q. was appointed under the “Momentum” Intramural CSIC program, through the project MMT24-IBFG-01 awarded to A.C-B. Author Contributions L.I. performed the experiments and carried out the processing and analysis of the sequencing data. A.C., C.D., L.P., and M.T.V. collaborated in the experimental work. N.M.Q. provided supervision for the design, development and optimization of the bioinformatic workflow and software development. A.C-B. conceived and designed the study, supervised experimental activities, contributed to data analysis and interpretation, and led the writing of the manuscript. All authors discussed the results and approved the final version of the manuscript. Declaration of interests The authors declare no competing interests. References Mimitou, E. P. & Symington, L. S. DNA end resection: many nucleases make light work. DNA repair 8 , 983-995, doi:10.1016/j.dnarep.2009.04.017 (2009). Gobbini, E., Cassani, C., Villa, M., Bonetti, D. & Longhese, M. P. Functions and regulation of the MRX complex at DNA double-strand breaks. Microb Cell 3 , 329-337, doi:10.15698/mic2016.08.517 (2016). Garcia, V., Phelps, S. E., Gray, S. & Neale, M. J. Bidirectional resection of DNA double-strand breaks by Mre11 and Exo1. Nature 479 , 241-244, doi:10.1038/nature10515 (2011). Mimitou, E. P. & Symington, L. S. Sae2, Exo1 and Sgs1 collaborate in DNA double-strand break processing. Nature 455 , 770-774, doi:10.1038/nature07312 (2008). Zhu, Z., Chung, W. H., Shim, E. Y., Lee, S. E. & Ira, G. Sgs1 helicase and two nucleases Dna2 and Exo1 resect DNA double-strand break ends. Cell 134 , 981-994, doi:10.1016/j.cell.2008.08.037 (2008). Dumont, A. et al. Mechanism of homology search expansion during recombinational DNA break repair in Saccharomyces cerevisiae. Mol Cell 84 , 3237-3253 e3236, doi:10.1016/j.molcel.2024.08.003 (2024). Chen, H., Lisby, M. & Symington, L. S. RPA coordinates DNA end resection and prevents formation of DNA hairpins. Mol Cell 50 , 589-600, doi:10.1016/j.molcel.2013.04.032 (2013). Wang, X. & Haber, J. E. Role of Saccharomyces single-stranded DNA-binding protein RPA in the strand invasion step of double-strand break repair. PLoS biology 2 , E21, doi:10.1371/journal.pbio.0020021 (2004). Krejci, L. et al. DNA helicase Srs2 disrupts the Rad51 presynaptic filament. Nature 423 , 305-309, doi:10.1038/nature01577 (2003). Veaute, X. et al. The Srs2 helicase prevents recombination by disrupting Rad51 nucleoprotein filaments. Nature 423 , 309-312, doi:10.1038/nature01585 (2003). Piazza, A. et al. Dynamic Processing of Displacement Loops during Recombinational DNA Repair. Mol Cell 73 , 1255-1266 e1254, doi:10.1016/j.molcel.2019.01.005 (2019). Piazza, A. & Heyer, W. D. Homologous Recombination and the Formation of Complex Genomic Rearrangements. Trends in cell biology 29 , 135-149, doi:10.1016/j.tcb.2018.10.006 (2019). Bantele, S. et al. Repair of DNA double-strand breaks leaves heritable impairment to genome function. Science 390 , eadk6662, doi:10.1126/science.adk6662 (2025). Sasaki, M., Tischfield, S. E., van Overbeek, M. & Keeney, S. Meiotic recombination initiation in and around retrotransposable elements in Saccharomyces cerevisiae. PLoS genetics 9 , e1003732, doi:10.1371/journal.pgen.1003732 (2013). Gilat, N. et al. From single-molecule to genome-wide mapping of DNA lesions: repair-assisted damage detection sequencing. Biophys Rep (N Y) 1 , None, doi:10.1016/j.bpr.2021.100017 (2021). Hussmann, J. A. et al. Mapping the genetic landscape of DNA double-strand break repair. Cell 184 , 5653-5669 e5625, doi:10.1016/j.cell.2021.10.002 (2021). de Alba, E. L. et al. A comprehensive genetic catalog of human double-strand break repair. Science 390 , eadr5048, doi:10.1126/science.adr5048 (2025). Sriramachandran, A. M. et al. Genome-wide Nucleotide-Resolution Mapping of DNA Replication Patterns, Single-Strand Breaks, and Lesions by GLOE-Seq. Mol Cell 78 , 975-985 e977, doi:10.1016/j.molcel.2020.03.027 (2020). Campos, A. et al. Cdc14 phosphatase counteracts Cdk-dependent Dna2 phosphorylation to inhibit resection during recombinational DNA repair. Nat Commun 14 , 2738, doi:10.1038/s41467-023-38417-5 (2023). Pfeifer, G. P. & Jin, S. G. Methods and applications of genome-wide profiling of DNA damage and rare mutations. Nat Rev Genet 25 , 846-863, doi:10.1038/s41576-024-00748-4 (2024). Teng, Y. et al. A novel method for the genome-wide high resolution analysis of DNA damage. Nucleic Acids Res 39 , e10, doi:10.1093/nar/gkq1036 (2011). Zhang, K., Zheng, D. Q., Sui, Y., Qi, L. & Petes, T. D. Genome-wide analysis of genomic alterations induced by oxidative DNA damage in yeast. Nucleic Acids Res 47 , 3521-3535, doi:10.1093/nar/gkz027 (2019). Rybin, M. J. et al. Emerging Technologies for Genome-Wide Profiling of DNA Breakage. Front Genet 11 , 610386, doi:10.3389/fgene.2020.610386 (2020). Ramos, F. et al. Genome-wide sequencing analysis of Sgs1, Exo1, Rad51, and Srs2 in DNA repair by homologous recombination. Cell Rep 38 , 110201, doi:10.1016/j.celrep.2021.110201 (2022). Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome biology 10 , R25, doi:10.1186/gb-2009-10-3-r25 (2009). Lo, Y. C. et al. Sgs1 regulates gene conversion tract lengths and crossovers independently of its helicase activity. Mol Cell Biol 26 , 4086-4094, doi:10.1128/MCB.00136-06 (2006). Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25 , 1754-1760, doi:10.1093/bioinformatics/btp324 (2009). Bradshaw, V. A. & McEntee, K. DNA damage activates transcription and transposition of yeast Ty retrotransposons. Mol Gen Genet 218 , 465-474, doi:10.1007/BF00332411 (1989). Farkash, E. A., Kao, G. D., Horman, S. R. & Prak, E. T. Gamma radiation increases endonuclease-dependent L1 retrotransposition in a cultured cell assay. Nucleic Acids Res 34 , 1196-1204, doi:10.1093/nar/gkj522 (2006). Staleva Staleva, L. & Venkov, P. Activation of Ty transposition by mutagens. Mutation research 474 , 93-103, doi:10.1016/s0027-5107(00)00165-2 (2001). Mitchel, K., Lehner, K. & Jinks-Robertson, S. Heteroduplex DNA position defines the roles of the Sgs1, Srs2, and Mph1 helicases in promoting distinct recombination outcomes. PLoS genetics 9 , e1003340, doi:10.1371/journal.pgen.1003340 (2013). Pohl, T. J. & Nickoloff, J. A. Rad51-independent interchromosomal double-strand break repair by gene conversion requires Rad52 but not Rad55, Rad57, or Dmc1. Mol Cell Biol 28 , 897-906, doi:10.1128/MCB.00524-07 (2008). So, A. et al. RAD51 protects against nonconservative DNA double-strand break repair through a nonenzymatic function. Nucleic Acids Res 50 , 2651-2666, doi:10.1093/nar/gkac073 (2022). Argueso, J. L. et al. Double-strand breaks associated with repetitive DNA can reshape the genome. Proc Natl Acad Sci U S A 105 , 11845-11850, doi:10.1073/pnas.0804529105 (2008). Putnam, C. D., Hayes, T. K. & Kolodner, R. D. Specific pathways prevent duplication-mediated genome rearrangements. Nature 460 , 984-989, doi:10.1038/nature08217 (2009). Gal, Z. et al. Hyper-recombination in ribosomal DNA is driven by long-range resection-independent RAD51 accumulation. Nat Commun 15 , 7797, doi:10.1038/s41467-024-52189-6 (2024). Qiu, Y. et al. Srs2 prevents Rad51 filament formation by repetitive motion on DNA. Nat Commun 4 , 2281, doi:10.1038/ncomms3281 (2013). Liu, B. et al. Structural variation discovery in the cancer genome using next generation sequencing: computational solutions and perspectives. Oncotarget 6 , 5477-5489, doi:10.18632/oncotarget.3491 (2015). Janke, C. et al. A versatile toolbox for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes. Yeast 21 , 947-962, doi:10.1002/yea.1142 (2004). Goldstein, A. L. & McCusker, J. H. Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15 , 1541-1553, doi:10.1002/(SICI)1097-0061(199910)15:143.0.CO;2-K (1999). Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26 , 841-842, doi:10.1093/bioinformatics/btq033 (2010). Freese, N. H., Norris, D. C. & Loraine, A. E. Integrated genome browser: visual analytics platform for genomics. Bioinformatics 32 , 2089-2095, doi:10.1093/bioinformatics/btw069 (2016). Wickham, H., François, R., Henry, L., Müller, K. & Vaughan, D. dplyr: a grammar of data manipulation. (2023). Wickham, H., Vaughan, D. & Girlich, M. . Tidyr: tidy messy data. . (2024). Wickham, H. Stringr: simple, consistent wrappers for common string operations. . (2025). Wickham, H. ggplot2: Elegant Graphics for Data Analysis. . Journal of the Royal Statistical Society Series A: Statistics in Society 174(1), 245–246 . Wickham, H., & Henry, L. . purrr: Functional Programming Tools. (2025). Xu, S. et al. Use ggbreak to Effectively Utilize Plotting Space to Deal With Large Datasets and Outliers. Front Genet 12 , 774846, doi:10.3389/fgene.2021.774846 (2021). Allaire, J., Xie, Y., Dervieux, C., McPherson, J., Luraschi, J., Ushey, K., Atkins, A., Wickham, H., Cheng, J., Chang, W., & Iannone, R. . rmarkdown: Dynamic Documents for R. (2024). Garrick, A.-B., Sievert, C., Iannone, R., & Allaire, J. . flexdashboard: R Markdown Format for Flexible Dashboards. (2025). Martin, M. Cutadapt Removes Adapter Sequences From High‐Throughput Sequencing Reads. EMBnet.journal 17 , doi:https://doi.org/10.14806/ej.17.1.200 (2011). Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34 , i884-i890, doi:10.1093/bioinformatics/bty560 (2018). Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25 , 2078-2079, doi:10.1093/bioinformatics/btp352 (2009). Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res 44 , W160-165, doi:10.1093/nar/gkw257 (2016). Robinson, J. T. et al. Integrative genomics viewer. Nature biotechnology 29 , 24-26, doi:10.1038/nbt.1754 (2011). Wagih, O. ggseqlogo: a versatile R package for drawing sequence logos. Bioinformatics 33 , 3645-3647, doi:10.1093/bioinformatics/btx469 (2017). Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nature methods 9 , 357-359, doi:10.1038/nmeth.1923 (2012). Otto, J. K., D. ggdensity: Interpretable Bivariate Density Visualization with ggplot2. https://jamesotto852.github.io/ggdensity/ (2023). Nakazawa, M. fmsb: Functions for Medical Statistics Book with some Demographic Data. (2010). Csárdi, G., & Nepusz, T. The igraph software package for complex network research. (2006). Rice, P., Longden, I. & Bleasby, A. EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet 16 , 276-277, doi:10.1016/s0168-9525(00)02024-2 (2000). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryFigureLegends.docx SupplementaryFigure7.jpg Supplementary Figure 7 SupplementaryFigure3.jpg Supplementary Figure 3 SupplementaryFigure4.jpg Supplementary Figure 4 RGPMVv9CHRIII.fasta.zip Related Manuscript File SupplementaryFigure1.jpg Supplementary Figure 1 RGPMVv9.fasta.zip Related Manuscript File SupplementaryFigure5.jpg Supplementary Figure 5 SupplementaryFigure8.jpg Supplementary Figure 8 SupplementaryFigure6.jpg Supplementary Figure 6 SupplementaryFigure2.jpg Supplementary Figure 2 SupplementaryTables.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8989286","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":611327482,"identity":"b13387b7-8974-452e-a2d3-9f25f2a44897","order_by":0,"name":"Andres Clemente-Blanco","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA80lEQVRIiWNgGAWjYHACAxAhA+XYADFzA1FaeMDMAwxpQJKRNC2HCWvhn9287cOHP3Y8/GKHjz3+UHE+cX77wQaGD39wa5G4c6x45sy2ZB7J2WnpBgfO3E7ccCaxgXFmGx5rbuQYM/M2HOAxuJ1jJnGwDaiFIbEBKIJbhzxIy58/B3jsb+d/kzj471zi/P6HDUAR3FoMQFoY2IC2SOewSRxsOJDYcANoCwMbbi2GN9KKGXuBfpG4nWYmceZYsvGGGw8bDvbi8YvcjeTNDD/+2Mnxz05+JlFRYyc7vz/54IMfeByGHRwgVcMoGAWjYBSMAlQAAEDJWSj5Yd1uAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-3674-0671","institution":"Instituto de Biología Funcional y Genómica","correspondingAuthor":true,"prefix":"","firstName":"Andres","middleName":"","lastName":"Clemente-Blanco","suffix":""},{"id":611327483,"identity":"ba97a503-c843-4199-8fa7-8bbe75e4a1c2","order_by":1,"name":"Lydia Iglesias","email":"","orcid":"https://orcid.org/0000-0002-6537-9002","institution":"Spanish National Research Council","correspondingAuthor":false,"prefix":"","firstName":"Lydia","middleName":"","lastName":"Iglesias","suffix":""},{"id":611327484,"identity":"daac78aa-2a3d-4a6a-b25c-25e1ac500bb5","order_by":2,"name":"Adrián Campos","email":"","orcid":"https://orcid.org/0000-0002-5451-8150","institution":"Spanish National Research Council","correspondingAuthor":false,"prefix":"","firstName":"Adrián","middleName":"","lastName":"Campos","suffix":""},{"id":611327485,"identity":"72a881b5-90b8-49d2-b958-aa3250fc9bfe","order_by":3,"name":"Celia Delgado","email":"","orcid":"https://orcid.org/0000-0002-3717-1378","institution":"Spanish National Research Council","correspondingAuthor":false,"prefix":"","firstName":"Celia","middleName":"","lastName":"Delgado","suffix":""},{"id":611327486,"identity":"c8e53037-26b1-4363-ba9c-a75ce3117228","order_by":4,"name":"Lydia Pulido","email":"","orcid":"","institution":"Spanish National Research Council","correspondingAuthor":false,"prefix":"","firstName":"Lydia","middleName":"","lastName":"Pulido","suffix":""},{"id":611327487,"identity":"710ead1d-a777-4896-b930-8a5fafdb74ee","order_by":5,"name":"Maria Teresa Villoria","email":"","orcid":"","institution":"Instituto de Biología Funcional y Genómica","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"Teresa","lastName":"Villoria","suffix":""},{"id":611327488,"identity":"8ee1797d-2c67-4d46-ad13-b22973ea4248","order_by":6,"name":"Narciso Quijada","email":"","orcid":"https://orcid.org/0000-0002-1956-7205","institution":"Spanish National Research Council - CSIC","correspondingAuthor":false,"prefix":"","firstName":"Narciso","middleName":"","lastName":"Quijada","suffix":""}],"badges":[],"createdAt":"2026-02-27 14:42:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8989286/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8989286/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105358364,"identity":"aaf4fb26-c9fb-419b-bb9c-b04a09183a67","added_by":"auto","created_at":"2026-03-25 07:18:10","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2626562,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSelected survivors of cells deficient in stage-specific DNA repair factors exhibited distinct gene conversion outcomes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Schematic representation of the endogenous \u003cem\u003eMAT\u003c/em\u003ea\u003cem\u003e locus\u003c/em\u003e in chromosome III, lacking the silent mating-type \u003cem\u003eloci\u003c/em\u003e \u003cem\u003eHMR\u003c/em\u003ea\u003cem\u003e \u003c/em\u003eand \u003cem\u003eHML\u003c/em\u003ea. A 1.3 kb \u003cem\u003eMAT\u003c/em\u003ea´ DNA fragment containing engineered polymorphisms approximately every 60 bps (purple lines) was inserted at the \u003cem\u003eARG5,6\u003c/em\u003e \u003cem\u003elocus\u003c/em\u003e on chromosome V to distinguish donor and recipient sequencing reads. Blue lines at the \u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003elocus\u003c/em\u003edepict polymorphisms that generate \u003cem\u003eEco\u003c/em\u003eRI sites, which are used to assess dynamics and symmetry of gene conversion in Southern blot experiments.\u003c/p\u003e\n\u003cp\u003eB) Diagram illustrating the experimental procedure used to assess genomic changes in response to a DSB in non-selected survivors (NSS) and selected survivors (SS), as well as in undamaged cells (UD) relative to the uninduced culture (T0). The time points for sample collection for each strain are indicated.\u003c/p\u003e\n\u003cp\u003eC) T0-normalized sequencing coverage profile across the -800 to +800 bp region flanking the HO/HO-\u003cem\u003einc\u003c/em\u003e sites (chromosomes III and V, respectively) for undamaged (UD), selected survivors (SS) and non-selected survivors (NSS) for wild-type, \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells.\u003c/p\u003e\n\u003cp\u003eD) Gene conversion was assessed in NSS and SS cells by analysing coverage profiles generated from 18-mer reads aligned using Bowtie in -k 1 mode (see Supplementary Fig. 2B). The size of the \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ peaks (distance from baseline to maximum coverage) was measured to quantify subtle changes in GC, reflecting differences in conversion frequency, extension and directionality. Concomitant loss of sequencing coverage on chromosome III and enrichment on chromosome V were plotted to illustrate the net transfer of genetic information between \u003cem\u003eMAT loci\u003c/em\u003e in each mutant background and cell stage.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/79f9063bbe19c448b539a5ac.jpg"},{"id":105564816,"identity":"9eb7363e-5dbf-4e14-b57f-9757ddbc9539","added_by":"auto","created_at":"2026-03-27 12:50:56","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1303782,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDiscordant reads at the \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMAT\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003ea/\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMAT\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003ea´ \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eloci\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e provide a robust method for detecting subtle variations in the frequency, directionality, and extent of individual GC events.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Schematic representation of the approach used to assess GC boundaries at the \u003cem\u003eMAT\u003c/em\u003e \u003cem\u003elocus\u003c/em\u003ethrough analysis of 75-mer discordant reads.\u003c/p\u003e\n\u003cp\u003eB) Densitometry map generated by plotting the \u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea´ coordinates of each 75-mer discordant read pair of SS cells, enabling precise inference of symmetry and extent of individual GC events. Percentage reflects the proportion of recombination events for each defined area.\u003c/p\u003e\n\u003cp\u003eC) Quantification of \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ discordant events depicted in B) provides a comparative assessment of the frequency, distribution, and characteristics of gene conversion events across the \u003cem\u003eMAT\u003c/em\u003e \u003cem\u003elocus\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eD) Schematic representation of the discordances expected for continuous and discontinuous GC events at each side of the HO break through analysis of 18-mer discordant reads.\u003c/p\u003e\n\u003cp\u003eE) Distribution of discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ and \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea 18-mer reads flanking the HO cleavage site. Discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ and \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea reads in the bottom-left and top-right quadrants correspond to the left and right boundaries of continuous GC events, respectively. In contrast, discordant \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ reads in the bottom-left and top-right quadrant, respectively, indicate left and right discontinuous GC events.\u003c/p\u003e\n\u003cp\u003eF) Quantification of continuous and discontinuous left and right GC events. Repair occurs predominantly via continuous GC to the left of the HO site, whereas only a minor fraction of events proceeds via continuous GC to the right or via discontinuous GC on either side. Discontinuities at the HO site are also asymmetric, occurring more frequently to the left than to the right.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/85d8e220b0a0ba65ac809318.jpg"},{"id":105358375,"identity":"906b59c3-368d-44b2-bb18-9f6c263b9272","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1807580,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStage-specific DNA repair mutants exhibit distinctive discordant \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMAT\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003ea/\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eMAT\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003ea´ gene conversion patterns.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA-D) Analysis of discordant read profiles in \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D \u003c/em\u003eand \u003cem\u003erad51D\u003c/em\u003e cells. The left, middle, and right panels show, respectively, densitometry of discordant reads, distribution of GC events along the \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003eloci\u003c/em\u003e, and the continuity of GC events.\u003c/p\u003e\n\u003cp\u003eE) Quantification of GC directionality and continuity in \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D \u003c/em\u003eand \u003cem\u003erad51D\u003c/em\u003e cells, highlighting differences in the frequency, symmetry, and presence of discontinuous events across mutants.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/a0ad62e95853580d5d6d8135.jpg"},{"id":105564680,"identity":"9b13aa8e-889c-43ca-9982-429b9a001e61","added_by":"auto","created_at":"2026-03-27 12:50:30","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2139306,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistinct DNA repair mutants display characteristic HO-associated mutagenic signatures and DNA repair pathway choices.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Schematic overview of the approach used to link DNA repair pathway choice to mutagenic outcomes. 150-mer reads mapping at the HO site were filtered based on polymorphisms at positions -64 (A/G\u003csup\u003e200689\u003c/sup\u003e, left of HO/HO-\u003cem\u003einc\u003c/em\u003e) and 0 (C/T\u003csup\u003e200753\u003c/sup\u003e, HO/HO-\u003cem\u003einc\u003c/em\u003e). Reads containing T\u003csup\u003e200753\u003c/sup\u003e/G\u003csup\u003e200689\u003c/sup\u003e were assigned to long-range HR, T\u003csup\u003e200753\u003c/sup\u003e/A\u003csup\u003e200689\u003c/sup\u003e to short-range HR, and C\u003csup\u003e200753\u003c/sup\u003e/A\u003csup\u003e200869\u003c/sup\u003e to NHEJ. All subsets were re-aligned to chromosome III, and mismatches were quantified at position -11 to +9 relative to the HO site.\u003c/p\u003e\n\u003cp\u003eB) Long-range and short-range HR-associated mutagenesis in SS cells. No increase in HO-induced mutations was observed in any mutant, indicating that defects in HR factors do not elevate the mutation rate. Note that the observed C to T substitution at position 0 reflects the intrinsic HR repair process rather than a true mutation. BS: base substitution; DEL: deletion; INS: insertion.\u003c/p\u003e\n\u003cp\u003eC) NHEJ-associated mutagenesis profiles in \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D \u003c/em\u003eand \u003cem\u003erad51D\u003c/em\u003e SS cells. In \u003cem\u003esgs1D\u003c/em\u003e cells, deletions accumulated between -1 and -3 nt, with insertions at -2nt. \u003cem\u003esrs2D \u003c/em\u003ecells showed insertions at -3 nt. \u003cem\u003erad51D\u003c/em\u003e mutants exhibited extensive mutagenesis, including base substitutions, insertions from -1 to -3 nt, and widespread deletions across the analysed segment, particularly between 0 and -3 nt. BS: base substitution; DEL: deletion; INS: insertion.\u003c/p\u003e\n\u003cp\u003eD) Sequence logo profiles of \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D \u003c/em\u003eand \u003cem\u003erad51D\u003c/em\u003e SS cells across the -11 to +9 region flanking the HO site. Reduced letter size near the HO recognition site indicates deletion events, which are more pronounced in \u003cem\u003esgs1D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells. Insertions observed in \u003cem\u003esrs2D\u003c/em\u003e have minimal impact on the overall mutagenic signature shown. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eE) Quantification of base substitutions, deletions and insertions in wild-type, \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D \u003c/em\u003eand \u003cem\u003erad51D\u003c/em\u003e SS cells.\u003c/p\u003e\n\u003cp\u003eF) DNA repair pathway usage in stage-specific DNA repair mutants. Wild-type, \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, and \u003cem\u003esrs2D \u003c/em\u003eSS cells preferentially repair the HO break via long-range HR, followed by short-range HR. In contrast, \u003cem\u003erad51D\u003c/em\u003e SS cells displayed increased reliance on NHEJ with reduced long-range HR, while short-range HR remained largely unaffected.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/1495cd9947c1b598be941aa8.jpg"},{"id":105358368,"identity":"26dafbb2-e402-4173-aaa8-674d2ae78665","added_by":"auto","created_at":"2026-03-25 07:18:10","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2561879,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenome-wide sequencing analysis reveals altered coverage of distinctive genomic features in response to an HO-induced DSB.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Genome-wide coverage analysis of genomic elements grouped into ORF, intergenic region, LTR, TEG, Ty, tRNA, rRNA, ncRNA, snRNA, snoRNA, ARS, centromere and subtelomeric region. For each genomic element, coverage value for each experiment was normalized to its respective T0 sample. Top panel, NSS wild-type cells following HO induction; middle panel, SS cells; bottom panel, UD raffinose-grown cells. Values above 1 indicate coverage gains, whereas values below 1 indicate losses relative to the uncut state. Note the pronounced enrichment of LTRs, TEGs and Ty elements in NSS cells, which is largely counter-selected in SS cells and absent in UD cells.\u003c/p\u003e\n\u003cp\u003eB) Violin plots representing the distribution and quantification of coverage-based enrichment across the genomic categories. Mean coverage values of all individual genomic features shown in A) were grouped according to their corresponding genomic category, and the distribution of these values was visualized as violin plots. For each category, the mean coverage ratio (normalized to T0) is calculated, allowing comparison of coverage variations across categories in NSS (top panel), SS (middle panel) and UD (bottom panel) samples.\u003c/p\u003e\n\u003cp\u003eC) Genomic elements showing significant coverage changes (\u0026lt; 0.8 in blue, and \u0026gt;1.2 in red) within each genomic category. For NSS, SS and UD samples, enrichment metrics were calculated as the percentage of genomic elements within each category displaying coverage ratios above 1.2 (red) or below 0.8 (blue), relative to the corresponding T0 sample. Variability associated with these metrics is shown. NSS cells exhibit a strong overrepresentation of coverage gains in LTRs, TEGs, Ty elements, tRNAs and subtelomeric regions, together with centromeric depletion. These alterations are largely suppressed in SS cells, although a residual enrichment of mobile elements persists, while UD cells display only minor variations, except for subtelomeric regions.\u003c/p\u003e\n\u003cp\u003eD) Chromosome-specific distribution of coverage alterations following HO induction. Genomic features were sorted by chromosome (I-XVI), and the percentage of elements per chromosome exceeding coverage thresholds (\u0026lt;0.8 or \u0026gt;1.2) was calculated. NSS cells display a global tendency toward coverage gains across most chromosomes, with marked chromosome-to-chromosome variability, indicating non-uniform genome reorganization in response to DNA damage (top panel). These alterations are strongly reduced in SS cells (middle panel) and largely absent in UD cells (bottom panel). \u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003eloci\u003c/em\u003e were excluded from the analysis to avoid homologous recombination-derived chromosome III/V artifacts.\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/8842b9a3ee4a379b594f46da.jpg"},{"id":105358378,"identity":"34931f3e-1ff9-4dac-b88d-a1e4c2b8b1c6","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":2647234,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eexo1D\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e, \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003esgs1D\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e, \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003esrs2D\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e and \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003erad51D\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e mutants display distinct genome-wide coverage alterations in response to DNA damage.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Genomic elements exhibiting significant coverage changes (\u0026lt; 0.8, blue; \u0026gt;1.2, red) across genomic categories in \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells under NSS (top panels), SS (middle panels) and UD (bottom panels) conditions.\u003c/p\u003e\n\u003cp\u003eB) Chromosome-specific distribution of coverage gains (\u0026gt;1.2, red) and losses (\u0026lt;0.8, blue) in \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells under NSS (top panels), SS (middle panels) and UD (bottom panels) conditions.\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/d9c151ef044458e91faf3fd9.jpg"},{"id":105564542,"identity":"79e747c7-0a04-487d-9eae-ffb2c30615aa","added_by":"auto","created_at":"2026-03-27 12:49:56","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2987806,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenome-wide discordant read analysis reveals global genome reorganizations following a DNA break.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Schematic representation of the bioinformatic filtering strategy used to distinguish true inter-chromosomal discordant read pairs from artifactual discordances. Sequencing or amplification errors (asterisk) may generate apparent discordant pairs due to misalignment of one mate. To ensure reliability, discordant read pairs were required to remain discordant after relaxed BLAST re-alignment. Read pairs were classified according to their ability to re-align concordantly, including pairs with no concordant alignment (NA), those becoming concordant after allowing 1, 2 or ³3 nucleotide variants and unfiltered reads (UF), from maximum to lower discordance confidence score, respectively. Discordant pairs detected at the T0 timepoint were removed to eliminate artifacts arising from differences between the reference genome and the PMV genetic background.\u003c/p\u003e\n\u003cp\u003eB) Genome-wide distribution of reliable inter-chromosomal discordances (NA group) in NSS and SS cells (left-bottom and right-top triangle, respectively) following an HO-induced DSB. Dot plots display pairwise association between genomic features involved in each rearrangement event, with signal intensity reflecting event frequency. Disc-normalization was applied to calculate discordant frequencies. The bottom-right graph provides an enlarged view highlighting the increased frequency of discordant events at mobile elements and subtelomeric regions in NSS cells. The top right graph shows the net number of NA discordance events per genomic category, revealing an enrichment of rDNA-associated discordances in SS cells.\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/6d4ee2c911e81c89b040cb24.jpg"},{"id":105358380,"identity":"537efbe5-eb72-4ad7-9fa6-2dfb3b283bee","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":2753126,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistinct DNA repair factors shape genome-wide reorganization patterns following a DSB.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) Genome-wide distribution of reliable NA inter-chromosomal discordances across genomic categories in \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells under NSS, SS and UD conditions. Disc-normalization was applied to calculate discordant reads distribution.\u003c/p\u003e\n\u003cp\u003eB) Genomic maps showing the frequency of NA discordant events for all genomic features across three biological replicates in NSS cells. Disc-normalization was applied to calculate discordant frequencies. The intensity of the colors represents the frequency of discordances.\u003c/p\u003e\n\u003cp\u003eC) Network representation of highly discordant genomic \u003cem\u003eloci\u003c/em\u003e (³6% discordance frequency) in NSS conditions. Nodes are colored according to the strains, with shaded backgrounds indicating their corresponding genomic category, and the edges representing recombination events, revealing strain-specific patterns of genome reorganization.\u003c/p\u003e\n\u003cp\u003eD) Quantification of the recombination frequencies between genomic categories of wild-type, \u003cem\u003eexo1D\u003c/em\u003e, \u003cem\u003esgs1D\u003c/em\u003e, \u003cem\u003esrs2D\u003c/em\u003e and \u003cem\u003erad51D\u003c/em\u003e cells from the interaction networks (NSS, SS and UD conditions). The graphs reveal strain and condition-specific patterns of genome reorganization, including transposons, rDNA rearrangements and ORF or intergenic-associated recombination events, providing a comprehensive view of the structural dynamics of the genome in response to DNA damage.\u003c/p\u003e\n\u003cp\u003eE) Comparison of coverage values and discordance ratios of NSS, SS and UD cells across genomic features. Coverage and discordance are generally correlated, indicating that translocations often involve genomic amplification. Transposons show the strongest correlation, consistent with their amplification enhancing mobility. Features with high coverage but low discordance (mainly Ty elements) likely undergo amplification without detectable integration, whereas features with low coverage but high discordance (largely rDNA NTS1-2 region) indicate translocation rather than expansion or contraction of the rDNA array.\u003c/p\u003e\n\u003cp\u003eF) Assessment of the homology requirements for HO-induced recombination events. The sequence identity of each actual discordant pair (red line) was compared to the random pairwise identity expected across the genome (blue line). The ratio between observed and random identity reflects the degree of homology dependence (HDR). ORFs, despite low random identity, exhibit high HDR ratios, indicating strong sequence selectivity during recombination. In contrast, LTRs, intergenic regions, and rDNA, which have higher random identity, display lower homology dependence. Retrotransposons show intermediate and variable HDR, consistent with frequent self-targeted insertions observed in the interaction networks shown in C).\u003c/p\u003e","description":"","filename":"Figure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/20d6df95c111d4d539a766d1.jpg"},{"id":108804086,"identity":"b8c44ed5-12a8-4971-b76f-69f8dcd5afba","added_by":"auto","created_at":"2026-05-08 15:15:31","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":19559684,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/6120da18-0797-436f-8b26-ab3d69e17155.pdf"},{"id":105565491,"identity":"16555432-6ee4-4d27-9114-5f8ae3ea3c83","added_by":"auto","created_at":"2026-03-27 12:53:24","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":19469,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementaryFigureLegends.docx","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/0ea27f654a3b0b08355db186.docx"},{"id":105358371,"identity":"8a7e4dfc-7443-4af1-b2b6-37d6c00147af","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":2626240,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 7\u003c/p\u003e","description":"","filename":"SupplementaryFigure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/ebae384053d04925ac0c2b6e.jpg"},{"id":105358366,"identity":"abe13a28-8658-4016-83d3-5d7f67ebd3e8","added_by":"auto","created_at":"2026-03-25 07:18:10","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1564880,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 3\u003c/p\u003e","description":"","filename":"SupplementaryFigure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/0106fbe3ef0a06a678fe7530.jpg"},{"id":105565120,"identity":"5f391d9c-4f97-4d20-a486-86bb8e1542ac","added_by":"auto","created_at":"2026-03-27 12:52:01","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":3300658,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 4\u003c/p\u003e","description":"","filename":"SupplementaryFigure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/d84c563c8fab8f7584fb6b07.jpg"},{"id":105565785,"identity":"4baa392e-347c-42fe-9745-7a3dad08ff15","added_by":"auto","created_at":"2026-03-27 12:54:23","extension":"zip","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":85935,"visible":true,"origin":"","legend":"\u003cp\u003eRelated Manuscript File\u003c/p\u003e","description":"","filename":"RGPMVv9CHRIII.fasta.zip","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/bbc2aa993d8db90bcb3213cf.zip"},{"id":105358370,"identity":"71f3bf78-2fb1-4cf6-84e8-1d99d5746110","added_by":"auto","created_at":"2026-03-25 07:18:10","extension":"jpg","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":1915390,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 1\u003c/p\u003e","description":"","filename":"SupplementaryFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/5a0d8fc7dd74963f36b3dff7.jpg"},{"id":105358377,"identity":"7e04c5ff-9757-47d8-adf7-633c82731f77","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"zip","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":3210171,"visible":true,"origin":"","legend":"\u003cp\u003eRelated Manuscript File\u003c/p\u003e","description":"","filename":"RGPMVv9.fasta.zip","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/cd1c64285327ec1acae6772d.zip"},{"id":105564973,"identity":"0e58033f-73a8-496d-b2d2-8db3a8ed6d67","added_by":"auto","created_at":"2026-03-27 12:51:29","extension":"jpg","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":2034337,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 5\u003c/p\u003e","description":"","filename":"SupplementaryFigure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/9ae1ee4e293e40837d3d9b60.jpg"},{"id":105358381,"identity":"6286070e-6343-4ab2-adbe-959d4591b090","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":2014538,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 8\u003c/p\u003e","description":"","filename":"SupplementaryFigure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/dd0e9352a3c7648e5f44f97e.jpg"},{"id":105358373,"identity":"d14f98ec-a314-4e2c-8240-2ff6f87d9c70","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":3032778,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 6\u003c/p\u003e","description":"","filename":"SupplementaryFigure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/0683f5ac88da832591b8c81b.jpg"},{"id":105358379,"identity":"fe3bcb26-f726-47cd-9fca-12bc8e495828","added_by":"auto","created_at":"2026-03-25 07:18:11","extension":"jpg","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":2732940,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 2\u003c/p\u003e","description":"","filename":"SupplementaryFigure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/9eabd607ec2af5767ea258c2.jpg"},{"id":105565492,"identity":"4d94e51d-1941-4ed4-9b01-1b4758a9c154","added_by":"auto","created_at":"2026-03-27 12:53:24","extension":"docx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":20684,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTables.docx","url":"https://assets-eu.researchsquare.com/files/rs-8989286/v1/ca4d6e1433fe169b08d01f6f.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Stage-specific Recombination Factors Differentially Regulate Double-strand Break Repair Fidelity and Influence Genome Stability","fulltext":[{"header":"Introduction","content":"\u003cp\u003eFaithful repair of DSBs is essential to maintain genome stability and prevent chromosomal rearrangements that could compromise cell survival. To counteract the detrimental effect of DSBs, cells employ evolutionarily conserved and highly coordinated repair pathways to restore these DNA lesions, preventing the transmission of unwanted mutation to the progeny. Non-homologous end joining (NHEJ) directly ligates broken DNA ends without requiring a homologous template, being often an error-prone mechanism that accumulates insertions, deletions or mutations at the break site. By contrast, HR constitutes one of the most faithful repair pathways, as it uses an intact homologous sequence as a template to recover the lost genetic information. HR is initiated by nucleolytic processing of DSB ends, which generates 3\u0026acute; single strand DNA (ssDNA) through a process known as resection \u003csup\u003e1\u003c/sup\u003e. Resection occurs in two sequential phases. In the initial step, the nuclease activity of Mre11 within the MRX complex (Mre11-Rad50-Xrs2), together with Sae2, introduces an endonucleolytic nick at the boundaries of the DSB, and degrades DNA in the 3\u0026acute;-5\u0026acute; direction toward the break, producing a short 3\u0026acute; ssDNA overhang \u003csup\u003e2,3\u003c/sup\u003e. This intermediate is subsequently recognized by the exonucleases Exo1 and Dna2, that together with the STR complex (Sgs1-Top3-Rmi1) promote extensive 5\u0026acute;-3\u0026acute; long-range resection, ultimately generating long ssDNA tracts \u003csup\u003e4,5\u003c/sup\u003e. Long-range resected tracts are coated with RPA and Rad51, which are necessary for ssDNA protection, homology search, strand displacement, and ultimately for promoting strand invasion, leading to the formation of the displacement loop (D-loop) \u003csup\u003e6-8\u003c/sup\u003e. After the missing genetic information has been copied, the D-loop is dismantled by the activity of helicases such as Sgs1 or Srs2 \u003csup\u003e9-11\u003c/sup\u003e, and the newly synthesized DNA re-anneals to its original chromosomal context. The exposing 3\u0026acute; ends then prime bidirectional DNA synthesis to complete repair. Because this recombination-mediated pathway involves unidirectional transfer of the sequence information from the donor to the recipient, it typically results in GC events.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlthough the core HR machinery has been well characterized over the past years, the specific contributions of HR factors to DNA repair, as well as their broader effects on genome-wide integrity, remains incompletely understood. This gap is particularly relevant because subtle perturbations in HR can generate complex genomic alterations, including gross chromosomal rearrangements, multi-strand invasions and chromothripsis, changes that are difficult to detect using classical genetic approaches \u003csup\u003e12\u003c/sup\u003e. Alongside, recent work has revealed that a single repaired DSB can induce long-lasting heritable alterations in chromatin topology and gene expression across large genomic regions, emphasizing that DNA repair outcomes extend beyond sequence restoration to persistent changes in genome organization and function \u003csup\u003e13\u003c/sup\u003e. Beyond sporadic or experimentally induced damage, DSBs also arise during meiosis within or near repetitive elements such as retrotransposons, increasing the risk of non-allelic recombination and genome rearrangements. High-resolution mapping of endogenous Ty elements in budding yeast has shown that these elements variably modulate local DSB formation, in some cases promoting nearby hotspots, highlighting the importance of the genomic context in recombination outcomes and genome stability \u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eIn this context, genome-wide methods have recently transformed the ability to study DNA repair and genomic instability at unprecedented depth and scale. Techniques such as RADD-seq (Repair-Assisted Detection sequencing) allow mapping of DNA damage and repair sites across the entire genome rather than restricting analysis to a few pre-selected foci \u003csup\u003e15\u003c/sup\u003e. Moreover, high-throughput functional screening strategies such as Repair-seq and REPAIRome, which assay the outcomes of DSBs under hundreds of genetic perturbations, have exposed unexpected pathway dependencies \u003csup\u003e16,17\u003c/sup\u003e. Single-strand breaks (SSBs) have also been the focus in recent years, with the discovery of GLOE-seq providing a powerful tool to capture these DNA lesions and demonstrating its applicability to the mapping of Okazaki fragments and to the analysis of spontaneous distribution of nicks in yeast and human chromatin \u003csup\u003e18\u003c/sup\u003e. Discordant reads have also been used as an approach to detect various types of recombination events, both at naturally occurring and experimentally induced DSBs, providing a powerful tool to study genomic changes and DNA repair \u003csup\u003e14,19\u003c/sup\u003e. These genome-wide approaches provide a powerful, unbiased framework to detect, classify and interpret not only canonical repair products, such as GC products, crossovers or indels, but also complex structural rearrangements, as copy number changes or complex genome-wide reorganizations \u003csup\u003e20-22\u003c/sup\u003e, providing deeper insight into the mechanisms driving mutagenesis, disease, and genome evolution \u003csup\u003e23\u003c/sup\u003e. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHere, we introduce Di-GRAPH (\u003cu\u003eD\u003c/u\u003eSB-\u003cu\u003ei\u003c/u\u003enduced \u003cu\u003eG\u003c/u\u003eenome-wide \u003cu\u003eR\u003c/u\u003eepair \u003cu\u003eA\u003c/u\u003enalysis and \u003cu\u003eP\u003c/u\u003erofiling of \u003cu\u003eH\u003c/u\u003eomologous recombination), a computational framework that integrates whole-genome sequencing with coverage and discordant read analyses to quantitatively characterize the frequency, directionality and tract length of individual GC events associated to the repair of a single and site-specific HO break. Di-GRAPH also analyses the relative contribution of short- and long-range HR \u003cem\u003evs\u003c/em\u003e NHEJ in repair and their associated mutagenic frequencies. At the genomic scale, Di-GRAPH detects and annotates genome-wide gross chromosomal rearrangements and associates them with specific genomic categories (ORF, intergenic region, LTR, TEG, Ty, tRNA, rRNA, ncRNA, snRNA, snoRNA, ARS, centromere and subtelomeric region). By comparing pre-damage, repair and survivor populations, Di-GRAPH enables the classification of lethal \u003cem\u003evs\u003c/em\u003e non-lethal rearrangements and separates DNA damage-dependent effects from cell cycle-associated genomic changes. The application of Di-GRAPH to four key HR mutants representing distinct stages of the pathway (Exo1, Sgs1, Srs2 and Rad51), revealed distinct stage-specific repair signatures, both at the break site and genome-wide, highlighting essential roles of these factors in maintaining genome stability. Notably, most genomic features showed a strong correlation between coverage and discordant-read signals, indicating that recombination-driven translocations are commonly associated with local amplification, whereas sequence identity analyses revealed feature-specific homology requirements, with ORFs displaying the strongest homology dependence and repetitive elements recombining under reduced sequence constraints.\u003c/p\u003e\n\u003cp\u003eTogether, these findings demonstrate that HR factors acting at different stages of the pathway differentially safeguard genome integrity, both locally and genome-wide, and highlight Di-GRAPH as a powerful tool to define genome behavior in response to DNA damage.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eSurvival after HO cleavage is uncoupled from DSB repair efficiency\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo examine how defects in specific steps of the HR pathway affect DSB repair and genome integrity, we focus on the role of the exonuclease Exo1 (long-range resection), the helicases Sgs1 and Srs2 (dissolution of recombination intermediates), and the recombinase Rad51 (homology search and strand invasion) in repairing an inducible HO break in the \u003cem\u003eS. cerevisiae\u0026nbsp;\u003c/em\u003ePMV background strain \u003csup\u003e24\u003c/sup\u003e. Briefly, this strain retains the endogenous wild-type \u003cem\u003eMAT\u003c/em\u003ea \u003cem\u003elocus\u003c/em\u003e on chromosome III but lacks the \u003cem\u003eHMR\u003c/em\u003e/\u003cem\u003eHML\u003c/em\u003e donor sequences, preventing intra-chromosomal repair after HO cleavage. Replacing the endogenous HO promoter with the inducible \u003cem\u003eGAL1\u0026nbsp;\u003c/em\u003epromoter enables galactose-dependent induction of a defined DSB at the \u003cem\u003eMAT\u003c/em\u003ea \u003cem\u003elocus\u003c/em\u003e. As a donor for recombination, a 1.3 kb \u003cem\u003eMAT\u003c/em\u003ea´ sequence is inserted at the \u003cem\u003eARG5,6\u003c/em\u003e \u003cem\u003elocus\u003c/em\u003e on chromosome V. This donor is identical to \u003cem\u003eMAT\u003c/em\u003ea except for 23 evenly spaced polymorphisms (~60 nt). To prevent cleavage of the \u003cem\u003eMAT\u003c/em\u003ea´ and re-cutting of gene conversion products, a C to G point mutation at the HO recognition site has been inserted (HO-\u003cem\u003einc\u003c/em\u003e) (Fig. 1A).\u003c/p\u003e\n\u003cp\u003eA single HO break was induced in wild-type, \u003cem\u003eexo1Δ\u003c/em\u003e, \u003cem\u003esgs1Δ\u003c/em\u003e, \u003cem\u003esrs2Δ\u003c/em\u003e, and \u003cem\u003erad51Δ\u0026nbsp;\u003c/em\u003ecells by adding galactose to the media. Samples were taken after 12 h for wild-type and \u003cem\u003eexo1Δ\u003c/em\u003e, and after 24 h for \u003cem\u003esgs1Δ\u003c/em\u003e, \u003cem\u003esrs2Δ\u003c/em\u003e, and \u003cem\u003erad51Δ\u003c/em\u003e. Coverage recovery of a 2–4 kb DNA region right to the HO break varied substantially among wild-type and HR-deficient strains, reflecting distinct capacities of each mutant to repair the DSB (Supplementary Fig. 1A, left). Nevertheless, despite varying extents of repair efficiency, all strains exhibited reduced cell viability relative to their repair capacity (Supplementary Fig. 1A, right). This indicates that survival after HO cleavage is not solely determined by the repair efficiency and differs between DNA repair mutants, suggesting that these factors may act beyond the DSB to support cell viability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHO-selected survivors of wild-type, \u003cem\u003eexo1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e, sgs1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e, srs2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003eand \u003cem\u003erad51\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003edisplay distinctive GC patterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the origin of cell lethality after DSB repair, we induced an HO cleavage and collected samples immediately after DNA damage checkpoint release (non-selected survivors, NSS hereafter), and after several rounds of culture dilution (selected survivors, SS hereafter). As control for genomic changes unrelated to HO cleavage, parallel samples were collected from un-induced raffinose cultures (undamaged cells, UD hereafter) (Fig. 1B).\u003c/p\u003e\n\u003cp\u003eWe first quantified HO repair efficiency in each strain at the specified time points. Three of the \u003cem\u003eMAT\u003c/em\u003ea´ polymorphisms generate \u003cem\u003eEco\u003c/em\u003eRI restriction sites (Supplementary Fig. 1B, left diagram), allowing the analysis of NHEJ \u003cem\u003evs\u0026nbsp;\u003c/em\u003eHR and GC directionality by Southern blotting \u003csup\u003e24\u003c/sup\u003e. Hybridization with the left and right probes showed that GC in wild-type, \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e, sgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eand \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003ewas asymmetric to the left side of the HO site, incorporating the two \u003cem\u003eEco\u003c/em\u003eRI sites to this side of the HO-\u003cem\u003einc\u0026nbsp;\u003c/em\u003e(Supplementary Fig. 1B, right panel). As expected, we detected a decrease in HR repair products in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003ecells, and a re-accumulation of the uncut band, indicative of DNA repair mainly by NHEJ in the survivor population (Supplementary Fig. 1B, right panel).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNext, Di-GRAPH was used to calculate the coverage surrounding the HO site (800 nt up- and downstream) by aligning 75-mer sequencing reads to the reference genome (bowtie \u003csup\u003e25\u003c/sup\u003e in -m 1 mode to exclude reads mapping to duplicated regions). Coverage at each time point was normalized to their respective T0 values (Fig. 1C). Profiles showed asymmetric left GC with concomitant gain/loss of chromosome V/III, as previously reported \u003csup\u003e24\u003c/sup\u003e. Coverage patterns differed between strains, reflecting strain-specific GC. Differences were also observed between NSS and SS samples, mainly in \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, likely due to repair efficiency differences (Fig. 1C, extended low coverage at CIII in NSS, and Supplementary Fig. 1B). Extending this analysis to the entire chromosome III revealed low-coverage regions around the HO break in NSS samples of these mutants (Supplementary Fig. 1C). Notably, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed a coverage loss between two transposon elements (85-150 kb) in SS, indicative of stable gross chromosomal rearrangements in the absence of this helicase. No coverage changes were observed in uninduced raffinose samples, confirming that these effects are DNA damage-dependent (Fig. 1C and Supplementary Fig. 1C).\u003c/p\u003e\n\u003cp\u003eTo precisely quantify GC profiles, we split reads in 18-mer and aligned them using bowtie in -k 1 mode. In this condition, reads within inter-polymorphism regions mapped randomly to \u003cem\u003eMAT\u003c/em\u003ea or \u003cem\u003eMAT\u003c/em\u003ea´, generating a basal coverage line, whereas reads overlapping polymorphisms mapped specifically to chromosome III or V \u003cem\u003eMAT\u0026nbsp;\u003c/em\u003eregions (Supplementary Fig. 2A). Coverage graphs were then generated for each mutant and time point (Supplementary Fig. 2B). By measuring the size of \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ peaks (distance between base and top/bottom), we quantified subtle changes in gain/loss at each polymorphism, reflecting small variations in GC, extension and directionality. NSS wild-type displayed declined \u003cem\u003eMAT\u003c/em\u003ea with a symmetrical \u003cem\u003eMAT\u003c/em\u003ea´ coverage enrichment extending up to 586 nt to the left of the break, with a preference for GC within the first 178 nt (Fig. 1D, left panel). No differences between NSS and SS were observed, indicating that GC tracts established during HO repair are stably inherited (Fig. 1D, right panel).\u003c/p\u003e\n\u003cp\u003eSimilar GC profile was observed in \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, including comparable GC length and directionality in NSS and SS, indicating that Exo1 does not influence GC in the PMV strain (Fig. 1D). In contrast, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e exhibited poor \u003cem\u003eMAT\u003c/em\u003ea´ gain and \u003cem\u003eMAT\u003c/em\u003ea loss in NSS, consistent with low repair efficiency at this time point (Fig. 1D, left panel). Among SS, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed extended \u003cem\u003eMAT\u003c/em\u003ea´ tracts compared to wild-type, suggesting that Sgs1 limits GC length (Fig. 1D, right panel), whereas \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed no major defects in GC length, symmetry, or efficiency (Fig. 1D, right panel). \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e survivors showed equal distribution of \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ polymorphisms, consistent with a complete loss of recombination between chromosomes III and V, except at the HO-\u003cem\u003einc\u003c/em\u003e, which retained an asymmetrical \u003cem\u003eMAT\u003c/em\u003ea´ gain and \u003cem\u003eMAT\u003c/em\u003ea loss pattern, indicating altered GC repair process (Fig. 1D, right panel). These GC profile variations were strictly HO-dependent, as un-induced raffinose cultures showed overlapping patterns with relative percentage near zero across all strains (Supplementary Fig. 3A).\u003c/p\u003e\n\u003cp\u003eIn summary, coverage profiling at the break site in the PMV background enables sensitive measurement of GC events, revealing stable strain-specific repair outcomes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscordant reads at the \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003eloci\u003c/em\u003e as a method to detect small variations in the frequency, directionality and extent of individual GC events\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDiscordant reads analysis has been established as a sensitive tool to determine small variations in the directionality and extent of GC products associated with the repair of an HO break by HR \u003csup\u003e19\u003c/sup\u003e. To gain further insight, we implemented this strategy in Di-GRAPH to assess GC efficiency in wild-type SS cells. Since the distribution of discordant reads delineates the two boundaries of GC events (Fig. 2A), we generated a densitometry map plotting the \u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea´ coordinates for each discordant read pair, which allowed us to precisely infer the symmetry and extent of individual GC events and to identify the preferred regions where recombination boundaries occur (Fig. 2B). GC events accumulated at the right flank of HO site and at a distal region to the left of the break, marking the boundaries of the recombination events. We also observed an accumulation of discordant reads proximal to the left side of the HO site, likely resulting from short-range DNA repair outcomes (i.e., NHEJ or short-range HR).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo facilitate the quantification of GC events distribution, we split reads in 18-mer, following the same approach described above. By quantifying the size of the \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ reads peaks (Supplementary Fig. 3B), we can accurately measure small variations in discordancy at each \u003cem\u003eMAT\u003c/em\u003ea´ polymorphism and thus detect subtle differences in the extension and directionality of GC events. Discordances at the right boundary of the HO break comprised the HO site (0 nt, \u003cem\u003eMAT\u003c/em\u003ea´) and the first polymorphism to the right (+64 nt, \u003cem\u003eMAT\u003c/em\u003ea) (Fig. 2C), indicating a drastic inhibition of GC to this side. Discordances to the left of the HO site ranged from -244 to the last polymorphism at -634 nt (Fig. 2C), reflecting variability in the length of GC products.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo detect the presence of discontinuities in GC events away from the HO break, we analyzed the distribution of discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ and \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea reads to the left and right of the HO break (Fig. 2D). Discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ and \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea reads in the bottom-left and top-right quadrants represent, respectively, the left and right boundaries of GC events (Fig. 2D and 2E, blue labels). Conversely, discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ and \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea reads in the top-right and bottom-left quadrants represent, respectively, right and left discontinuous GC events (Fig. 2D and 2E, red labels). Repair occurred preferentially by continuous GC to the left, while only a small fraction repaired by continuous GC to the right (Fig. 2F). The frequency of discontinuities at the HO site was also asymmetric, occurring more often to the left than to the right.\u003c/p\u003e\n\u003cp\u003eOverall, discordant read analysis enables precise mapping of GC tracts, revealing a strong leftward asymmetry at the HO cleavage site, with rare discontinuities also preferentially accumulating on this site of the break.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDistinctive DNA repair mutants show unique discordant \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea´ GC patterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo determine the influence of distinctive genes implicated at different levels of the HR repair pathway in GC efficiency, we applied the same approach described above to SS cells of \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants and assessed variation in frequency, length, symmetry and continuity of GC events.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLoss of Exo1 displayed a profile similar to that of wild-type cells in the accumulation of discordant reads, covering nearly all polymorphisms in the left flank of the HO site, with a pronounced accumulation at the last two polymorphisms (Fig. 3A, left graph and Supplementary Fig. 3B). Consistently, the discordant read density map showed a profile comparable to that of wild-type cells, with increased density at proximal right and distal left of the HO site at the same coordinates (Fig. 3A, middle graph). No changes were observed in the frequency of the -64 polymorphism, indicating a similar proportion of cells undergoing short-range DNA repair. Likewise, the analysis of GC continuity revealed no differences relative to the wild-type strain (Fig. 3A, right graph and Fig. 3E).\u003c/p\u003e\n\u003cp\u003eBoth \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD,\u003c/em\u003e and to a lesser extent, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003emutants showed a substantial accumulation of discordant reads in the left distal -634 to -586 region (Fig. 3B and 3C, middle graphs and Supplementary Fig. 3B). This result was supported by the increase in discordant reads associated with the left GC boundary observed in the discordant reads density maps of both mutants (Fig. 3B and 3C, left graphs), indicating that Sgs1 and Srs2 are implicated in restraining GC length, as previously suggested for Sgs1 \u003csup\u003e26\u003c/sup\u003e. The absence of Sgs1 or Srs2 did not alter the frequency of the -64 polymorphism, indicating no effect on the frequency of short-range DNA repair (Fig. 3B and 3C, middle graphs). Notably, both mutants exhibited an accumulation of discontinuous GC events, an effect particularly pronounced on the right side of the HO break (Fig. 3B and 3C right graphs and Fig. 3E), suggesting that these helicases are required to prevent HO-associated gaps in the recombinant products.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAbsence of Rad51 limited the appearance of discordant reads at distal regions of the HO break (Fig. 3D, left graph and Supplementary Fig. 3B), reflecting its inability to sustain long-range HR. However, we detected discordant reads spanning the HO site and its two flanking polymorphisms (-64 and +64), indicating that in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003ecells a fraction of cells can still incorporate the HO\u003cem\u003e-inc\u003c/em\u003e from chromosome V to chromosome III. This result is further supported by the presence of \u003cem\u003eMAT\u003c/em\u003ea´ coverage near the break site in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells (Fig. 1D, bottom right graph). Accordingly, the density map showed an accumulation of discordant reads at the HO/HO\u003cem\u003e-inc\u003c/em\u003e interface, with only a few events detected distal to the break (Fig. 3D, left graph). Discordancy symmetry was similar to that observed in wild-type cells, suggesting that the fraction of cells capable to sustain HR retains the ability to asymmetrically copy to the left side of the HO break (Fig. 3E). Finally, we observed a near absence of GC discontinuities (Fig. 3E), suggesting that these recombinant outcomes depend on Rad51.\u003c/p\u003e\n\u003cp\u003eAltogether, these results show that discordant reads are sensitive markers to reveal how different DNA repair mutants alter the frequency, length, symmetry and continuity of GC during HR.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHR repair mutants show unique HO associated mutagenic patterns upon DNA repair\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe have seen above that \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e, sgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e, srs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eand \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003eSS cells displayed different GC profiles upon HO repair. To assess the accuracy of this pathway and to determine the contribution of each DNA repair mutant to the associated mutagenic pattern, we devised an approximation within Di-GRAPH to link DNA repair by long/short-range HR or NHEJ to specific patterns of HO mutagenesis (Fig. 4A). We used the BWA aligner \u003csup\u003e27\u003c/sup\u003e, since its flexibility enables reliable mapping of insertions, deletions, and structural variants. 150-mer reads mapping at the HO break were filtered based on the presence of A/G\u003csup\u003e200689\u003c/sup\u003e at position -64 (first polymorphism left to the HO/HO-\u003cem\u003einc\u003c/em\u003e site) and C/T\u003csup\u003e200753\u003c/sup\u003e at position 0 (HO/HO-\u003cem\u003einc\u003c/em\u003e site) (Fig. 4A, top left diagram). Reads containing the combination T\u003csup\u003e200753\u003c/sup\u003e/G\u003csup\u003e200689\u003c/sup\u003e were associated with long-range HR (Fig. 4A, bottom left diagram), T\u003csup\u003e200753\u003c/sup\u003e/A\u003csup\u003e200689\u0026nbsp;\u003c/sup\u003ewith short-range HR (Fig. 4A, bottom middle diagram) and C\u003csup\u003e200753\u003c/sup\u003e/A\u003csup\u003e200689\u003c/sup\u003e with NHEJ (Fig. 4A, bottom right diagram). It is important to note that for all types of DNA repair, the combination T\u003csup\u003e200753\u003c/sup\u003e/G\u003csup\u003e200689\u003c/sup\u003e also represents reads derived from the donor \u003cem\u003eMAT\u003c/em\u003ea´ sequence on chromosome V. We re-aligned the three subsets of filtered reads to chromosome III and quantified the percentage of mismatches for any given position within the HO recognition site (-11 and +9 relative to the HO site), both before HO induction and in SS cells.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRepresentation of reads containing G\u003csup\u003e200689\u003c/sup\u003e and A\u003csup\u003e200689\u003c/sup\u003e at T0 for all strains revealed no associated mutations in any strain tested, indicating intact HO/HO\u003cem\u003e-inc\u003c/em\u003e sequences prior HO induction. Importantly, the proportions of C\u003csup\u003e200753\u003c/sup\u003e and T\u003csup\u003e200753\u003c/sup\u003e reads dropped to 50%, accurately reflecting the equal representation of reads mapping to the \u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003eloci\u003c/em\u003e, respectively (Supplementary Fig. 3C). \u0026nbsp;We did not detect long-range or short-range HR-associated mutations in the SS samples analyzed for any of the mutants assayed, indicating that the inherent HR defects linked to these mutants do not increase the HO mutation rate (Fig. 4B). Note that the high levels of base substitution at coordinate 0 nt reflect the C-to-T substitution intrinsic to HR repair, rather than a true mutation. We did not find A\u003csup\u003e200689\u003c/sup\u003e/C\u003csup\u003e200753\u003c/sup\u003e reads in wild-type and \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, indicating a complete absence of NHEJ repair in these strains (Fig. 4F). By contrary, NHEJ-associated mutations increased in \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e strains, and were markedly elevated in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants, with each strain displaying distinct mutagenic signatures (Fig. 4C, 4D and 4E). For instance, NHEJ-associated deletions in \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003ecells were accumulated between positions -1 nt and -3 nt, together with insertions at -2 nt, while \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed insertions at -3 nt (Fig. 4C and 4E). The \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutant exhibited extensive mutagenesis, including base substitutions and insertions from -1 to -3 nt, as well as widespread deletions across the entire DNA segment analyzed, most prominently between positions 0 and -3 nt (Fig. 4C and 4E).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRegarding DNA repair pathway choice, \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants preferentially repaired the HO break via long-range HR, followed by short-range HR. By contrast, \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants showed increased NHEJ and reduced long-range HR (Fig. 4F), ratifying the asymmetric \u003cem\u003eMAT\u003c/em\u003ea´/\u003cem\u003eMAT\u003c/em\u003ea reads distribution obtained in figure 1D. Interestingly, short-range HR was unaffected in the absence of Rad51, suggesting that this protein is not required to sustain short GC tracts.\u003c/p\u003e\n\u003cp\u003eTaken together, these results demonstrate that SS cells from different HR mutants display distinct HO-associated mutagenic signatures and DNA repair pathways choices. \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells preferentially use long-range HR, whereas \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e shifts toward error-prone NHEJ without affecting short-range HR, highlighting the differential impact of HR factors on DNA repair accuracy.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenome-wide sequencing analysis reveals altered coverage of distinctive genomic features in response to an HO break\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe have shown above that HO repair levels in the PMV background do not correlate with cell viability (Supplementary Fig. 1A), suggesting that there might exist other sources of genomic instability that affect cell recovery after HO repair. To address this, we used the genome-wide sequencing data to look for genomic coverage changes between NSS and SS wild-type cells within the Di-GRAPH framework. The rationale behind this experimental approach is that lethal genomic rearrangements occurring during the repair of the HO break in NSS might be counter-selected when screening SS cells that have overcome the DNA lesion. We also include in the analysis an undamaged raffinose culture to identify DNA damage-independent genomic reorganizations. We normalized the mean coverage of each of the 14518 genetic elements in NSS, SS and UD cells to their respective uninduced T0 coverage values. Coverage ratios exceeding 1 indicate a gain, whereas ratios below 1 indicate a loss, relative to the uncut HO state. For downstream bioinformatic analysis and data visualization, all genomic elements were subsequently grouped into 13 genomic categories comprising ORF, intergenic region, LTR, TEG, Ty, tRNA, rRNA, ncRNA, snRNA, snoRNA, ARS, centromere and subtelomeric region (Supplementary Fig. 3D).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnalysis of NSS cells showed that most of the genomic features within each category were very stable (as denoted for coverage values close to 1), except for the LTR (\u003cu\u003eL\u003c/u\u003eong \u003cu\u003eT\u003c/u\u003eerminal \u003cu\u003eR\u003c/u\u003eepeats), TEG (\u003cu\u003eT\u003c/u\u003eransposable \u003cu\u003eE\u003c/u\u003element \u003cu\u003eG\u003c/u\u003eenes) and Ty elements (Ty retrotransposons), where a drastic increment in coverage levels was detected (Fig. 5A, top graph). This coverage enrichment was counter-selected in SS cells (Fig. 5A, middle graph), suggesting that elevated levels of LTR, TEG and Ty during the damage response affects cellular viability. This graph resembles that obtained from UD cells (Fig. 5A, bottom graph), indicating a specific DNA damage-dependent effect on the genomic changes observed during the induction of the HO break.\u003c/p\u003e\n\u003cp\u003eTo precisely quantify DNA damage-specific genome-wide coverage changes, we plotted the ratio of NSS and SS samples relative to their corresponding UD sample. NSS displayed a 1.84% increase (C\u003csub\u003eWt\u003c/sub\u003e \u0026gt; 1.2) and a 0.73% decrease (C\u003csub\u003eWt\u003c/sub\u003e \u0026lt; 0.8) across all genomic elements analyzed (Supplementary Fig. 5A, top graphs), revealing an amplification of genome size, likely due to the hyper-recombinogenic state of DNA damaged cells. These values decreased to 1.23% and 0.3%, respectively, in SS cells (Supplementary Fig. 5A, bottom graphs), suggesting that most of these alterations are incompatible with cell viability.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo determine whether these changes were randomly distributed across the categories analyzed, we examined both the distribution of raw values from these samples (Fig. 5B) and the percentage of genomic changes within each genomic category (Fig. 5C). Most of the coverage enrichments in NSS cells originated from increases in LTRs, TEGs, Ty elements, and tRNAs, likely reflecting retrotransposition activity associated with DSB repair \u003csup\u003e28-30\u003c/sup\u003e (Fig. 5B,C, top graphs). We also observed a slight increase in subtelomeric regions and a reduction at centromeres. All these coverage changes were restored in SS cells (Fig. 5B,C, middle graphs), indicating that lethality of NSS is primarily associated with alterations in mobile elements, centromeres and subtelomeric regions. Importantly, although reduced in SS cells, mobile elements still displayed a residual signal, suggesting that DNA damage-dependent retrotransposon mobility might be buffered by the cell. These changes were DNA damage-dependent, as only small variations were detected in UD cells (Fig. 5B, bottom graph), except for subtelomeric regions (Fig. 5C, bottom graph).\u003c/p\u003e\n\u003cp\u003eNext, we asked whether the genomic changes observed in NSS cells were uniformly distributed across chromosomes. Consistent with the global coverage increase in coverage following HO induction (Supplementary Fig. 5A, top graph), most chromosomes exhibited a tendency toward increase in coverage levels, with some chromosomes more prone to enrichment than others (Fig. 5D, top graph). It is important to note that both \u003cem\u003eMAT\u003c/em\u003ea and \u003cem\u003eMAT\u003c/em\u003ea´ \u003cem\u003eloci\u003c/em\u003e were excluded from the analysis to avoid chromosome III/V coverage artifacts resulting from HR repair. These results indicate that genome reorganization in response to a HO break is not equally distributed across all chromosomes. As before, although a residual tendency toward coverage enrichment was still detectable, most of these variations were reduced in SS and UD cells (Fig. 5D, middle and bottom graph).\u003c/p\u003e\n\u003cp\u003eOverall, these data suggest that the genome exhibits a degree of plasticity in response to a DSB. Genomic changes vary across genomic categories and chromosomes, implying that certain regions are more susceptible to alteration, likely due to local genomic context or regulatory influences during DNA repair.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eexo1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e,\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and \u003cem\u003erad51\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eD\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;cells exhibit distinct genomic coverage profiles in response to DNA damage\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate the genomic consequences of an HO-induced break in the absence of stage-specific factors of the HR pathway, we applied the same experimental approximation described above to the \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants. NSS \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed distinct coverage changes compared with the wild-type (C\u003csub\u003eexo1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 3; C\u003csub\u003eexo1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 0.57), with a general tendency toward increased coverage levels (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs), particularly at mobile elements and centromeres (Fig. 6A and Supplementary Fig. 5B). Conversely, decreased coverage was observed at rDNA \u003cem\u003eloci\u003c/em\u003e and, irregularly, at the subtelomeric regions (Fig. 6A and Supplementary Fig. 5B). Gain in chromosome coverage appeared randomly distributed, with a slight enrichment on chromosome I (Fig. 6B, top graphs). As in wild-type cells, SS \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e exhibited a pronounced reduction in coverage alterations (C\u003csub\u003eexo1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 1.19; C\u003csub\u003eexo1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 0.34) when compared with NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), across all chromosomes (Fig. 6B, middle graphs), and most notably at mobile elements (Fig. 6A and Supplementary Fig. 5B). This suggests that retrotransposition in the \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e background contributes to cell lethality following DNA repair. Notably, the reduction in rDNA coverage was maintained in SS cells (Fig. 6A and Supplementary Fig. 5B), indicating that this feature does not compromise cell viability in response to DNA damage. No major coverage alterations were detected in UD \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), except for a minor, randomly distributed signal at subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs). These findings suggest that Exo1 is necessary to prevent mobile element transposition events and to maintain rDNA copy number following DNA damage.\u003c/p\u003e\n\u003cp\u003eNSS \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells exhibited coverage profiles that differed from those of the wild-type, reflecting highly dynamic coverage variations (C\u003csub\u003esgs1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 1.77; C\u003csub\u003esgs1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 1.59) (Supplementary Fig. 5A, top graphs). Increased coverage was detected at mobile elements (Fig. 6A and Supplementary Fig. 5B), randomly affecting all chromosomes (Fig. 6B, top graphs). Interestingly, SS \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells still displayed pronounced coverage alterations (C\u003csub\u003esgs1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 1.7; C\u003csub\u003esgs1\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 0.86) compared with NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), suggesting the accumulation of non-lethal genomic alterations. Mobile elements and subtelomeric regions displayed a markedly reduced coverage, whereas tRNA, ncRNA and rRNA showed pronounced increases (Fig. 6A and Supplementary Fig. 5B), both patterns occurring randomly distributed across all chromosomes (Fig. 6B, middle graphs), indicating widespread genomic imbalances. Unlike the wild-type, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells also showed pronounced coverage changes in UD conditions (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B). These variations spanned multiple genomic categories, with particularly strong effects at rDNA and subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs). Collectively, these results demonstrate that Sgs1 is critical for preserving genome integrity, even in the absence of DNA damage.\u003c/p\u003e\n\u003cp\u003eNSS \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed pronounced coverage changes compared with the wild-type (C\u003csub\u003esrs2\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 2.28; C\u003csub\u003esrs2\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 2.39), with both gains and losses in coverage (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs). Coverage increases were enriched at mobile elements and subtelomeric regions (Fig. 6A and Supplementary Fig. 5B), whereas losses were predominately observed at centromeres (Fig. 6A and Supplementary Fig. 5B). Chromosome coverage gains appeared largely random, with a slight enrichment on chromosome I (Fig. 6B, top graphs). SS \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed a marked reduction in coverage alterations (C\u003csub\u003esrs2\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 1.29; C\u003csub\u003esrs2\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 0.20) relative to NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), for all chromosomes (Fig. 6B, middle graphs) and most evident at mobile elements (Fig. 6A and Supplementary Fig. 5B), indicating that retrotransposition in the absence of Srs2 contributes to cell lethality following DNA repair. Similar to \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, SS \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed an increase in rDNA coverage (Fig. 6A and Supplementary Fig. 5B), suggesting that although both helicases suppress rDNA expansion, this property does not impair cell viability under DNA damage. No major coverage alterations were detected in UD \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), aside from a minor decrease at centromeres and a localized increase at the subtelomeric regions (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs), with the strongest effect on the recipient chromosome III (Fig. 6B, bottom panel). Taken together, these results demonstrate that Srs2 plays a critical role in preventing retrotransposition and maintaining centromere stability in response to DNA damage.\u003c/p\u003e\n\u003cp\u003eNSS \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed the most pronounced coverage changes of all strains analyzed (C\u003csub\u003erad51\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 2.13; C\u003csub\u003erad51\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 2.87), with both gains and losses occurring in a pattern similar to that observed in \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003ecells (Supplementary Fig. 4A and Supplementary Fig. 5A, top graphs). Coverage gains were enriched at mobile elements and subtelomeric regions (Fig. 6A and Supplementary Fig. 5B), whereas losses were more prominent at snoRNAs, ARSs, centromeres and subtelomeric regions, and to a lesser extent to other genomic categories (Fig. 6A and Supplementary Fig. 5B). Chromosome coverage gains appeared largely random, with a slight enrichment on chromosome VI, while coverage losses were distributed uniformly across all chromosomes (Fig. 6B, top graphs). SS \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed a marked reduction in coverage alterations (C\u003csub\u003erad51\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026gt; 1.2 = 0.92; C\u003csub\u003erad51\u003c/sub\u003e\u003csub\u003eD\u003c/sub\u003e \u0026lt; 0.8 = 0.43) relative to NSS cells (Supplementary Fig. 4B and Supplementary Fig. 5A, bottom graphs), extending across all chromosomes (Fig. 6B, middle graphs), while subtelomeric regions still exhibited slightly high coverage levels (Fig. 6A and Supplementary Fig. 5B). Notably, unlike the other strains analyzed, mobile elements in SS cells were completely downregulated, indicating a severe impact of retrotransposition on lethality in the absence of Rad51 (Fig. 6A and Supplementary Fig. 5B). No major coverage alterations were detected in UD \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells (Fig. 6A, 6B, Supplementary Fig. 4C and Supplementary Fig. 5B), aside from a minor decrease at centromeres and an increase at subtelomeric regions, again resembling the \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutant (Fig. 6A, bottom graphs and Supplementary Fig. 5B, bottom graphs), with no bias against any specific chromosome (Fig. 6B, bottom panel). Taken together, these results indicate that Rad51 deficiency produces a coverage profile similar to that of \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, suggesting that both mutants have similar consequences for genome stability in response to DNA damage.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenome-wide analysis of discordant reads reveals global genome reorganizations in response to a DSB\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCoverage-based approaches are useful for identifying prominent biases in the duplication or depletions of genomic elements but cannot determine their relocation within the genome. Moreover, such changes must accumulate iteratively at the same \u003cem\u003eloci\u003c/em\u003e for coverage gains or losses to be detectable and attributable to specific genomic elements. To address this limitation, we implemented Di-GRAPH with a module to apply the same discordant reads analysis \u003csup\u003e19\u003c/sup\u003e used above to map GC boundaries, this time with the aim of detecting global recombination events across the entire genome. Reads from T0, NSS, SS and UD wild-type cells were aligned using Bowtie2 avoiding multimapping, and discordant reads were filtered by requiring both mates of the pair to align to different chromosomes with no mismatches allowed. For internal normalization, we used a custom reference genome containing an intact \u003cem\u003eADE3\u003c/em\u003e gene on chromosome VII and an artificial chromosome carrying the \u003cem\u003epGal1-10-HO-SSB1-vector\u0026nbsp;\u003c/em\u003esequence (chromosome Disc), identical to the construct integrated at the \u003cem\u003eADE3\u003c/em\u003e gene at chromosome VII in the PMV background (Supplementary Fig. 6A). In addition, the endogenous \u003cem\u003epGAL1-10\u003c/em\u003e and \u003cem\u003eHO-SSB1\u003c/em\u003e sequences were replaced with Ns to prevent removal of duplicated reads mapping to both \u003cem\u003eloci\u003c/em\u003e in the reference genome. Under this set up, concordant reads at the \u003cem\u003eADE3-Vector\u003c/em\u003e and \u003cem\u003epGAL1-10-ADE3\u003c/em\u003e interfaces appear as discordant reads between chromosome VII and the chromosome Disc (Supplementary Fig. 6A), thereby enabling the quantitative normalization of discordances across the genome (Disc-normalization). This normalization was applied to all subsequent analysis of discordant reads.\u003c/p\u003e\n\u003cp\u003ePolymerization errors during library amplification or sequencing mistakes might generate erroneous reads that, if misalign relative to their pair, may produce false discordant reads (Fig. 7A). To minimize these artifacts, we implemented a bioinformatic approach in which, for a true discordant pair, both reads should not concordantly re-align under relaxed BLAST alignment conditions (Fig. 7A). In this scenario, concordant re-alignments increase the likelihood of identifying a false discordant pair (Fig. 7A). We filtered read pairs and classified them based on their ability to remain discordant (no alignment, hereafter NA), along with those containing\u0026nbsp;³3, 2 or 1 nucleotide variants that rendered them to a concordant state, in order to evaluate the reliability of these discordances (Fig. 7A). Finally, artifactual discordances arising from variations between the reference genome and the genetic PMV background were filtered by removing discordant read pairs detected at the T0 timepoint in any of the three experimental replicates.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEach read of the pair from the NA group was plotted in a dot array comparing NSS and SS samples (Fig. 7B, left panel, left-bottom and right-top triangle respectively), mapping the two genomic features involved in each recombination event, with signal intensity reflecting the frequency of the rearrangement. The most abundant discordances appeared between two ORFs, followed by ORF-intergenic regions. Consistent with the increase in mobile elements and subtelomeric coverage observed in NSS cells (Fig. 5B and 5C), we detected higher frequency of discordancy in these regions (Fig. 7B, bottom right graph) compared to SS cells. Each genomic category showed a prevalence of discordances at ORFs and intergenic regions, followed by the rDNA and ARS sequences. Interestingly, rDNA discordances increased specifically in SS cells, indicating non-lethal translocation events in this genomic region in response to DNA damage (Fig. 7B, top right graph). Successive relaxation of the discordance confidence score slightly increased discordances at mobile elements (Ty, TEG ad LTR) and subtelomeric regions only in the NSS sample, again supporting the presence of lethal transposition events during the damage response (Supplementary Fig. 7A).\u003c/p\u003e\n\u003cp\u003eTogether, these results indicate that discordancy analysis is a reliable method to assess genome rearrangements during the damage response and suggest that ORF, intergenic regions, rDNA, transposons, ARS and subtelomeric translocations are the most frequent genomic rearrangements occurring during the induction of a DSB. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenome-wide analysis of discordant reads reveals distinctive patterns of genome reorganization in the absence of DNA repair factors\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate how different steps of the repair pathway influence genome integrity, we applied the experimental workflow described above to \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells under NSS, SS and UD conditions (Fig. 8A and Supplementary Fig. 6B). To minimize false-positive discordance, we restricted the analysis to filtered pairs from the NA group. Additionally, to visualize the genomic localization of these events, we plotted genomic maps representing the discordance frequency of every genomic feature that was detected in our three replicates (Fig. 8B and Supplementary Fig. 6C, D).\u003c/p\u003e\n\u003cp\u003eSupporting the tendency toward rDNA coverage decay, loss of Exo1 resulted in low discordances at this \u003cem\u003elocus\u003c/em\u003e in NSS cells (Fig. 8A). We also detected a slight increase in discordance levels at ORF and intergenic regions, especially in UD cells, supporting a DNA damage-independent role for the nuclease in maintaining genome stability (Fig. 8A). This discordancy was evenly distributed along all genomic features comprising these structures (Supplementary Fig. 6D). Transposons showed a modest accumulation in SS samples when the discordance confidence score was relaxed (Supplementary Fig. 7B), indicating certain accumulation of Ty transposition following DNA repair. By contrast, the absence of Sgs1 leads to a drastic accumulation of ORF and ARS discordances in SS cells, accompanied by a rDNA reduction in NSS conditions (Fig. 8A,B and Supplementary Fig. 6C,D), indicating a specific role for Sgs1 in maintaining these genomic structures, particularly during the damage response. Discordances in \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells were uniformly distributed across the entire genome, apart from a strong single hotspot located in an intergenic region that appeared under both NSS and SS conditions but was absent in the UD sample (Fig. 8B and Supplementary Fig. 6C,D). Supporting coverage data, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells also led to accumulation of Ty and TEG elements in NSS cells (Fig. 8A,B and Supplementary Fig. 6B), an effect that became dramatically enhanced under relaxed confidence score settings (Supplementary Fig. 7C, left panel). This elevated transposition activity is likely deleterious, as it was drastically reduced in SS cells (Fig 8A, Supplementary Fig. 6B,C and Supplementary Fig. 7C middle panel). Similarly to \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, a \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutant also exhibited high levels of transposition events under NSS conditions (Fig. 8A, Supplementary Fig. 6B). However, unlike \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells retained elevated transposon-associated discordances in SS samples under relaxed confidence score conditions (Supplementary Fig. 7D), suggesting that this helicase is essential for preventing retrotransposition in response to a DSB. Consistent with the coverage records, loss of Rad51 resulted in increased Ty, TEG and LTR discordances mainly under NSS conditions (Fig 8A,B, Supplementary Fig. 6B and Supplementary Fig. 7E), indicating that transposition becomes lethal when HR is impaired. Importantly, some Rad51-dependent NSS discordances generated strong hotspots across distinct ORFs and intergenic regions, indicating fragile genome sites in the absence of this repair factor (Fig. 8B). \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells also displayed a high frequency of ORF-, ARS- and ncRNA-associated discordances in UD samples (Fig. 8A), highlighting a genome-stabilizing function of Rad51 even in the absence of DNA damage.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIdentifying strain-specific genome rearrangement networks via discordant reads analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo detect genomic regions with a high prevalence of rearrangements, we filtered the reads used above to retain only those mapping to \u003cem\u003eloci\u003c/em\u003e with discordant frequency higher than 6%. Each discordant pair from the assessed strains was assigned a distinct color, allowing the identification of strain-specific discordance patterns. We then connected the two reads of each pair to construct a discordance network of NSS, SS and UD (Fig. 8C, Supplementary Fig. 8A,B). To quantify the interactions observed in this network, we calculated the percentage for each discordant pair, generating an interaction map that illustrates the recombination probability between all pairs of genomic categories for each condition tested (Fig. 8D).\u003c/p\u003e\n\u003cp\u003eConsistent with the high levels of retrotransposition observed earlier, several Ty and TEG elements occupied central positions in the NSS network, particularly in the \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e mutants (Fig. 8C), being transposon-ORF the most frequent recombination events (Fig. 8D). This effect was particularly evident in the \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e strain, which also showed transposon-transposon recombination (Fig. 8D). Recombination at the rDNA cluster was also common, occurring in all strains except \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e. The rDNA cluster appeared largely separated from the rest of the network, suggesting that rDNA rearrangements occur independently of the elevated transposition state characteristic of the NSS condition (Fig. 8C). The primary recombination partners for rDNA were ORFs and intergenic regions (Fig. 8D). ARS-associated discordances in \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e predominantly involved ORF and LTRs (Fig. 8D). ORFs and intergenic regions were located at the periphery of the network (Fig. 8C), often in a strain-specific manner, typically recombining with other ORFs, and in the case of \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e strains, also with transposons. Interestingly, \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e and \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells showed a marked reduction in ORF-associated discordance (Fig. 8D), implicating these factors in the rearrangements of coding regions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast to NSS conditions, SS cells exhibited minimal Ty-associated recombination, indicating that retrotransposition in response to DNA damage is largely lethal (Fig. 8D and Supplementary Fig. 8A). Notably, the rDNA cluster shifted to the center of the network in SS cells, a pattern observed across all strains (Supplementary Fig 8A). The stabilization of rDNA recombination in SS cells confirms that these events are not inherently lethal. The absence of Sgs1 led to a pronounced increase in ORF-ORF and ORF-intergenic recombination (Fig. 8D), indicating that this helicase in essential for preventing ectopic rearrangements at these genomic regions following a DSB.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSimilarly to SS conditions, UD cells also showed reduced transposition activity (Fig. 8D and Supplementary Fig. 8B), reinforcing DNA damage as the primary source of retrotransposition. Consistent with the elevated discordance frequencies observed in \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells in the absence of damage, this mutant dominated the UD network, confirming a DNA damage-independent role for Exo1 in genome maintenance. \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed enhanced ORF-ORF and ORF-intergenic recombination, with similar, though less pronounced, patterns in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells (Fig. 8D). Interestingly, \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003ecells exhibited a marked reduction in ORF-associated discordances under UD conditions (Fig. 8D), suggesting that this helicase may be important for facilitating or regulating the ectopic recombination events that arise in the absence of DNA damage.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIntegration of coverage and discordant read data enables precise characterization of recombination identity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs shown above, both coverage measurements and discordant-reads analysis can detect copy number variations of specific genomic features and track their translocation across the genome. To evaluate whether these approaches could be integrated to determine the nature of the recombination events, we grouped NSS, SS and UD data and collectively compared coverage values with discordance ratios (Fig. 8E). Coverage and discordance were generally well correlated, indicating that translocations commonly involve genomic amplification arising from recombination. Most correlated features corresponded to transposons, whose amplification during HO induction may enhance their translocation potential. Some ORFs and intergenic regions also showed strong correlations. Notably, the presence of paralogues for these coding regions in the genome contributes to their recombination bias. A subset of features showed high coverage but low discordance (Fig. 8E), consisting mainly of Ty elements whose elevated amplification does not necessarily result in integration, or whose rearrangements may be difficult to detect using our discordance-based assay. Conversely, features displaying low coverage but high discordance were largely represented by the non-transcribed rDNA region NTS1-2 (Fig. 8E). This pattern suggests that rDNA changes during the DNA damage response reflect translocation of rDNA units rather than expansion or contraction of the rDNA array.\u003c/p\u003e\n\u003cp\u003eTo evaluate whether the HO-induced global rearrangements observed across the analyzed genomic categories depended on homology-directed recombination (HDR), we examined the set of 110 highly discordant features used before (Fig. 8C and Supplementary Fig. 8A,B) and computed their random pairwise sequence identity relative to every other feature in the genome (Fig. 8F, blue line). We then calculated the sequence identity for each actual discordant pair, representing the true homology between the two features involved in the rearrangement (Fig. 8F, red line). The ratio between the observed and random identity values reflected the degree of HDR. Higher ratios correspond to rearrangements requiring greater sequence similarity, whereas lower ratios indicated recombination events that tolerate reduced homology. Strikingly, ORFs, which displayed the lowest random identity, showed the highest HDR ratios, suggesting that coding regions are highly selective with respect to sequence homology during recombination. In contrast, LTRs, intergenic regions and rDNA, all of which exhibited higher random identity values, displayed lower homology dependence, implying that recombination at these \u003cem\u003eloci\u003c/em\u003e requires comparatively less sequence similarity. Retrotransposons showed a more variable pattern but generally trended toward higher HDR, consistent with the notion that most of their insertion events involve other retrotransposons and in agreement with the interaction network shown in figure 8C and 8D.\u003c/p\u003e\n\u003cp\u003eIn summary, these results demonstrate that integrating coverage profiles with discordant read analysis provides a powerful framework for characterizing the nature of recombination events at a genome-wide scale. This combined approach automated within Di-GRAPH does not only enables the detection and classification of rearrangements with high resolution but also reveals how distinct genomic features behave during the cellular response to DNA damage. By capturing both the structural outcomes and the underlying homology relationships, this method offers a comprehensive view of the recombination landscape and the dynamic contributions of different genomic elements during repair.\u0026nbsp;\u003c/p\u003e\n"},{"header":"Discussion","content":"\u003cp\u003eRepair of DSBs by HR requires the coordinated execution of multiple steps, yet how defects in individual HR components shape genome integrity beyond the DNA break remains poorly understood. Using a defined HO-induced break in the \u003cem\u003eS. cerevisiae\u003c/em\u003e PMV background, we developed Di-GRAPH, a user-friendly bioinformatic software that integrates DSB mutational signature analysis, repair pathway choice, coverage profiling, and discordant read mapping to quantitatively measure both local and genome-wide structural rearrangements. By applying Di-GRAPH to cells deficient in Exo1, Sgs1, Srs2 and Rad51, we determined the influence of each stage of the HR pathway at the cleavage site as well as genome-wide.\u003c/p\u003e\n\u003cp\u003eHigh-resolution Di-GRAPH profiling of GC tracts demonstrates that these outcomes are not only strain-specific, but that the associated HO repair signatures are stabilized in the survivor population. Wild-type and \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells displayed nearly identical GC tract lengths and asymmetry, indicating that Exo1-mediated long range resection is largely dispensable for GC formation, in line with previous results \u003csup\u003e24\u003c/sup\u003e. \u0026nbsp;In contrast, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003eand to a lesser extent \u003cem\u003esrs2\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003emutants, accumulate extended GC tracts, consistent with their anti-recombinogenic activities in disassembling D-loop intermediates\u0026nbsp;\u003csup\u003e9-11\u003c/sup\u003e. Notably, both mutants also displayed elevated discontinuities within GC tracts, particularly on the right flank of the break, suggesting that Sgs1 and Srs2 are required to prevent inappropriate template switching intermediates during HR. Nevertheless, since both helicases have been reported to bias DNA repair toward non-crossover outcomes\u0026nbsp;\u003csup\u003e31\u003c/sup\u003e, it is plausible that these GC discontinuities represent crossover products that accumulate at the HO cleavage boundary during DNA repair. As expected, long-range GC events were drastically reduced in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells, although these cells retained the characteristic left-side asymmetry, indicating that Rad51 is not required for generating GC asymmetry. Consistent with the established role of Rad51 in HR, NHEJ products accumulated in SS cells in this mutant. Remarkably, short-range GC at the HO site was unaffected by the absence of Rad51, demonstrating that this type of repair is Rad51-independent. In agreement with these findings, previous work has shown that the loss of Rad51 stimulates allelic GC and break-induced replication by approximately 30-fold in a Rad52-dependent manner\u0026nbsp;\u003csup\u003e32\u003c/sup\u003e. Determining whether short-range GC depends on Rad52 remains an important question for future investigation.\u003c/p\u003e\n\u003cp\u003eAnother feature of Di-GRAPH is that it enables the quantification of repair pathway choice and the mutational patterns associated with long- and short-range HR, as well as with NHEJ. Wild-type and \u003cem\u003eexo1\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003ecells repaired the HO break almost exclusively through HR, without detectable mutagenesis. In contrast, \u003cem\u003esgs1\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003eand\u003cem\u003e\u0026nbsp;srs2\u003c/em\u003e\u003cem\u003eD\u0026nbsp;\u003c/em\u003emutants displayed a slight increase in NHEJ usage accompanied by distinct mutational signatures (short deletions and insertions localized to the -1 to -3 region) suggesting that these helicases might also function as NHEJ inhibitors. Importantly, and in line with recent discoveries showing that binding of Rad51 to the break prevents single-strand annealing (SSA) or alternative end-joining (A-EJ) DSB repair \u003csup\u003e33\u003c/sup\u003e, \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e cells exhibited the most severe shift toward NHEJ and accumulated extensive base substitutions and indels along the HO cleavage site. These results highlight that loss of stage-specific HR factors induces distinct mutational patterns, emphasizing that genome integrity is determined not only by whether breaks are repaired, but also by how cells redistribute repair between alternative pathways when canonical HR steps are compromised.\u003c/p\u003e\n\u003cp\u003eIn addition to analyzing DSB repair at the break site, Di-GRAPH is integrated with a genome-wide coverage analysis that identifies DNA damage-dependent changes of coverage across distinctive genomic categories. This analysis revealed widespread alterations in rDNA, mobile elements, centromeres, tRNAs, ARSs, and subtelomeric regions following HO induction. NSS samples exhibited extensive gains in LTR, Ty elements and TEGs, consistent with the activation of retrotransposition during the DNA damage response \u003csup\u003e28\u003c/sup\u003e. The dramatic reduction of these rearrangements indicates that retrotransposition-induced lesions are frequently lethal and subjected to negative selection during DSB recovery \u003csup\u003e34\u003c/sup\u003e. Subtelomeric expansions and centromere losses, observed in NSS but absent in SS, further highlight the inherent fragility of repetitive or structurally constrained regions during HR-mediated repair. Nevertheless, each HR mutant exhibited a distinct genome-wide signature, indicating that these factors differentially contribute to genome stability at sites distant form the DSB. Loss of Exo1 showed pronounced increases in mobile elements and a reduction in rDNA coverage, suggesting that this nuclease both restrains retrotransposition and preserves rDNA structure following DNA damage. In agreement with the role of Sgs1 in preventing duplication-mediated gross chromosomal rearrangements \u003csup\u003e35\u003c/sup\u003e, loss of this helicase displayed persistent genome imbalance even in undamaged conditions. Interestingly, this role was particularly pronounced for the rDNA, which exhibited the most substantial alterations among all features analyzed in this mutant. In this context, it has been reported that rDNA is permissive to RAD51 accumulation in the absence of BLM (the human homolog of Sgs1), leading to rDNA micronucleation and potentially contributing to global genome instability \u003csup\u003e36\u003c/sup\u003e, a phenotype reminiscent of the coverage profile observed in the absence of Sgs1. The absence of Srs2 or Rad51 led to extensive coverage gains in mobile elements and losses in centromeric regions, displaying highly similar profiles, suggesting shared vulnerabilities in suppressing genome rearrangements in these mutants. Since Srs2 is required to remove Rad51 from ssDNA \u003csup\u003e37\u003c/sup\u003e, it is plausible that the absence of Srs2 exhausts the pool of free Rad51, providing a potential explanation for the similar phenotypes observed in both mutants in genome-wide analyses. Interestingly, the complete suppression of these elevated mobile element coverage signal in \u003cem\u003erad51\u003c/em\u003e\u003cem\u003eD\u003c/em\u003e SS cells indicates that such events are lethal to the cell. These data demonstrate that genome instability arising during DSB repair is not uniform but is instead strongly influenced by the specific HR step that is impaired, with mobile elements, subtelomeric regions, and centromeres emerging as major contributors to cell lethality.\u003c/p\u003e\n\u003cp\u003eApplying discordant read mapping across the genome allowed direct detection of recombination events, overcoming limitations of coverage-based approaches, particularly their low sensitive and lack of spatial information. Several bioinformatic approaches have been developed to use discordant read as a source to detect genomic rearrangements \u003csup\u003e14,38\u003c/sup\u003e. Di-GRAPH incorporates a novel algorithm that not only identifies chromosomal rearrangements based on the paired-end reads discordance, but also evaluates the probability that these events represent true rearrangements rather than false-positive discordant reads generated during library preparation or DNA sequencing. Moreover, the use of a Disc-normalization allows precise quantification of these events, making this approach well suited for assessing differences in genome-wide rearrangements across distinct strains or experimental conditions. This analysis reveals widespread translocations involving ORFs, intergenic regions, rDNA, ARS sequences and mobile elements during the repair of a single HO break. The strong accumulation of discordant reads in mobile elements and subtelomeric regions in NSS, but not in SS cells, parallels the coverage analysis, validating this approach and suggesting that element mobility frequently results in duplications of the recombination regions involved. Importantly, rDNA rearrangements were enriched specifically in SS cells, suggesting that certain rDNA translocations are tolerated or even selectively retained following repair. In contrast, rearrangements involving mobile elements were largely eliminated in SS samples, consistent with their detrimental impact on cell survival following DNA repair \u003csup\u003e14\u003c/sup\u003e. Discordant read mapping in HR mutants further emphasized the diversity in genome reorganizations that arise when specific repair steps are compromised, reveling \u003cem\u003elocus\u003c/em\u003e- and chromosome-specific patterns. Finally, our analysis of HDR reveals key insight into the mechanisms driving HO-induced genome rearrangements. By comparing observed versus random sequence identity of discordant pairs, we found that ORFs, despite their low baseline similarity, exhibited the highest HDR ratios, indicating that recombination in coding regions is strongly constrained by sequence homology. In contrast, LTRs, intergenic regions, and rDNA, which have higher inherent similarity, showed lower HDR dependency, suggesting that recombination at these \u003cem\u003eloci\u003c/em\u003e tolerates reduced homology. Retrotransposons displayed a more variable pattern but generally relied on HDR, consistent with recombination among related elements, as supported by the discordant interaction networks. Thus, the interplay between local sequence similarity and the genomic context emerges as a central factor governing recombination outcomes.\u003c/p\u003e\n\u003cp\u003eOverall, our findings show that the consequences of a single DSB extend well beyond the break site and are strongly and distinctively influenced by stage-specific HR factors. By integrating local repair signatures with genome-wide alterations we have revealed that global genome instability, rather than DSB repair efficiency alone, is an important determinant for cell survival following DNA damage. Finally, our results stablish Di-GRAPH as a powerful framework for dissecting the consequences of a DSB both at the break site and genome-wide, providing a versatile framework for studying genome maintenance in diverse biological contexts.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eYeast strains and growing conditions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe strains used in this work are listed in Supplementary Table 1. Disruption of endogenous genes was achieved by gene targeting using PCR products as described previously \u003csup\u003e39,40\u003c/sup\u003e. Details of PMV construction have been described in Ramos et al., 2022 \u003csup\u003e24\u003c/sup\u003e. PMV was HO-induced by adding 2% galactose to cells growing in YP with 2% raffinose. Samples were collected for DNA analysis before and at distinctive timings after adding galactose to the media.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChemiluminescent Southern blotting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCell cultures (10 ml) adjusted to an OD\u003csub\u003e600\u003c/sub\u003e of 0.4 were collected by centrifugation and subsequently washed with 1 ml of PBS. Following a second centrifugation step, pellets were flash frozen and stored at -80\u0026ordm;C. Cell lysis was carried out by incubating the pellets for 10 min with 40 units of lyticase in DNA extraction buffer (1% SDS, 100 mM NaCl, 50 mM Tris-HCl, 10 mM EDTA). Genomic DNA was isolated by extracting the lysates with phenol:chloroform:isoamylalcohol (25:24:1) for 10 min. After centrifugation, the aqueous phase was ethanol-precipitated and the DNA pellet was resuspended in water. Purified genomic DNA was digested with \u003cem\u003eEco\u003c/em\u003eRI, resolved on 1% agarose gels, and analyzed by Southern blotting. Hybridization probes were generated by labeling a PCR-amplified DNA fragment using a nucleotide mixture containing fluorescein-12-dUTP (Fisher Scientific, 10354280). The oligonucleotides used for probe synthesis are listed in Supplementary Table 2. Detection was performed with an anti-fluorescein antibody (F(ab\u0026acute;)2 Anti-FLUORESCEIN Alkaline Phosphatase Conjugated, Life Technology, 700-105-096) diluted 1:200,000. Membranes were developed using the CDP-Star chemiluminescent substrate (Amersham, RPN3682) and visualized by exposure to autoradiography films.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenomic sequencing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGenomic DNA (1mgr) was used as input for sequencing library construction and subsequently sequenced on a DNBSEQ G400 platform in 150 paired-end mode (BGI) to obtain 2GB data per sample, accounting for 8M median reads per sample. In brief, DNA molecules were enzymatically fragmented and processed through standard end-repair and 3\u0026acute;-adenylation steps prior to adaptor ligation. Adaptor-modified fragments were amplified by PCR, after which the products were converted into single-stranded circular templates and subjected to rolling-circle amplification to generate DNA nanoballs (DNBs). Sequencing was carried out using the combinatorial probe-anchor synthesis (cPAS) method. Raw reads were filtered to remove adaptor-contaminated or low-quality sequences using SOAPnuke with the manufactured-recommended parameters (-n 0.001 -l 20 -q 0.4 --adaMis 3 --rmdup \u0026ndash;minReadLen 150). Sequenced reads were aligned to the \u003cem\u003eS. cerevisiae\u003c/em\u003e PMV reference genome (Supplementary Data, RG_PMV_v9.fasta) for downstream analysis. For visualization of the BedGraph \u003csup\u003e41\u003c/sup\u003e coverage profiles we used the Integrative Genomics Browser IGB (v9.0.2) \u003csup\u003e42\u003c/sup\u003e. Figures and analyses were done using the following R packages: dplyr (v1.1.4) \u003csup\u003e43\u003c/sup\u003e, tidyr (v1.3.1) \u003csup\u003e44\u003c/sup\u003e, stringr (v1.6.0)\u003csup\u003e45\u003c/sup\u003e, ggplot2 (v3.5.2) \u003csup\u003e46\u003c/sup\u003e, purrr (v1.0.4) \u003csup\u003e47\u003c/sup\u003e and ggbreak (v0.1.6) \u003csup\u003e48\u003c/sup\u003e. HTML report was generated using the following R packages: rmarkdown (v2.29) \u003csup\u003e49\u003c/sup\u003e and flexdashboard (v0.6.2) \u003csup\u003e50\u003c/sup\u003e. For each sample assessed, three independent experimental replicates were sequenced.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSoftware development and data analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe developed Di-GRAPH, a user-friendly software that automates i) the analysis of gene conversion efficiency, extent and symmetry at an HO-induced break; ii) the assessment of repair pathway choice an HO-associated mutagenic patterns; iii) genome-wide coverage profiling to evaluate genome instability in response to HO induction; and iv) genomic discordance analysis to identify and map genome-wide chromosomal rearrangements occurring during HO repair. Di-GRAPH is open source and publicly available at https://github.com/acb-lab/Di_GRAPH, together with instruction for its installation and usage. For this study, Di-GRAPH was run as follows:\u003c/p\u003e\n\u003cp\u003eDi-GRAPH.sh -b $CONDA_PREFIX/files/BLAST -c $CONDA_PREFIX/files/Categories -g $CONDA_PREFIX/files/RG -i working_directory/ -r $CONDA_PREFIX/files/Report_files\u003c/p\u003e\n\u003cp\u003eThe following sections of the methods describe in detail the different steps and software that compose Di-GRAPH pipeline.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCoverage and Polymorphism Analysis. \u003c/em\u003eThe sequencing reads for each sample and experimental replicate, FASTQ files were concatenated and reads were sequentially trimmed at fixed-length fragments using cutadapt \u003csup\u003e51\u003c/sup\u003e (v5.1) to generate two datasets: 1) 75nt fragments for coverage profiling, and 2) 18nt fragments for polymorphism quantification. Quality filtering was performed using fastp (v0.25.0) \u003csup\u003e52\u003c/sup\u003e (Q30 threshold) and reads were then aligned to \u003cem\u003eS. cerevisiae\u003c/em\u003e PMV reference genome (Supplementary Data, RG_PMV_v9.fasta) using Bowtie (v1.3.1) \u003csup\u003e25\u003c/sup\u003e (75nt: -m 1 -v 0; 18nt -k 1 -v 0). SAM files were converted to sorted BAM format with SAMtools (v1.22) \u003csup\u003e53\u003c/sup\u003e, and genome coverage profiles were computed using bamCoverage (RPGC normalization, Deeptools v3.5.6 \u003csup\u003e54\u003c/sup\u003e). BedGraph files were sorted using Bedtools, and per-base coverage values were extracted for chromosome III and V. Coverage profiles for whole-chromosome views were generated by binning data into 100 bp windows. Finally, combined datasets for each chromosome and condition were averaged, and coverage heatmaps and polymorphism plots were generated. \u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWGS coverage profiling. \u003c/em\u003eCoverage-based enrichment across genomic features categories was quantified using Di-GRAPH. For each strain and time-point (T0, NSS, SS and UD), coverage profiles generated from 75nt mapping reads (see \u0026ldquo;Coverage and Polymorphism Analysis\u0026rdquo; section) were cross-examined against predefined genomic annotations. Genomic features sets were provided as category-specific BED/TSV files. For each experiment (E1-E3), per-chromosome coverage vectors were reconstructed by mapping read-depth values to the full coordinate span of each chromosome. Then, per-base coverage for each annotated feature region was calculated, aggregating values within each feature class. For each category and chromosome, the summed coverage of NSS, SS and UD samples was normalized to its corresponding T0 value to produce experiment-specific ratios. Ratios were then averaged across biological replicates, and standard deviations were calculated. Category enrichment was visualized as barplots with error bars and violin/boxplot distributions by using ggplot2 under R v4.4.3 environment. Coverage distribution along the entire genome was displayed using genome-wide barplots with positional segmentation, including annotated boundaries between major genomic categories. Threshold-based enrichment metrics were computed to quantify the proportion of sites exceeding ratio \u0026gt;1.2 or falling below \u0026lt;0.8 within each category, as well as their associated variability. Chromosome-specific enrichment analysis were further performed by sorting feature data by chromosome (I-XVI), and calculating the percentage of genomic features above or below coverage thresholds per chromosome.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHO-associated mutagenic profile. \u003c/em\u003ePair-ended 150nt sequencing reads were quality-filtered (Q\u0026sup3;30) using fastp and subsequently aligned to the \u003cem\u003eS. cerevisiae \u003c/em\u003ePMV chromosome III reference genome (Supplementary Data, RG_PMV_v9_CHRIII.fasta) using BWA-MEM (v0.7.19) \u003csup\u003e27\u003c/sup\u003e. Reads mapping specifically to the HO cleavage site (position 200753 on chromosome III) were extracted with SAMtools. Aligned reads were further processed to isolate those containing specific nucleotides at positions 200689 and 200753 (see figure 4A for details) using pysam (v0.23.1) \u003csup\u003e53\u003c/sup\u003e. Insertions, deletions and base substitutions were quantified with IGVtools (v2.17.3) \u003csup\u003e55\u003c/sup\u003e and BCFtools \u003csup\u003e53\u003c/sup\u003e, (v1.22), and variant frequencies were normalized to total read counts. Repair pathway contributions were inferred by calculating the proportion of reads corresponding to NHEJ (A\u003csup\u003e200689\u003c/sup\u003e/C\u003csup\u003e200753\u003c/sup\u003e), short-range HR(A\u003csup\u003e200689\u003c/sup\u003e/T\u003csup\u003e200753\u003c/sup\u003e) and long HR (G\u003csup\u003e200689\u003c/sup\u003e/T\u003csup\u003e200753\u003c/sup\u003e). Data aggregation, quantification analysis, and visualization of the repair pathway distributions and mutagenic rates calculation and data visualization were performed using the R packages specified above. DNA sequence logos were performed using ggseqlogo R package (v0.2) \u003csup\u003e56\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eDiscordant-reads-based analysis of Gene Conversion.\u003c/em\u003e For each sample and experimental replicate, quality-controlled reads were sequentially trimmed into 75nt and 18nt fixed-length fragment using cutadapt to generate two datasets: 1) R1-R2 75nt paired fragments for gene conversion analysis, and 2) 18nt paired fragments within the same read (~ 60nt spaced) for polymorphism analysis. Quality filtering was performed using fastp (Q30 threshold) and datasets were aligned to \u003cem\u003eS. cerevisiae\u003c/em\u003e reference genome twice: (i) using Bowtie (-m 1 -v 0), and (ii) using Bowtie2 \u003csup\u003e57\u003c/sup\u003e under restrictive scoring conditions to emulate -m 1 -v 0 behavior (--no-1mm-upfront --score-min C,0,0 -N 0 --end-to-end --fr). Multi-mapping reads from Bowtie were identified and discarded from the inter-chromosomal discordant read pairs of Bowtie2 dataset. Reads spanning the \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea\u0026acute; \u003cem\u003eloci\u003c/em\u003e were extracted based on coordinate filtering. SAM files were converted to sorted BAM format with SAMtools, and \u003cem\u003eMAT\u003c/em\u003ea/\u003cem\u003eMAT\u003c/em\u003ea\u0026rsquo; coverage profiles were computed using SAMtools and bamCoverage. BedGraph files were sorted using Bedtools, and per-base coverage values were extracted. The resulting chromosome-resolved, base-level coverage tables for each replicate were merged, positions were aligned relative to the HO cut site, and statistics were computed. Densitometry plots were performed using ggdensity R package (v1.0.0) \u003csup\u003e58\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eGenome-wide discordance assessment. \u003c/em\u003eGlobal genome recombination analysis was performed using Di-GRAPH. 75nt inter-chromosomal discordant read pairs from the dataset described in the previous section were used. Per-base discordance coverage was intersected with each annotating genomic feature to define inter-chromosomal discordant pairs. To ensure discordance reliability, a BLAST alignment was performed to guarantee no concordant re-alignment (see figure 7A for details). Artifactual inter-chromosomal discordant pairs arising from variations between the reference genome and the PMV genetic background were removed. Disc-normalization was applied to calculate discordant frequencies at each genomic region. The number of the discordant reads between the \u003cem\u003eADE3\u003c/em\u003e gene and the \u003cem\u003eVector-SSB1-HO-pGAL1-10\u003c/em\u003e sequence was calculated for each timepoint and replicate. A normalization factor was computed using the discordant values obtained for the reference strain and each mutant tested. Frequencies were normalized using these correction factors allowing comparison between timepoints, replicates and strains. Radars plots were represented using fmsb R package (v0.7.6) \u003csup\u003e59\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eInter-chromosomal discordant network\u003c/em\u003e. To generate inter-chromosomal discordant networks, we set genomic features with high discordance frequency (\u0026sup3;6%) as nodes and plotted all genomic features they showed a rearrangement with. Both \u003cem\u003eloci\u003c/em\u003e were connected to generate the distance network using igraph R package (v2.1.4) \u003csup\u003e60\u003c/sup\u003e. To quantify the interactions observed in the network, we calculated the percentage of each genomic category combination for each timepoint (NSS, SS and UD).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCoverage and discordant read data correlation. \u003c/em\u003eTo evaluate the nature of recombination events, we compared coverage levels and discordancy frequency for each genomic feature with high discordance frequency (\u0026sup3;6%). Genomic features showing both high coverage levels and discordancy frequency involve genomic amplification arising from recombination.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eHomology-directed recombination. \u003c/em\u003eTo assess whether the global rearrangements observed following an HO induction were dependent on homology-directed recombination, we calculated the sequence identity of every genomic feature exhibiting high discordancy frequency (\u0026sup3;6%) using EMBOSS (v6.6.0) \u003csup\u003e61\u003c/sup\u003e (needleall -gapopen 10 -gapextend 0.5). For each genomic feature, identity percentages were computed for two different genomic datasets: (i) all genomic features annotated for PMV strain, to calculate random identity; and (ii) genomic features from inter-chromosomal discordant pairs, to calculate discordant pairs identity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistics and reproducibility\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data presented in this manuscript was generated from three independent biological replicates. For each analysis, mean values and standard deviations were calculated using the Di-GRAPH computational pipeline, which was applied consistently across all datasets to ensure statistical robustness and reproducibility.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGenomic datasets and are deposited in the Sequence Read Archive (SRA) database under accession code PRJNA1380437 (https://dataview.ncbi.nlm.nih.gov/object/PRJNA1380437?reviewer=dis753khgl1a9urvk7v1qjlrn6).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Di-GRAPH code is available on GitHub, together with instructions for installation and usage, at https://github.com/acb-lab/Di_GRAPH. To ensure reproducibility, a test dataset has been publicly released to Zenodo (https://doi.org/10.5281/zenodo.18470248) and instructions for Di-GRAPH analysis are available at https://github.com/acb-lab/Di_GRAPH.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank J. Haber for kindly providing the background strains JKM139, from which the PMV strain was derived. We thank members of our laboratory for helpful discussions and comments on the manuscript. This work was supported by the projects PID2021-125290NB-I00 and PID2024-160235NB-I00, funded by the MCIN/AEI/10.13039/501100011033/ and by the \u0026ldquo;FEDER, Una manera de hacer Europa\u0026rdquo;, awarded to A. C-B. This work was also supported by the institutional grant \u0026ldquo;Deep-MaX-2024 (CSIC)\u0026rdquo;. L.I. and A.C. were recipients of a predoctoral fellowship from the \u0026ldquo;Junta de Castilla y Le\u0026oacute;n\u0026rdquo;. L.I. was the recipient of a \u0026ldquo;Jae-INTRO\u0026rdquo; grant from the CSIC. C.D. and L.P. were recipients of a predoctoral fellowship from the \u0026ldquo;Universidad de Salamanca\u0026rdquo;. C.D. was recipient of a predoctoral fellowship form the FPU \u0026ldquo;Formaci\u0026oacute;n de Profesorado Universitario\u0026rdquo; from the \u0026ldquo;Ministerio de Ciencia, Innovaci\u0026oacute;n y Universidades\u0026rdquo;. N.M.Q. was appointed under the \u0026ldquo;Momentum\u0026rdquo; Intramural CSIC program, through the project MMT24-IBFG-01 awarded to A.C-B.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eL.I. performed the experiments and carried out the processing and analysis of the sequencing data. A.C., C.D., L.P., and M.T.V. collaborated in the experimental work. N.M.Q. provided supervision for the design, development and optimization of the bioinformatic workflow and software development. A.C-B. conceived and designed the study, supervised experimental activities, contributed to data analysis and interpretation, and led the writing of the manuscript. All authors discussed the results and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMimitou, E. P. \u0026amp; Symington, L. S. DNA end resection: many nucleases make light work. \u003cem\u003eDNA repair\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 983-995, doi:10.1016/j.dnarep.2009.04.017 (2009).\u003c/li\u003e\n\u003cli\u003eGobbini, E., Cassani, C., Villa, M., Bonetti, D. \u0026amp; Longhese, M. P. Functions and regulation of the MRX complex at DNA double-strand breaks. \u003cem\u003eMicrob Cell\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 329-337, doi:10.15698/mic2016.08.517 (2016).\u003c/li\u003e\n\u003cli\u003eGarcia, V., Phelps, S. E., Gray, S. \u0026amp; Neale, M. J. Bidirectional resection of DNA double-strand breaks by Mre11 and Exo1. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e479\u003c/strong\u003e, 241-244, doi:10.1038/nature10515 (2011).\u003c/li\u003e\n\u003cli\u003eMimitou, E. P. \u0026amp; Symington, L. S. Sae2, Exo1 and Sgs1 collaborate in DNA double-strand break processing. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e455\u003c/strong\u003e, 770-774, doi:10.1038/nature07312 (2008).\u003c/li\u003e\n\u003cli\u003eZhu, Z., Chung, W. H., Shim, E. Y., Lee, S. E. \u0026amp; Ira, G. Sgs1 helicase and two nucleases Dna2 and Exo1 resect DNA double-strand break ends. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e134\u003c/strong\u003e, 981-994, doi:10.1016/j.cell.2008.08.037 (2008).\u003c/li\u003e\n\u003cli\u003eDumont, A.\u003cem\u003e et al.\u003c/em\u003e Mechanism of homology search expansion during recombinational DNA break repair in Saccharomyces cerevisiae. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e84\u003c/strong\u003e, 3237-3253 e3236, doi:10.1016/j.molcel.2024.08.003 (2024).\u003c/li\u003e\n\u003cli\u003eChen, H., Lisby, M. \u0026amp; Symington, L. S. RPA coordinates DNA end resection and prevents formation of DNA hairpins. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 589-600, doi:10.1016/j.molcel.2013.04.032 (2013).\u003c/li\u003e\n\u003cli\u003eWang, X. \u0026amp; Haber, J. E. Role of Saccharomyces single-stranded DNA-binding protein RPA in the strand invasion step of double-strand break repair. \u003cem\u003ePLoS biology\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, E21, doi:10.1371/journal.pbio.0020021 (2004).\u003c/li\u003e\n\u003cli\u003eKrejci, L.\u003cem\u003e et al.\u003c/em\u003e DNA helicase Srs2 disrupts the Rad51 presynaptic filament. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e423\u003c/strong\u003e, 305-309, doi:10.1038/nature01577 (2003).\u003c/li\u003e\n\u003cli\u003eVeaute, X.\u003cem\u003e et al.\u003c/em\u003e The Srs2 helicase prevents recombination by disrupting Rad51 nucleoprotein filaments. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e423\u003c/strong\u003e, 309-312, doi:10.1038/nature01585 (2003).\u003c/li\u003e\n\u003cli\u003ePiazza, A.\u003cem\u003e et al.\u003c/em\u003e Dynamic Processing of Displacement Loops during Recombinational DNA Repair. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 1255-1266 e1254, doi:10.1016/j.molcel.2019.01.005 (2019).\u003c/li\u003e\n\u003cli\u003ePiazza, A. \u0026amp; Heyer, W. D. Homologous Recombination and the Formation of Complex Genomic Rearrangements. \u003cem\u003eTrends in cell biology\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 135-149, doi:10.1016/j.tcb.2018.10.006 (2019).\u003c/li\u003e\n\u003cli\u003eBantele, S.\u003cem\u003e et al.\u003c/em\u003e Repair of DNA double-strand breaks leaves heritable impairment to genome function. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e390\u003c/strong\u003e, eadk6662, doi:10.1126/science.adk6662 (2025).\u003c/li\u003e\n\u003cli\u003eSasaki, M., Tischfield, S. E., van Overbeek, M. \u0026amp; Keeney, S. Meiotic recombination initiation in and around retrotransposable elements in Saccharomyces cerevisiae. \u003cem\u003ePLoS genetics\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e1003732, doi:10.1371/journal.pgen.1003732 (2013).\u003c/li\u003e\n\u003cli\u003eGilat, N.\u003cem\u003e et al.\u003c/em\u003e From single-molecule to genome-wide mapping of DNA lesions: repair-assisted damage detection sequencing. \u003cem\u003eBiophys Rep (N Y)\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, None, doi:10.1016/j.bpr.2021.100017 (2021).\u003c/li\u003e\n\u003cli\u003eHussmann, J. A.\u003cem\u003e et al.\u003c/em\u003e Mapping the genetic landscape of DNA double-strand break repair. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e184\u003c/strong\u003e, 5653-5669 e5625, doi:10.1016/j.cell.2021.10.002 (2021).\u003c/li\u003e\n\u003cli\u003ede Alba, E. L.\u003cem\u003e et al.\u003c/em\u003e A comprehensive genetic catalog of human double-strand break repair. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e390\u003c/strong\u003e, eadr5048, doi:10.1126/science.adr5048 (2025).\u003c/li\u003e\n\u003cli\u003eSriramachandran, A. M.\u003cem\u003e et al.\u003c/em\u003e Genome-wide Nucleotide-Resolution Mapping of DNA Replication Patterns, Single-Strand Breaks, and Lesions by GLOE-Seq. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e78\u003c/strong\u003e, 975-985 e977, doi:10.1016/j.molcel.2020.03.027 (2020).\u003c/li\u003e\n\u003cli\u003eCampos, A.\u003cem\u003e et al.\u003c/em\u003e Cdc14 phosphatase counteracts Cdk-dependent Dna2 phosphorylation to inhibit resection during recombinational DNA repair. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 2738, doi:10.1038/s41467-023-38417-5 (2023).\u003c/li\u003e\n\u003cli\u003ePfeifer, G. P. \u0026amp; Jin, S. G. Methods and applications of genome-wide profiling of DNA damage and rare mutations. \u003cem\u003eNat Rev Genet\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 846-863, doi:10.1038/s41576-024-00748-4 (2024).\u003c/li\u003e\n\u003cli\u003eTeng, Y.\u003cem\u003e et al.\u003c/em\u003e A novel method for the genome-wide high resolution analysis of DNA damage. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e39\u003c/strong\u003e, e10, doi:10.1093/nar/gkq1036 (2011).\u003c/li\u003e\n\u003cli\u003eZhang, K., Zheng, D. Q., Sui, Y., Qi, L. \u0026amp; Petes, T. D. Genome-wide analysis of genomic alterations induced by oxidative DNA damage in yeast. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 3521-3535, doi:10.1093/nar/gkz027 (2019).\u003c/li\u003e\n\u003cli\u003eRybin, M. J.\u003cem\u003e et al.\u003c/em\u003e Emerging Technologies for Genome-Wide Profiling of DNA Breakage. \u003cem\u003eFront Genet\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 610386, doi:10.3389/fgene.2020.610386 (2020).\u003c/li\u003e\n\u003cli\u003eRamos, F.\u003cem\u003e et al.\u003c/em\u003e Genome-wide sequencing analysis of Sgs1, Exo1, Rad51, and Srs2 in DNA repair by homologous recombination. \u003cem\u003eCell Rep\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e, 110201, doi:10.1016/j.celrep.2021.110201 (2022).\u003c/li\u003e\n\u003cli\u003eLangmead, B., Trapnell, C., Pop, M. \u0026amp; Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. \u003cem\u003eGenome biology\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, R25, doi:10.1186/gb-2009-10-3-r25 (2009).\u003c/li\u003e\n\u003cli\u003eLo, Y. C.\u003cem\u003e et al.\u003c/em\u003e Sgs1 regulates gene conversion tract lengths and crossovers independently of its helicase activity. \u003cem\u003eMol Cell Biol\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 4086-4094, doi:10.1128/MCB.00136-06 (2006).\u003c/li\u003e\n\u003cli\u003eLi, H. \u0026amp; Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 1754-1760, doi:10.1093/bioinformatics/btp324 (2009).\u003c/li\u003e\n\u003cli\u003eBradshaw, V. A. \u0026amp; McEntee, K. DNA damage activates transcription and transposition of yeast Ty retrotransposons. \u003cem\u003eMol Gen Genet\u003c/em\u003e \u003cstrong\u003e218\u003c/strong\u003e, 465-474, doi:10.1007/BF00332411 (1989).\u003c/li\u003e\n\u003cli\u003eFarkash, E. A., Kao, G. D., Horman, S. R. \u0026amp; Prak, E. T. Gamma radiation increases endonuclease-dependent L1 retrotransposition in a cultured cell assay. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, 1196-1204, doi:10.1093/nar/gkj522 (2006).\u003c/li\u003e\n\u003cli\u003eStaleva Staleva, L. \u0026amp; Venkov, P. Activation of Ty transposition by mutagens. \u003cem\u003eMutation research\u003c/em\u003e \u003cstrong\u003e474\u003c/strong\u003e, 93-103, doi:10.1016/s0027-5107(00)00165-2 (2001).\u003c/li\u003e\n\u003cli\u003eMitchel, K., Lehner, K. \u0026amp; Jinks-Robertson, S. Heteroduplex DNA position defines the roles of the Sgs1, Srs2, and Mph1 helicases in promoting distinct recombination outcomes. \u003cem\u003ePLoS genetics\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e1003340, doi:10.1371/journal.pgen.1003340 (2013).\u003c/li\u003e\n\u003cli\u003ePohl, T. J. \u0026amp; Nickoloff, J. A. Rad51-independent interchromosomal double-strand break repair by gene conversion requires Rad52 but not Rad55, Rad57, or Dmc1. \u003cem\u003eMol Cell Biol\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 897-906, doi:10.1128/MCB.00524-07 (2008).\u003c/li\u003e\n\u003cli\u003eSo, A.\u003cem\u003e et al.\u003c/em\u003e RAD51 protects against nonconservative DNA double-strand break repair through a nonenzymatic function. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 2651-2666, doi:10.1093/nar/gkac073 (2022).\u003c/li\u003e\n\u003cli\u003eArgueso, J. L.\u003cem\u003e et al.\u003c/em\u003e Double-strand breaks associated with repetitive DNA can reshape the genome. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e105\u003c/strong\u003e, 11845-11850, doi:10.1073/pnas.0804529105 (2008).\u003c/li\u003e\n\u003cli\u003ePutnam, C. D., Hayes, T. K. \u0026amp; Kolodner, R. D. Specific pathways prevent duplication-mediated genome rearrangements. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e460\u003c/strong\u003e, 984-989, doi:10.1038/nature08217 (2009).\u003c/li\u003e\n\u003cli\u003eGal, Z.\u003cem\u003e et al.\u003c/em\u003e Hyper-recombination in ribosomal DNA is driven by long-range resection-independent RAD51 accumulation. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 7797, doi:10.1038/s41467-024-52189-6 (2024).\u003c/li\u003e\n\u003cli\u003eQiu, Y.\u003cem\u003e et al.\u003c/em\u003e Srs2 prevents Rad51 filament formation by repetitive motion on DNA. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 2281, doi:10.1038/ncomms3281 (2013).\u003c/li\u003e\n\u003cli\u003eLiu, B.\u003cem\u003e et al.\u003c/em\u003e Structural variation discovery in the cancer genome using next generation sequencing: computational solutions and perspectives. \u003cem\u003eOncotarget\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 5477-5489, doi:10.18632/oncotarget.3491 (2015).\u003c/li\u003e\n\u003cli\u003eJanke, C.\u003cem\u003e et al.\u003c/em\u003e A versatile toolbox for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes. \u003cem\u003eYeast\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 947-962, doi:10.1002/yea.1142 (2004).\u003c/li\u003e\n\u003cli\u003eGoldstein, A. L. \u0026amp; McCusker, J. H. Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. \u003cem\u003eYeast\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1541-1553, doi:10.1002/(SICI)1097-0061(199910)15:14\u0026lt;1541::AID-YEA476\u0026gt;3.0.CO;2-K (1999).\u003c/li\u003e\n\u003cli\u003eQuinlan, A. R. \u0026amp; Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 841-842, doi:10.1093/bioinformatics/btq033 (2010).\u003c/li\u003e\n\u003cli\u003eFreese, N. H., Norris, D. C. \u0026amp; Loraine, A. E. Integrated genome browser: visual analytics platform for genomics. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 2089-2095, doi:10.1093/bioinformatics/btw069 (2016).\u003c/li\u003e\n\u003cli\u003eWickham, H., Fran\u0026ccedil;ois, R., Henry, L., M\u0026uuml;ller, K. \u0026amp; Vaughan, D. dplyr: a grammar of data manipulation. (2023).\u003c/li\u003e\n\u003cli\u003eWickham, H., Vaughan, D. \u0026amp; Girlich, M. . Tidyr: tidy messy data. . (2024).\u003c/li\u003e\n\u003cli\u003eWickham, H. Stringr: simple, consistent wrappers for common string operations. . (2025).\u003c/li\u003e\n\u003cli\u003eWickham, H. ggplot2: Elegant Graphics for Data Analysis. . \u003cem\u003eJournal of the Royal Statistical Society Series A: Statistics in Society\u003c/em\u003e \u003cstrong\u003e174(1), 245\u0026ndash;246\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eWickham, H., \u0026amp; Henry, L. . purrr: Functional Programming Tools. (2025).\u003c/li\u003e\n\u003cli\u003eXu, S.\u003cem\u003e et al.\u003c/em\u003e Use ggbreak to Effectively Utilize Plotting Space to Deal With Large Datasets and Outliers. \u003cem\u003eFront Genet\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 774846, doi:10.3389/fgene.2021.774846 (2021).\u003c/li\u003e\n\u003cli\u003eAllaire, J., Xie, Y., Dervieux, C., McPherson, J., Luraschi, J., Ushey, K., Atkins, A., Wickham, H., Cheng, J., Chang, W., \u0026amp; Iannone, R. . rmarkdown: Dynamic Documents for R. (2024).\u003c/li\u003e\n\u003cli\u003eGarrick, A.-B., Sievert, C., Iannone, R., \u0026amp; Allaire, J. . flexdashboard: R Markdown Format for Flexible Dashboards. (2025).\u003c/li\u003e\n\u003cli\u003eMartin, M. Cutadapt Removes Adapter Sequences From High‐Throughput Sequencing Reads. \u003cem\u003eEMBnet.journal\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, doi:https://doi.org/10.14806/ej.17.1.200 (2011).\u003c/li\u003e\n\u003cli\u003eChen, S., Zhou, Y., Chen, Y. \u0026amp; Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, i884-i890, doi:10.1093/bioinformatics/bty560 (2018).\u003c/li\u003e\n\u003cli\u003eLi, H.\u003cem\u003e et al.\u003c/em\u003e The Sequence Alignment/Map format and SAMtools. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 2078-2079, doi:10.1093/bioinformatics/btp352 (2009).\u003c/li\u003e\n\u003cli\u003eRamirez, F.\u003cem\u003e et al.\u003c/em\u003e deepTools2: a next generation web server for deep-sequencing data analysis. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, W160-165, doi:10.1093/nar/gkw257 (2016).\u003c/li\u003e\n\u003cli\u003eRobinson, J. T.\u003cem\u003e et al.\u003c/em\u003e Integrative genomics viewer. \u003cem\u003eNature biotechnology\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 24-26, doi:10.1038/nbt.1754 (2011).\u003c/li\u003e\n\u003cli\u003eWagih, O. ggseqlogo: a versatile R package for drawing sequence logos. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 3645-3647, doi:10.1093/bioinformatics/btx469 (2017).\u003c/li\u003e\n\u003cli\u003eLangmead, B. \u0026amp; Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. \u003cem\u003eNature methods\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 357-359, doi:10.1038/nmeth.1923 (2012).\u003c/li\u003e\n\u003cli\u003eOtto, J. K., D. ggdensity: Interpretable Bivariate Density Visualization with ggplot2. \u003cem\u003ehttps://jamesotto852.github.io/ggdensity/\u003c/em\u003e (2023).\u003c/li\u003e\n\u003cli\u003eNakazawa, M. fmsb: Functions for Medical Statistics Book with some Demographic Data. (2010).\u003c/li\u003e\n\u003cli\u003eCs\u0026aacute;rdi, G., \u0026amp; Nepusz, T. The igraph software package for complex network research. (2006).\u003c/li\u003e\n\u003cli\u003eRice, P., Longden, I. \u0026amp; Bleasby, A. EMBOSS: the European Molecular Biology Open Software Suite. \u003cem\u003eTrends Genet\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 276-277, doi:10.1016/s0168-9525(00)02024-2 (2000).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Double-strand break, DNA repair, homologous recombination, genome-wide stability","lastPublishedDoi":"10.21203/rs.3.rs-8989286/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8989286/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Homologous recombination (HR) repairs DNA double-strand breaks (DSBs) with high fidelity, yet how defects at distinct steps of the recombination pathway influence gene conversion (GC) outcomes and contribute to genome instability remains poorly defined. Here, we develop Di-GRAPH, a computational framework that integrates coverage profiling and discordant read mapping to systematically analyze mutational signatures, repair pathway choice, and GC architecture. At a genomic scale, Di-GRAPH detects, quantifies and classifies DNA damage-dependent gross chromosomal rearrangements across diverse genomic contexts. Applying Di-GRAPH to mutants defective in resection (exo1Δ), recombination intermediate processing (sgs1Δ, srs2Δ), and strand invasion (rad51Δ), we identify stage-specific GC patterns, mutation-specific repair signatures and distinct patterns of genome-wide rearrangements. Loss of Exo1 minimally affects GC formation but severely compromises genome stability in undamaged cells. Absence of Sgs1 and Srs2 leads to extended and discontinued GCs, accompanied by widespread ectopic recombination and retrotransposon mobilization following DSB repair. Rad51 deficiency results in severely impaired GC outcomes, strong reliance on non-homologous end joining, elevated break-proximal mutagenesis, and extensive genome-wide rearrangements in both DSB survivors and undamaged cells. Coverage and discordant-read signals were strongly correlated across most genomic features, indicating that recombination-associated translocations are frequently coupled to local DNA amplification. Comparative sequence identity analyses reveal distinct homology requirements, with ORFs showing the highest homology dependence, whereas repetitive elements recombine under more relaxed constraints. Together, this integrated analysis reveals that defects at discrete steps of the HR pathway impose distinct and predictable genome-wide instability signatures, providing a mechanistic framework for understanding how recombination fidelity is enforced across the DNA repair process.","manuscriptTitle":"Stage-specific Recombination Factors Differentially Regulate Double-strand Break Repair Fidelity and Influence Genome Stability","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 07:18:05","doi":"10.21203/rs.3.rs-8989286/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"42038bdb-37f1-48b4-9137-3e0025080017","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Reject after peer review","date":"2026-05-05T17:10:04+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-04T19:06:58+00:00","index":2,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":65043965,"name":"Biological sciences/Molecular biology/DNA damage and repair/Double-strand DNA breaks"},{"id":65043966,"name":"Biological sciences/Molecular biology/DNA damage and repair/Homologous recombination"}],"tags":[],"updatedAt":"2026-05-05T17:15:18+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 07:18:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8989286","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8989286","identity":"rs-8989286","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}