A high-resolution meiotic crossover map from single-nucleus ATAC-seq reveals insights into the recombination landscape in mammalian sperm

14 Meiotic crossovers promote correct chromosome segregation and the shuffling of genetic 15 diversity. However, the measurement of crossovers remains challenging, impeding our ability 16 to decipher the molecular mechanisms that are necessary for their formation and regulation . 17 Here we demonstrate a novel repurposing of the single -nucleus Assay for Transposase 18 Accessible Chromatin with sequencing (snATAC-seq) as a simple and high-throughput method 19 to identify and characterise meiotic crossovers from sperm nuclei . We first validate the 20 feasibility of obtaining genome-wide coverage from snATAC-seq by using ATAC-seq on bulk 21 haploid mouse sperm, ensuring adequate variant detection for haplotyping. Subsequently, we 22 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 2 adapt droplet-based snATAC-seq for crossover detection, revealing over 25,000 crossovers in 23 F1 hybrid mice. Comparison between wildtype and a hyper -recombinogenic Fancm-deficient 24 mutant mouse model confirmed an increase in crossover rates in this genotype, however a 25 distribution which was unchanged. We also find that regions with the highest rate of crossover 26 formation are enriched for DMC1 and PRDM9, with a subset that is further enriched for DMC1 27 in Fancm-deficient mice. Our findings demonstrate the utility of snATAC-seq as a robust and 28 scalable tool for high-throughput crossover detection, offering insights into meiotic crossover 29 dynamics and elucidating the underlying molecular mechanisms. It is possible that the research 30 presented here with snATAC-seq of haploid sperm could be extended into fertility -related 31 diagnostics. 32 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 3

33 Meiosis is a specialised cell division which is necessary for the formation of haploid gametes 34 in sexually reproducing species. Meiosis halves ploidy through one round of DNA replication 35 being followed by two consecutive rounds of chromosome division. The first round of 36 chromosome segregation (meiosis I) is a reductional division, where homologues are pulled to 37 opposite poles. The second round of chromosome segregation (meiosis II) is an equational 38 division where sister chromatids are pulled to opposite poles (Hunter, 2015). At the onset of 39 meiosis, hundreds of DNA double -strand breaks (DSBs) are formed (Lam & Keeney, 2015) . 40 Homologous recombination is used to repair the DSBs as either crossovers (COs) or non -41 crossovers (NCO) (Gray & Cohen, 2016; Hunter, 2015; Szostak et al., 1983) . Meiotic 42 crossovers are large reciprocal exchanges between homologous chromosomes. For 43 chromosomes to segregate correctly at the first meiotic division, all homologue pairs must have 44 at least one “obligate” crossover. Loss of the obligate crossover leads to aneuploid gametes and 45 a reduction in fertility (Hunter, 2015) . Aneuploid gametes used in fertilisation can lead to 46 aneuploid karyotypes, such as trisomy 21 in humans (Hassold & Sherman, 2000; Sherman et 47 al., 1991; Warren et al., 1987) . The obligate crossover is linked to a phenomenon known as 48 crossover interference, where the occurrence of one chromosome reduces the probability of 49 another crossover on the same chromosome in the same meiosis in a distance dependent manner 50 (Berchowitz & Copenhaver, 2010; Hunter, 2015). 51 52 Assaying for genome-wide crossovers remains challenging, despite more than a century since 53 the discovery of meiotic crossing over in drosophila and how it informs on physical 54 chromosome structure and enables genetic map construction (Morgan, 1910; Sturtevant, 1913). 55 These detection challenges mostly relate to the need for generating and genotyping pedigrees 56 that require at least three generations and many offspring. 57 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 4 58 Reporter assays continue to play an essential role in deciphering the mechanisms of double-59 strand break repair and crossover formation. Some examples include DSB hotspot analysis, 60 linked phenotypic markers and tetrad analysis across an array of species (Cole et al., 2014; 61 Fogel & Mortimer, 1971; Francis et al., 2007; McClintock, 1931; Morgan, 1910; Schwacha & 62 Kleckner, 1994; Yelina et al., 2013). These assays can allow quantitative analysis of a variety 63 of parameters at a single locus or adjacent loci e.g. formation of DSBs , crossovers, non-64 crossovers and, repair on sister chromatids. A strength of their methods is the reproducibility 65 once established and potential for comparisons across experimental conditions . On the other 66 hand, cytological analysis of DSB and crossover formation inform on chromosome-scale and 67 genome-wide events. However, these methods, while providing rich and important data tend 68 to be more labour intense, particularly for quantification of events being scored. 69 Whole-genome single-gamete sequencing provides the opportunity to combine the advantages 70 of different styles of assays for crossover detection and quantification . Single-gamete 71 sequencing for genome-wide crossover analysis can reduce the number of generations required 72 to access the recombined products of meiosis. Interest in understanding crossover regulation 73 has driven the adaptation of various technologies to sequence individual gametes (Lu et al., 74 2012; Bell et al., 2019; Hinch et al., 2019; Luo et al., 2019; Sun et al., 2019; J. Campoy et al., 75 2020; Tsui, et al., 2023; Xie et al., 2023) and the development of methods for analysis of the 76 data (Carioscia et al., 2022; Lyu et al., 2022) . Notably, droplet -based gamete sequencing 77

are a powerful tool for investigating genome -wide crossover formation, because 78 despite the low level of coverage typically obtained per cell, the coverage is sufficient to 79 determine haplotypes (J. Campoy et al., 2020; Leung et al., 2021; Lyu et al., 2022; Tsui, et al., 80 2023). However, despite advances in single gamete sequencing approaches, specialised kits 81 used in previous studies have been discontinued which results in established methods becoming 82 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 5 unavailable. We therefore sought to establish a method that we consider should remain 83 accessible for longer or could be re-established with in-house reagents in the case of product 84 discontinuation. The Assay for Transposase Accessible Chromatin using sequencing (ATAC-85 seq) was developed for profiling chromatin accessibility at a genome-wide scale (Buenrostro 86 et al., 2013) . This method of DNA library preparation could be adapted for single-nucleus 87 sperm sequencing for crossover detection applications, which w e predict as it can generate 88 reads genome-wide – in some cell lines and tissues – and has already been used for single cell 89 sequencing assays (Cusanovich et al., 2015). Further, open-source methods for Tn5 expression 90 and purification are well established which can facilitate in -house development of ATAC -91 based methods. However, it is less clear the extent to which gametes can be sequenced with 92 ATAC-based methods given the differ ent chromatin structure – which is protamine rich – in 93 mammalian sperm compared to somatic tissue. 94 We demonstrate here that snATAC-seq libraries can be used to sequence genomic DNA for 95 crossover quantification and analyses. Using hybrid mice, from a cross between the two strains 96 C57BL/6J and FVB, we isolated haploid sperm and prepared libraries using modified 97 snATAC-seq methods. We obtained high quality libraries which had ample coverage for 98 variant calling, phasing and crossover detection . Further the number of high -quality cells 99 recovered was far greater than other methods that we have used. Using an established hyper-100 crossover mouse model, we could also demonstrate that this robust method can detect changes 101 in crossover frequencies, opening scope for other applications in research and diagnostics. 102 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 6

103 Bulk ATAC-sequencing libraries using sperm nuclei produce genome-wide coverage 104 ATAC-seq assays use the enzymatic activity of Tn5 to generate DNA libraries for high 105 throughput sequencing. Tn5 simultaneously cuts accessible regions of double stranded DNA 106 while ligating sequencing adapters compatible with high-throughput whole-genome 107 sequencing platforms (Adey et al., 2010; Buenrostro et al., 2013; Goryshin & Reznikoff, 1998). 108 However, sperm chromatin is particularly condensed and protamine rich (Hammoud et al., 109 2009; Oliva, 2006) , and it was unclear if libraries could generate read coverage suitable for 110 haplotyping and CO measurement. Therefore, we first sought to test if sufficient coverage of 111 sperm DNA could be obtained in bulk ATAC -seq samples to justify using ATAC -seq with 112 single-sperm nuclei. 113 We first performed ATAC -sequencing by isolating 50,000 bulk-sorted haploid sperm nuclei 114 extracted from F1 (C57BL/6 J x FVB/N) hybrid mice , using our SSNIP -seq protocol 115 (Novakovic et al., 2022) for high quality sperm nuclei isolation . We expressed and purified 116 Tn5 (Supplementary Figure 1) using a plasmid previously described (Picelli et al., 2014) and 117 then generated tagmented sequencing libraries using our home-made and commercially 118 available hyperactive Tn5 enzymes . Our sequencing libraries served as a proof -of-concept, 119 demonstrating genome wide sequencing coverage derived from sperm nuclei is attainable via 120 ATAC-sequencing. 121 Our bulk ATAC-sequencing libraries were used to generate approximately 152 to 189 million 122 paired-end reads, with an average coverage of approximately 6 to 7x across the mouse genome 123 (Table 1). Our libraries exhibited minimal GC-bias and covered 42-50% of the informative 124 SNPs between C57BL/6J and FVB backgrounds (Table 1), which we used for haplotyping and 125 crossover detection previously (Tsui et al., 2023). While our ATAC-sequencing data does not 126 map homogenously throughout the genome at nucleotide resolution (Table 1, Figure 1A-C) – 127 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 7 as is expected with ATAC-seq – with the high number of SNPs in the genome we predicted it 128 would be sufficient for haplotyping and crossover calling. Our prediction was based on the 129 observation that when considering “bins” throughout the genome (Figure 1A -C), reads map 130 relatively homogenously to broader regions, even if at the resolution of a gene there is skewing 131 of reads towards regulatory elements and exons (Figure 1C). We therefore proceeded to test 132 and repurpose single-gamete ATAC-sequencing methods for meiotic crossover detection. 133 134 Table 1. Bulk-sequencing results from sperm nuclear DNA prepared with Tn5. Sequence 135 output metrics are provided for bulk ATAC-seq libraries produced from haploid sperm DNA. 136 137 138 Tn5 source Genotype Total Mapped read pairs Read length (bp) Average genome-wide coverage %GC Proportion of informative SNPs with a least one mapped read* Home-made Fancm+/+ 188,998,455 150 6.7 49 49.7% Commercial Fancm+/+ 151,991,673 150 5.5 50 41.5% Home-made Fancm∆2/∆2 179,633,072 150 6.1 51 45.4% Commercial Fancm∆2/∆2 162,885,127 150 5.8 51 41.5% .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 8 Figure 1: Bulk sequencing of tagmented sperm nuclei generates fragments that 139 map broadly with genome -wide coverage. Library were generated by Tn5 140 fragmentation of haploid sperm nuclei from wild type and mutant F1 hybrid mouse 141 testis, with reads aligned to the mouse reference genome ( mm39). Tracks show 142 reads from wildtype (orange) and mutant (blue) samples , generated with home -143 made Tn5, aligned to: A) The length of chromosome 12 (64 kb bins). B) A zoomed-144 in ~580 kb region (chr12: 64,834,440 to 65,414,128; 10 kb bins). C) A zoomed-in 145 ~72 kb region (chr12: 65,116,871 to 65,18 9,331; 100 b p bins). The g rey box 146 highlights the locus of exon 2 of Fancm; indicated by a red arrow for clarity. 147 *Maximum displayed coverage threshold set to 20. 148 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 9 snATAC-seq library generation, pre-processing and filtering. 149 To assay meiotic crossovers with high throughput, we generated single-sperm ATAC -150 sequencing libraries (10x Genomics) using FACS-sorted haploid spermatocytes isolated from 151 six F1 hybrid male mice; three F1.Fancm+/+ and three F1.FancmD2/D2. Coverage estimates of the 152 snATAC data were performed through assessing the number of fragments at common regions 153 of open chromatin across samples. Read coverage showed that the snATAC seq data from 154 haploid sperm had relatively homogeneous coverage across bins of the genome (Figure 2A), 155 replicating the genome-wide coverage observed in the bulk ATAC -seq data (Figure 1; Table 156 1). 157 158 Given that sperm chromatin is more condensed than in other cell types, bioinformatic tools 159 Signac (v1.13.0) (Stuart et al., 2021) and Seurat (v5.0.3) (Stuart et al., 2019) were employed to 160 assess chromatin accessibility and evaluate the quality of the dataset. We calculated the number 161 of fragments per cell, which provided a measure of chromatin accessibility for each cell in the 162 sample group. Additionally, Transcription Start Site (TSS) enrichment scores were computed 163 to indicate the accessibility of chromatin near TSS regions, reflecting the regulatory activity in 164 these areas. These metrics were visualized using a density scatter plot (Figure 2B). The relative 165 number of counts per cell aligns with the known inaccessibility of sperm chromatin (Oliva, 166 2006). snATAC-seq libraries were produced from three biological replicates of each genotype 167 (F1.Fancm+/+ and F1.FancmD2/D2) in two batches and sequenced in multiple runs (Figure 2C-E). 168 When assessing the percentage of read s in peaks (Figure 2C), the data segregated into two 169 groups, indicating possible batch effects , however this does not affect the ability to detect 170 crossovers (below). When exploring the overall structure of the data in a UMAP plot, the 171 samples from the first sequencing batch appeared to have an increased number of artefacts, 172 indicated by the large number of smaller clusters consisting of a few cells (Figure 2D). The 173 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 10 artefacts were removed, and the top genetic markers were identified from each cluster, for each 174 genotype. When comparing wild-type and mutant samples from the filtered data set , gene set 175 enrichment analysis (see Methods) indicated that cluster 0 corresponds to late-stage 176 spermatocytes, while cluster 1 represents early-stage sperm atocytes (Figure 2E). Cluster 2 177 could not be identified , indicating it may consist of contamination, unfiltered artefacts, or an 178 unannotated cell state. However, without single-cell RNA-sequencing or cytological analysis, 179 confirming the exact identify of this cluster remains challenging . Despite this ambiguity, the 180

confirm the correct cell types were sequenced with minimal contamination, making the 181 data suitable for further processing and crossover calling analysis. 182 183 The single sperm sequencing library wa s filtered for barcodes which were present with >10k 184 high quality fragments and excluded likely doublets, and cells with an incomplete complement 185 of chromosomes. The final dataset included more than 3800 haploid sperm genomes from 186 Fancm+/+ and FancmD2/D2 samples (Table 2, Supplementary table 1). Each cell was sequenced to 187 an average depth of 0.01x of the haploid genome, covering approximately 30 million base pairs 188 per cell on average. About 0.65% of the heterozygous SNPs between the two mouse strains 189 had at least one mapped read, which is sufficient for haplotyping (Table 2). To assess alignment 190 bias and sequencing bias we analysed allelic segregation ratios and found that throughout the 191 genome most loci segregated very close to a ratio of 1:1, in the three F1.Fancm+/+ and three 192 F1.FancmD2/D2 samples (Figure 3) . Some telomeric skewing remains on a limited number of 193 chromosomes after filtering repetitive regions, typically representing a mapping bias towards 194 the reference genome. 195 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 11 Figure 2: Quality control and validation of the haploid sperm snATAC-seq libraries. 196 A) Comparison of genome coverage across a ~60 kb window of chromosome 12 (Chr12: 197 65,120,000 to 65,180,000) from each mouse single gamete library. The locus contains Fancm, 198 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 12 with no reads mapping to the exon 2 of the F1.FancmD2/D2, which carry biallelic deletions (grey) 199 of exon 2. B) scatter plot of snATAC -seq data displaying quality control metrics (unique 200 fragments per cell and TSS enrichmint). Low TSS enrichment highlights the condensed nature 201 of sperm chromatin. C) Violin plot showing the percentage of reads that cluster into peaks for 202 each wild type and mutant replicate. Three replicates (wild-type / mutant pairs) were sequenced 203 across two sequencing runs (batch 1 and 2) . D) UMAP visualisations of snATAC -seq of 204 haploid sperm, from sequencing batches 1 and 2, showing cells grouped in 9 clusters. E) UMAP 205 of snATAC data for wildtype and mutant samples, post-filtering. 206 207 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 13 Figure 3. Marker segregation for the 19 autosomal chromosomes in Fancm wild type and 208 mutant snATAC-seq data . Genome-wide patterns of marker segregation from gametes 209 produced by F1(C57BL/6J x FVB) mouse with two haplotypes were calculated in chromosome 210 bins (of size 10 Mb) and found to match Mendelian segregation expectations, except for sub-211 telomeric regions (excluded from analysis due to mapping biases in repetitive regions ). 212 Hypothesis testing using a binomial test was performed to evaluate if marker segregation ratios 213 differ from 0.5; no significant differences were observed in any chromosomes. The y-axis units 214 of haplotype state ratio represent FVB or C57BL/6J ratios for given genomic bins, “1” 215 represents the alternate allele , which is FVB , and “0” represents the reference a llele 216 C57BL/6J. Most genomic regions have a haplotype state ratio close to 0.5, which is consistent 217 with chromosome segregation showing expected Mendelian ratios. 218 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 14 Meiotic crossover detection with snATAC-seq libraries 219 Using sgcocaller and comapr (Lyu et al., 2022) , we identified >25,000 meiotic crossovers 220 within our combined snATAC-sequencing samples. We found that crossover detection was 221 robust and was not affected by coverage within the ranges of read depth in our samples (Figure 222 4). 223 224 225 226 227 228 229 230 231 232 233 234 Figure 4: Analysis of genetic distance (cM) as a function of coverage reveals a robust 235 assay for crossover detection. Coverage is defined as reads per million [RPM] mapped reads 236 per sample. We find no significant correlation between genetic distance and coverage for both 237 wildtype and mutant F1 mouse samples using snATAC-seq, in 10kb bins. 238 239 Notably, the average number of crossovers inferred per sperm were higher in F1.FancmD2/D2 240 samples compared to F1.Fancm+/+ controls, with a median of 12 and 10 crossovers identified 241 per sperm, respectively (Two-sided Wilcoxon signed rank test with continuity correction, p < 242 2.2×10-16; Figure 5A), and the genetic map lengths were significantly longer in mutant (1,172.4 243 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 15 cM) than wild type ( 1,015.2 cM ) samples (Two-sided Wilcoxon signed rank test with 244 continuity correction, p < 2.2×10-16; Figure 5B). These findings align with our previous work 245 indicating Fancm acts as a meiotic crossover suppressor in mammals (Tsui et al., 2023). Most 246 individual chromosomes had genetic lengths of at least 50 cM, indicat ing good variant 247 detection along the length of each chromosome (Figure 5C). The differences in genetic distance 248 found in Fancm mice were observed along the length of individual chromosomes in bins of 10 249 Mb (Supplementary table 2) . Crossover numbers were elevated in mutant mice at a 250 chromosome level, with adjusted p -values above 0.01 in all chromosomes except 4, 5, 9, 15 251 and 19, as shown by both Kruskal-Wallis and permutation testing with false discovery rate 252 (FDR) correction (Supplementary table 2) . Although each chromosome had increased 253 crossovers in the mutant mice, the general crossover distribution – at any given locus – 254 remained comparable between both genotypes, with no significant difference between the 10 255 kb bins ( Permutation testing with FDR correction, p > 0.05; Figure 5D). These results are 256 consistent with our previous findings using single -sperm, bulk sequencing, and PCR -based 257

for crossover detection (Tsui et al., 2023) , suggesting that ATAC-seq provides a 258 robust tool for high-throughput crossover detection. 259 260 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 16 Table 2: Summary of crossovers detected from ATAC-sequencing of single sperm. Data 261 summarized from ATAC-sequencing results . Total numbers of COs from Fancm+/+ and 262 FancmΔ2/Δ2 samples were broken down into chromosomes with single, double , and >2 COs. 263 *Individual gametes were identified and filtered for a unique barcode which was represented 264 by >10 k high quality fragments (see Methods). 265 Genotype Cells* Single crossovers Double crossovers >2 crossovers Total crossovers Fancm+/+ 2,251 14,778 1,402 63 16,243 FancmD2/D2 1,581 10,496 1.617 104 12,217 266 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 17 267 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 18 Figure 5: snATAC-seq profiling of meiotic crossovers. Crossovers were assayed using 268 snATAC-sequencing of isolated haploid mouse sperm. A) Distribution of CO frequency 269 assayed per haploid sperm (n = 3 animals per genotype). Boxplot represents median, upper, 270 and lower quartile; whiskers represent the lowest and highest values within the 1.5 interquartile 271 range. B) Recombination rates measured for F1.Fancm+/+ and F1.FancmD2/D2 samples. 272 Observed crossover fractions were converted into genetic distances (centiMorgans) via the 273 Kosambi mapping function and presented as cumulative centiMorgans across the genome. C) 274 Average recombination frequency (cM) per chromosome per mutant and wild type samples. 275 D) Recombination rates (in cM) measured per 10 kb window along each chromosome position 276 (M, megabases) for Fancm+/+ (top) and FancmD2/D2 (bottom, flipped ) autosomes reveal the 277 increased crossover rate in mutant mice. 278 Genome-wide crossover distributions detected with snATAC-seq 279 Additionally, we investigated whether crossover interference could be detected with the 280 snATAC-seq method of crossover detection. Permutation testing via label swapping was used 281 to generate random null distributions of inter-crossover distances to simulate the absence of 282 crossover interference (null hypothesis), as we conducted previously (Tsui et al., 2023) . We 283 next filtered our dataset to chromosomes with only two crossovers (1,402 wild type and 1,617 284 mutant chromosomes; Table 2) and calculated the inter-crossover distance for each pair of 285 crossovers. Analysis of the spacing between two crossover events on the same chromosome 286 revealed that inter -crossover distance did not fit a random distribution in both wild type and 287 mutant samples, indicative of a functioning crossover interference mechanism (Figure 6A-C). 288 The observed median inter -crossover distance was 88.8 Mb in F1.Fancm+/+, and reduced to 289 82.6 Mb in F1.FancmD2/D2; pairwise comparisons using Wilcoxon rank sum test with continuity 290 correction, p < 2×10-7 (Figure 6A-C). The increase in crossovers in the absence of Fancm is 291 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 19 consistent with previous findings (Blary et al., 2018; Crismani et al., 2012; Fernandes et al., 292 2017; Girard et al., 2014, 2015; Lorenz et al., 2012; Mieulet et al., 2018; Séguéla -Arnaud et 293 al., 2015, 2016; Tsui et al., 2023) , suggesting the additional crossovers generated in gametes 294 lacking Fancm are likely to arise from the class II (non-interfering) crossover pathway. These 295 findings suggest that snATAC-seq libraries generated with methods here and using SSNIP-seq 296 (Novakovic et al., 2022) can be used for high-throughput and genome-wide measurements of 297 crossover interference. 298 299 300 301 302 303 304 305 306 Figure 6: Analysis of crossover interference in haploid sperm using single nuclei ATAC-307 seq data. The distance of observed crossovers was compared to the null hypothesis generated 308 from permutation via label swapping to simulate the absence of crossover interference . A) 309 Median distances for observed double crossover chromosomes from wild type samples was 310 approximately 88.8 Mb, compared to 0.024 Mb in the null hypothesis. Pairwise comparisons 311 using Wilcoxon rank sum test with continuity correction, p = 2×10-16. B) Median distances for 312 observed double crossover chromosomes from mutant samples was approximately 82.6 Mb, 313 compared to 0.0103 Mb in the null hypothesis. Pairwise comparisons using Wilcoxon rank sum 314 test with continuity correction, p < 2×10-16. C). Median inter-crossover distances were reduced 315 by 6.2 Mb in mutant crossovers, with an interquartile range of 5.9 Mb. These findings suggest 316 the additional crossovers in Fancm-deficient mice are likely to be derived from the non-317 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 20 interfering (type II) CO pathway. Pairwise comparisons using Wilcoxon rank sum test with 318 continuity correction , p < 2×10-7. Non-parametric statistical testing was used due to non -319 normal data distribution. 320 Crossover hotspots are associated with PRDM9 and DMC1 recombination hotspots 321 Crossovers identified in the snATAC -seq dataset here correlate with published PRDM9 322 binding locations, identified through ChIP-seq analysis from C57BL/6 mice (Biot et al., 2024). 323 Permutation testing was conducted by comparing the mean signal at overlapping regions with 324 the null distribution, generated by shuffling COs . P-values were calculated by comparing the 325 observed mean signal to the null distribution, testing if the association differs from a random 326 distribution. The results indicated there is a significant association between the COs, pooled 327 from both genotypes, and PRDM9 binding locations (p < 1×10-4; Figure 7A) , indicating a 328 broad requirement for PRDM9 for crossover localisation . COs from both genotypes were 329 pooled as no differences in distributions were detected (Fig. 5D) , even if there are generally 330 more COs in Fancm-deficient mice. However, there was no association between the COs (this 331 study) and C57BL/6 DMC1 binding sites (Biot et al., 2024), (permutation test, p = 0.3; Figure 332 7B). 333 334 COs hotspots were identified by sub -setting the data for the top 10% of COs in both 335 F1.FancmD2/D2 and F1.Fancm+/+ genotypes. Hotspots were enriched for PRDM9 ChIP -seq 336 signals; one-sided Wilcoxon signed rank test with continuity correction, p = 0.02 (Figure 7C-337 D). There was a strong, significant, association between the hotspots and DMC1 ChIP -seq 338 signal (one-sided Wilcoxon signed rank test with continuity correction , p = 1.5×10-7; Figure 339 7E-F). However, t here were no significant differences in either DMC1 or PRDM9 peak 340 association with CO sites when considering genotype, F1.FancmD2/D2 and F1.Fancm+/+, as a 341 variable. 342 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 21 343 The F1.FancmD2/D2 mutants have an increased number of COs, which are consistent with type 344 II CO/NCO distributions (Figure 6A-C). When assessing the ChIP-seq signal in the loci with 345 increased crossover frequencies in F1.FancmD2/D2 compared to all F1.Fancm+/+, there was not 346 a significant association with PRDM9 crossovers (Figure 7G-H). Permutation testing, however, 347 indicated that there was a significant association between the extra crossovers, compared to the 348 null distribution. This is likely due to the strong association with PRDM9 and all COs, observed 349 in Figure 7A. However, DMC1 showed enrichment for the ‘extra’ COs identified in 350 F1.FancmD2/D2 compared to F1.Fancm+/+ (one-sided Wilcoxon signed rank test with continuity 351 correction, p = 0.02, which was confirmed by permutation testing, p = 0.03’; Figure 7I-J). 352 353 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 22 354 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 23 Figure 7: Crossover sites in (FVBxC57BL/6J) have a strong association with PRDM9 355 binding sites, and to a lesser extent . A) Permutation test to compare observed means with 356 null distribution of PRDM9 ChIP -seq with all COs identifie d in snATAC-seq samples, p < 357 1×10-4. B) Permutation test to compare observed means with null distribution of DMC1 ChIP-358 seq with all COs identified in snATAC-seq samples; p = 0.3. C) One-sided Wilcoxon signed 359 rank test for PRDM9 ChIP-seq signals compared to the top 10% of COs identified in snATAC-360 seq; p = 0.02. D) Visual representation of the top 10% of COs and overlapping PRDM9 ChIP-361 seq peaks. E) One-sided Wilcoxon signed rank test DMC1 ChIP-seq signals compared to the 362 top 10% of COs (genetic distance multiplied by 10 for plotting) identified in snATAC-seq; p 363 < 1.5×10-7. F) Visual representation of the top 10% of COs (genetic distance multiplied by 10 364 for plotting) and overlapping DMC1 ChIP-seq peaks. G) One-sided Wilcoxon signed rank test 365 for PRDM9 ChIP -seq signals compared to the additional COs identified in F1.FancmD2/D2 366 samples, compared to all other COs; p = 0.14. H) Permutation test to compare observed means 367 with null distribution of PRDM9 ChIP-seq with additional F1.FancmD2/D2 COs; p < 1×10-4. I) 368 One-sided Wilcoxon signed rank test for DMC1 ChIP-seq signals compared to the additional 369 COs identified in F1.FancmD2/D2 samples, compared to all other COs; p = 0.002. J) Permutation 370 test to compare observed means with null distribution of DMC1 ChIP -seq with additional 371 F1.FancmD2/D2 COs; p = 0.03. 372 373

374 Tn5-based single gamete library preparation can be used for sequencing library 375 production 376 The detection of meiotic crossover events has historically relied on pedigree data, large 377 population cohorts, or cytogenetics. However, single-gamete library sequencing methods allow 378 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 24 one individual to serve as the exclusive source of data for high-throughput and accurate 379 crossover detection. These single-gamete approaches are scalable with several studies having 380 developed methods for library preparation and crossover calling software (Bell et al., 2019; 381 Campoy et al., 2020; Carioscia et al., 2022; Hinch et al., 2019; Lu et al., 2012; Lyu et al., 2022; 382 Sun et al., 2019; Tsui et al., 2023). Despite these advantages, there are some practical risks than 383 can limit the uptake of single-gamete sequencing for investigating crossover regulation, which 384 we have aimed to partially mitigate in this study : some published methods employ reagents 385 that are no longer commercially available (Bell et al., 2019; Tsui et al., 2023) . Therefore, we 386 sought to demonstrate proof -of-concept for single gamete library preparation with a view to 387 then use this data for crossover detection. We also suggest that it is important to be able to 388 adapt methods with home -made reagents to protect against flux of commercial kits. There is 389 precedent for such an approach , e.g., plate-based methods have been established for library 390 generation with somatic cells where bulk-tagmented nuclei are sorted into individual wells of 391 384-well plate for unique barcoding (Xu et al., 2021) . We anticipate that single -gamete 392 sequencing with ATAC -seq is amenable to plate -based approaches, using home -made Tn5 393 (Picelli et al., 2014), allowing for a cheaper alternative to commercial approaches for a smaller 394 number of samples, typically in the hundreds. 395 snATAC adaption for crossover calling from sperm 396 Here, we repurposed a single-nuclei ATAC-library preparation and sequencing method as a 397 straightforward, reproducible, and highly scalable approach for genome -scale haplotyping of 398 haploid sperm genomes. For sample preparation, we used our previously published SSNIP-seq 399 nuclei isolation protocol (Novakovic et al., 2022) , which allows for isolation of high -quality 400 nuclei suitable as input for ATAC -sequencing. The library preparation assay itself can be 401 completed in a single day with the collection of the sample, preparation to library generation 402 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 25 and amplification and QC. We sequenced the genomes of 3,842 haploid mouse sperm, 403 achieving uniform coverage across the genome at the kilobase scale . Using our establish ed 404 bioinformatic approaches (Lyu et al., 2022) we analysed the sequencing data to detect over 405 30,000 crossovers. We validated the utility of this novel approach by using our established 406 mouse model, F1.FancmD2/D2, which has increased crossover rates (Tsui et al., 2023), whereby 407 we showed that the snATAC-seq method detects this altered crossover behaviour. We also 408 showed that they assay is highly robust for crossover detection, as there was no observed 409 correlation between coverage and crossover rates, within the range of coverage that we 410 obtained. Further, with the high number of nuclei that were sequenced, this study offers the 411 best resolution to date of the effect of loss of function of Fancm in mammals, compared to our 412 previous work. We show that there is a genome-wide increase in crossover rates in the absence 413 of Fancm, these crossovers occur in the same loci as in the wildtype, and the effect in the 414 mutant appears to be an increase in amplitude of what occurs in the wildtype. This suggests 415 that the same DSB formation mechanisms are used, but the fate of some DSB repair 416 intermediates progress down a crossover formation rather than non-crossover pathway. 417 The association between CO hotspots, PRDM9 and DMC1 binding has previously been 418 established (Baudat et al., 2010; Hinch et al., 2019; R. Li et al., 2019; Myers et al., 2010; 419 Parvanov et al., 2010). It is therefore not unexpected that the COs observed in this study show 420 a similar association, however it is important to note the degree of shared crossover hotspots 421 in our data (Figure 7D, 7F) with the binding sites of pro -crossover factors from independent 422 work (Biot et al., 2024). The loci that account for the additional type II COs observed in Fancm-423 deficient m ice are located in regions enriched for DMC1, indicating they are repaired by 424 canonical crossover pathways. Although there was no specific enrichment for PRDM9 at these 425 COs, compared to other COs, the strong association with PRDM9 and the complete dataset of 426 COs might reduce the sensitivity to detect a further increase of PRDM9 at these extra 427 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 26 crossovers in Fancm-deficient mice. DMC1 is only enriched in regions containing high levels 428 of COs, and it is currently unclear why this is. It may be that it is simply easier to detect an 429 increase in DMC1signal from its baseline levels. Another possibility is that increases in DMC1 430 levels, when pushed upwards, can further increase crossover rates. However, further 431 experiments in future studies would be required to test these hypotheses. 432 433 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 27

and future perspectives 434 Our development of snATAC -seq for F1 sperm, coupled with bioinformatic analysis, 435 establishes this as a reliable method for studying crossover events and the mechanisms required 436 for their formation . This approach has revealed that Fancm plays a key role in limiting the 437 formation of class II crossovers mediated by intrinsic levels of DMC1. We anticipate th ese 438 findings and tools will have broader applications in fertility research and diagnostics, 439 potentially aiding in the quantification of sperm aneuploidy linked to chromosome segregation 440 errors (Templado et al., 2013). Future studies could further examine the interaction of Fancm 441 with other crossover pathways, advancing our understanding of recombination and 442 reproductive health. 443

444 We are grateful to all members of the Crismani laboratory for helpful comments on the 445 manuscript. Thank you to Tim Semple for helpful discussions about ATAC-seq methods and 446 to Adam Thomas for assistance with Tn5 production. WC and DJM receive Fellowships and 447 funding related to this work from the Australian National Health and Medical Research Council 448 (GNT1129757, GNT1112681, GNT1185387). 449 Authors’ contributions 450 All authors wrote, reviewed and discussed the manuscript. SN and CH prepared the figures. 451 All authors read and approved the final manuscript. 452 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 28 Ethics statement 453 All experimental procedures were approved by the St. Vincent’s Hospital Melbourne Animal 454 Ethics Committee. 455 Competing interests 456 The authors declare that they have no competing interests. 457

458 Animals 459 All animal experiments were approved by the Animal Ethics Committees at St Vincent’s 460 Hospital Melbourne and conducted in accordance with Australian NHMRC Guidelines on 461 Ethics in Animal Experimentation. All mice were housed at the BioResources Centre , St. 462 Vincent’s Hospital, in a controlled environment with a 12-hour light/dark cycle, and with food 463 and water provided ad libitum . The FancmΔ2/Δ2 mice used in this study were previous ly 464 described (Tsui, et al., 2023). 465 Isolation of haploid nuclei 466 Haploid nuclei were isolated from testicular single -nucleus suspensions and enriched for 467 haploids using FACS as previously described (Novakovic et al., 2022) . Briefly, testes were 468 dissected out from adult mice and placed in 1 mL of chilled Nuclei EZ Lysis Buffer (Sigma) . 469 The testes were gently squeezed with pointed tweezers to releases seminiferous tubule s into 470 the solution. Simultaneously, the spleen was also dissected and homogenised with 2 mL 471 Dulbecco's Phosphate Buffered Saline (DBPS) before being passed through a 40 μm strainer; 472 this sample served as a diploid control for flow cytometry. A 300 μL aliquot of splenic 473 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 29 homogenate was added to 1 mL Nuclei EZ Lysis Buffer and both the testis and splenic samples 474 were incubated for 5 minutes on ice, with mixing by gentle inversion after 3 minutes. To avoid 475 somatic contaminant, 600 μL of the testis suspension was transferred to a fresh 1.5 mL Lo -476 Bind microcentrifuge tube. Suspensions were centrifuged at 500 × g for 5 minutes at 4oC. The 477 supernatant was removed, and the pellet was resuspended with 1 mL of Nuclei EZ Lysis Buffer. 478 After a 5 minute incubation on ice, the samples were again centrifuged at 500 g for 5 minutes 479 at 4oC. Without disrupting the pellet, 1 mL of Nuclei Wash and Resuspension Buffer (NWRB; 480 DPBS with 1% [v/v] BSA) was slowly added and allowed to buffer exchange for 5 minutes on 481 ice. The pallet was resuspended by gentle pipetting up and down 10 times using a 1 mL wide-482 bore pipette tip. Samples were centrifuged at 500 × g for 5 minutes at 4oC, and most of the 483 supernatant was removed, leaving behind just enough to cover the pellet. Each sample was 484 resuspended with 300 μL NWRB supplemented with DAPI (10 μg / μL), and then filtered using 485 a 40 μm Flowmi Cell Strainer (Bel-Art) into a 5 mL polypropylene tube. Samples were FACS 486 sorted and haploid nuclei (peaks with DAPI content half of diploid control peaks) were 487 collected. 488 Bulk ATAC-sequencing of haploid sperm nuclei 489 Bulk tagmentation was performed using either commercial pre-assembled Tn5 transposase 490 (Diagenode) or with home-made Tn5 transposases produced and assembled using previously 491 published methods (Picelli et al., 2014) (Supplementary figure 1). 492 Libraries were prepared using previously published methods (Corces et al., 2017) with minor 493 modifications. Approximately 50,000 haploid sperm nuclei were FACS sorted into a 5 mL 494 polypropylene tube containing 300 uL NWRB with 0.1% (w/v) BSA. Nuclei were centrifuged 495 at 500 × g for 5 minutes at 4oC and the supernatant was aspirated. Without disrupting the pellet, 496 1 mL of OMNI -ATAC RSB buffer (10 mM Tris –HCl pH 8.0, 10 mM NaCl, 3 mM MgCl 2) 497 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 30 was slowly added and the sample was incubated for 5 minutes, then resuspended gently using 498 a wide bore tip. Nuclei were centrifuged at 500 × g for 5 minutes at 4 oC and the supernatant 499 was aspirated. Nuclei were resuspended in 100 μL chilled OMNI -ATAC RSB buffer 500 containing 0.1% (v/v) Tween 20, 0.1% (v/v) IGEPAL, and 0.01% (v/v) Digitonin and 501 incubated on ice for 2 minutes. Nuclei were centrifuged again at 500 × g for 5 minutes at 4oC 502 and the supernatant was aspirated. Nuclei were resuspended in 50 μL freshly prepared 503 transposition mix (33 mM Tris –HCl pH 8.0, 66 mM Potassium Acetate, 10 mM Magnesium 504 Acetate, 15% (v/v) N,N -dimethylformamide, 0.01% (v/v) Digitonin, 2.5 μL Tn5 assembled 505 enzyme) and mixed by pipetting up and down ten times. Transposition reactions were 506 incubated at 37°C for 45 minutes in a thermocycler with gentle mixing by tapping the tube at 507 5-minute intervals. Transposition reaction was stopped with 50 μL 2xATAC-STOP buffer (10 508 mM Tris–HCl pH 8.0, 20 mM EDTA). Afterwards, the samples were purified using a Monarch 509 Nuclei Acid Purification Kit (NEB) as per manufacturer’s instructions and eluted in 20 μL of 510 nuclease free water . All libraries were amplified with barcoded primers listed below, using 511 previously described methods (Buenrostro et al., 2015). 512 Mouse ID Line Genotype Tn5 Source Barcoding oligo ID Barcode 2024 F1(C57BL/6 x FVB) Fancm+/+ Home-made WC226 TAAGGCGA 2024 F1(C57BL/6 x FVB) Fancm+/+ Commercial (Diagenode) WC227 CGTACTAG 2025 F1(C57BL/6 x FVB) Fancm D2/D2 Home-made WC228 AGGCAGAA 2025 F1(C57BL/6 x FVB) FancmD2/D2 Commercial (Diagenode) WC229 TCCTGAGC 513 Single haploid sperm ATAC-seq library preparation and sequencing 514 Preparation of single-gamete ATAC-sequencing libraries was performed with modifications to 515 our previously published protocol (Novakovic et al., 2022) (ref SSNIP-seq paper DOI: 516 10.1371/journal.pone.0275168). Briefly, for each single -gamete ATAC -sequencing library, 517 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 31 haploid sperm nuclei were FACS sorted into a 5 mL polypropylene tube containing 300 µL 518 NWRB with 0.1% (w/v) BSA. Nuclei were centrifuged at 500 × g for 5 minutes at 4oC and the 519 supernatant was aspirated. Nuclei were resuspended in NWRB with 0.1% (w/v) BSA 520 supplemented with 0.01% (v/v) Digitonin and incubated on ice for 5 minutes. Nuclei were 521 centrifuged again at 500 × g for 5 minutes at 4oC and the supernatant was aspirated. 500 μL of 522 Diluted Nuclei Buffer (10x Genomics) was slowly added to the nuclei without disrupting the 523 pellet and allowed to buffer exchange for 5 minutes on ice. The nuclei were resuspended in 10 524 μL of Diluted Nuclei Buffer (10x Genomics) and quantified using a hemocytometer. Nuclei 525 concentration was adjusted to attain a capture number of around 1,000 nuclei, accounting for 526 loss during subsequent steps. Prepared nuclei were added to the Transposition Mix (10x 527 Genomics) and transposition, library preparation and clean -up for single -cell ATAC -528 sequencing was performed according to the manufacturers protocol using the Chromium 529 Single-Cell ATAC Kit v1.1 (10x Genomics). Each library was sequenced on the NovaSeq 6000 530 (Illumina) using a 50/8/16/50 bp read configuration. 531 Data processing 532 Bulk ATAC-sequencing fastq files were trimmed using cutadapt (v3.4) (Martin, 2011) and 533 mapped to the mm39 reference genome using bwa-mem with default settings (H. Li, 2013) . 534 Duplicate reads were removed using MarkDuplicates (picard tools). 535 For snATAC-sequencing, fastq files were processed using Cell Ranger ATAC (v2.1.0; 10x 536 Genomics) to align reads to the mm39 reference genome. Outputs from Cell Ranger were used 537 in downstream analysis. 538 Sequencing coverage analysis 539 Coverage file s for bulk and snATAC experiments were generated from BAM files using 540 samtools depth (H. Li et al., 2009). Data were summarised in 10 kb bins based on chromosome 541 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 32 lengths from mm39. Genomic tracks were visualised with Gviz, with gene annotations added 542 from UCSC genome browser (Hahne & Ivanek, 2016). 543 Single nuclei ATAC-sequencing analysis with Signac and Seurat 544 The outputs from Cell Ranger containing peak and fragment information were used to generate 545 coverage estimates of the snATAC-seq data by identifying a set of common peaks across all 546 samples, indicative of open chromatin regions. Barcodes with fewer than 500 fragments per 547 cell were filtered out. Fragment counts were then normalised using term -frequency inverse-548 document-frequency (TF-IDF) to result in the logarithmic ratio of the total number of reads per 549 cell, relative to total cells. 550 The normalised data was used to assess coverage in specified genetic regions, as well as to 551 assess various QC metrics, including: TSS enrichment scores, fragment counts ( nCounts), 552 nucleosome signal, and percentage of reads in peaks. Dimensionality reduction was performed 553 using Singular Value Decomposition (SVD) , and Uniform Manifold Approximation and 554 Projection (UMAP) plots were generated to visualise clusters, with Latent Semantic Indexing 555 (LSI) applied for data reduction. Spare clusters were filtered, and UMAP plots were separated 556 into sequencing batches. Gene activities were annotated using the Ensembl database and the 557 mm39 reference genome. Cluster-specific markers with positive fold changes between clusters 558 were identified, and Enrichr was used to identify putative cell types based on the top 559 differentially accessible genes. 560 Single nuclei ATAC-sequencing crossover calling with sgcocaller and comapr 561 Crossover calling was conducted on the Cell Ranger BAM output file using sgcocaller (Lyu et 562 al., 2022) , with the following parameters : --cmPmb 0.1 --maxDP 20 --maxTotalDP 450 --563 minTotalDP 0 --minDP 0 --thetaREF 0.2 --thetaALT 0.8. FVB/N-specific variants differing 564 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 33 from the mm 10 reference genome were obtained from the Mouse Genome Project 565 (FVB_NJ.mgp.v5.snps.dbSNP142.vcf) and lifted over for compatibility with the mm39 566

genome. 567 The outputs from Cell Ranger were filtered to generate a list of barcodes from intact cells with 568 greater than 10 ,000 reads. To refine crossover calls , the filtered barcode file and output of 569 sgcocaller were processed utilising comapr (Lyu et al., 2022), generating genetic distance maps 570 with the Kosambi mapping function, and identifying inter-crossover distances. The following 571 thresholds were set in comapr: minSNP = 10, minCellSNP = 100, maxR awCO = 5, 572 minLogllRatio = 30, bpDist = 1e5. Artefacts with double crossovers occurring within 3 bp were 573 removed from the analysis. 574 Marker segregation for single nuclei ATAC-sequencing 575 Segregation states of SNPs were analysed in 10 Mb bins, based on chromosome lengths from 576 the mm39 reference genome, using output files generated by sgcocaller. SNP positions and 577 their underlying segregation states were used to calculate the ratio of BL6 and FVB/N genetics, 578 where 1 indicated FVB/N and 0 indicated BL6. Haplotype ratios were computed, and binomial 579 distribution tests were performed to assess ed bias in marker segregation. Telomeric regions 580 were removed due to low coverage, which resulted in inaccurate state imputation. 581 Permutation tests and null distributions generation 582 Permutation tests using label swapping and null distribution generation were conducted 583 following previously described methods (Tsui et al., 2023). 584 ChIP-seq compared to CO analysis 585 C57BL/6 PRDM9 and DMC1 ChIP-seq data (Biot et al., 2024) were downloaded from NCBI’s 586 Sequence Read Archive (SRA), using the SRA toolkit. ChIP-seq reads were aligned to the 587 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 34 mm39 reference genome using bowtie2 (Langmead & Salzberg, 2012) and filtered to remove 588 low-quality reads and duplicates using samtools (fixmate, sort, and markdup; (H. Li et al., 589 2009)). PRDM9 and DMC1 knockout samples were excluded from wild-type analysis, and 590 ChIP-seq peaks were called using MACS2 (Zhang et al., 2008). 591 CO positions were derived from the snATAC-seq data, processed using sgcocaller and comapr. 592 Overlapping and ChIP-seq and CO positions were identified using the findOverlaps function 593 from GenomicRanges (Lawrence et al., 2013) . Statistical comparisons were conducted using 594 one-sided Wilcoxon signed-rank tests, and as permutation tests, comparing observed ChIP-seq 595 signals to the null distribution. 596 Oligonucleotide sequences used in this study. 597 Oligo name Oligonucleotide Sequence (5′-3′) WC226 CAAGCAGAAGACGGCATACGAGATTCGCCTTAGTCTCGTGGGCTCGGAGATGT WC227 CAAGCAGAAGACGGCATACGAGATCTAGTACGGTCTCGTGGGCTCGGAGATGT WC228 CAAGCAGAAGACGGCATACGAGATTTCTGCCTGTCTCGTGGGCTCGGAGATGT WC229 CAAGCAGAAGACGGCATACGAGATGCTCAGGAGTCTCGTGGGCTCGGAGATGT 598 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 1

599 Adey, A., Morrison, H. G., (no last name), A., Xun, X., Kitzman, J. O., Turner, E. H., 600 Stackhouse, B., MacKenzie, A. P., Caruccio , N. C., Zhang, X., & Shendure, J. (2010). 601 Rapid, low-input, low-bias construction of shotgun fragment libraries by high -density in 602 vitro transposition. Genome Biology, 11(12), R119. https://doi.org/10.1186/gb-2010-11-603 12-r119 604 Baudat, F., Buard, J., Grey, C., Fledel -Alon, A., Ober, C., Przeworski, M., Coop, G., & de 605 Massy, B. (2010). PRDM9 Is a Major Determinant of Meiotic Recombination Hotspots 606 in Humans and Mice. Science, 327(5967), 836 –840. 607 https://doi.org/10.1126/science.1183439 608 Bell, A. D., Mello, C. J., Nemesh, J., Brumbaugh, S. A., Wysoker, A., & McCarroll, S. A. 609 (2019). Insights about variation in meiosis from 31,228 human sperm genomes. Nature, 610 2019(May 2019), 625202. https://doi.org/10.1101/625202 611 Berchowitz, L. E., & Copenhaver, G. P. (2010). Genetic interference: don’t stand so close to 612 me. Current Genomics, 11(2), 91–102. https://doi.org/10.2174/138920210790886835 613 Biot, M., Toth, A., Brun, C., Guichard, L., de Massy, B., & Grey, C. (2024). Principles of 614 chromosome organization for meiotic recombination. Molecular Cell , 84(10), 1826 -615 1841.e5. https://doi.org/10.1016/j.molcel.2024.04.001 616 Blary, A., Gonzalo, A., Eber, F., Bérard, A., Bergès, H., Bessoltane, N., Charif, D., Charpentier, 617 C., Cromer, L., Fourment, J., Genevriez, C., Le Paslier, M. -C., Lodé, M., Lucas, M. -O., 618 Nesi, N., Lloyd, A., Chèvre, A. -M., & Jenczewski, E. (2018). FANCM Limits Meiotic 619 Crossovers in Brassica Crops. Frontiers in Plant Science , 9. 620 https://doi.org/10.3389/fpls.2018.00368 621 Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y., & Greenleaf, W. J. (2013). 622 Transposition of native chromatin for fast and sensitive epigenomic profiling of open 623 chromatin, DNA -binding proteins and nucleosome position. Nature Methods , 10(12), 624 1213–1218. https://doi.org/10.1038/nmeth.2688 625 Buenrostro, J. D., Wu, B., Chang, H. Y., & Greenleaf, W. J. (2015). ATAC‐seq: A Method for 626 Assaying Chromatin Accessibility Genome‐Wide. Current Protocols in Molecular 627 Biology, 109(1). https://doi.org/10.1002/0471142727.mb2129s109 628 Campoy, J., Sun, H., Goel, M., Jiao, W.-B., Folz-Donahue, K., Wang, N., Rubio, M., Liu, C., 629 Kukat, C., Ruiz, D., Huettel, B., & Schneeberger, K. (2020). Chromosome-level and 630 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 2 haplotype-resolved genome assembly enabled by high-throughput single-cell sequencing 631 of gamete genomes. https://doi.org/10.1101/2020.04.24.060046 632 Carioscia, S. A., Weaver, K. J., Bortvin , A. N., Pan, H., Ariad, D., Bell, A. D., & McCoy, R. 633 C. (2022). A method for low -coverage single -gamete sequence analysis demonstrates 634 adherence to Mendel’s first law across a large sample of human sperm. ELife, 11. 635 https://doi.org/10.7554/eLife.76383 636 Cole, F., Baudat, F., Grey, C., Keeney, S., De Massy, B., & Jasin, M. (2014). Mouse tetrad 637 analysis provides insights into recombination mechanisms and hotspot evolutionary 638 dynamics. Nature Genetics, 46(10), 1072–1080. https://doi.org/10.1038/ng.3068 639 Corces, M. R., Trevino, A. E., Hamilton, E. G., Greenside, P. G., Sinnott -Armstrong, N. A., 640 Vesuna, S., Satpathy, A. T., Rubin, A. J., Montine, K. S., Wu, B., Kathiria, A., Cho, S. 641 W., Mumbach, M. R., Carter, A. C., Kasowski, M., Orloff, L. A., Risca, V. I., Kundaje, 642 A., Khavari, P. A., … Chang, H. Y. (2017). An improved ATAC -seq protocol reduces 643

and enables interrogation of frozen tissues. Nature Methods, 14(10), 959–644 962. https://doi.org/10.1038/nmeth.4396 645 Crismani, W., Girard, C., Froger, N., Pradillo, M., Santos, J. L. L., Chelysheva, L., Copenhaver, 646 G. P., Horlow, C., & Mercier, R. (2012). FANCM Limits Meiotic Crossovers. Science, 647 336(6088), 1588–1590. https://doi.org/10.1126/science.1220381 648 Cusanovich, D. A., Daza, R., Adey, A., Pliner, H. A., Christiansen, L., Gunderson, K. L., 649 Steemers, F. J., Trapnell, C., & Shendure, J. (2015). Multiplex single -cell profiling of 650 chromatin accessibility by combinatorial cellular indexing. Science, 348(6237), 910–914. 651 https://doi.org/10.1126/science.aab1601 652 Fernandes, J. B., Séguéla -Arnaud, M., Larchevêque, C., Lloyd, A. H., & Mercier, R. (2017). 653 Unleashing meiotic crossovers in hybrid plants. Proceedings of the National Academy of 654 Sciences, 115(10), 2431–2436. https://doi.org/10.1073/pnas.1713078114 655 Fogel, S., & Mortimer, R. K. (1971). RECOMBINATION IN YEAST. Annual Review of 656 Genetics, 5(1), 219–236. https://doi.org/10.1146/annurev.ge.05.120171.001251 657 Francis, K. E., Lam, S. Y., Harrison, B. D., Bey, A. L., Berchowitz, L. E., & Copenhaver, G. 658 P. (2007). Pollen tetrad -based visual assay for meiotic recombination in Arabidopsis. 659 Proceedings of the National Academy of Sciences of the United States of America , 660 104(10), 3913–3918. https://doi.org/10.1073/pnas.0608936104 661 Girard, C., Chelysheva, L., Choinard, S., Froger, N., Macaisne, N., Lehmemdi, A., Mazel, J., 662 Crismani, W., & Mercier, R. (2015). AAA -ATPase FIDGETIN -LIKE 1 and Helicase 663 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 3 FANCM Antagonize Meiotic Crossovers by Distinct Mechanisms. PLoS Genetics, 11(7). 664 https://doi.org/10.1371/journal.pgen.1005369 665 Girard, C., Crismani, W., Froger, N., Mazel, J., Lemhemdi, A., Horlow, C., & Mercier, R. 666 (2014). FANCM-associated proteins MHF1 and MHF2, but not the other Fanconi anemia 667 factors, limit meiotic crossovers. Nucleic Acids Research , 42(14). 668 https://doi.org/10.1093/nar/gku614 669 Goryshin, I. Y., & Reznikoff, W. S. (1998). Tn5 in Vitro Transposition. Journal of Biological 670 Chemistry, 273(13), 7367–7374. https://doi.org/10.1074/jbc.273.13.7367 671 Gray, S., & Cohen, P. E. (2016). Control of Meiotic Crossovers: From Double -Strand Break 672 Formation to Designation. In Annual Review of Genetics (Vol. 50, pp. 175–210). Annual 673 Reviews Inc. https://doi.org/10.1146/annurev-genet-120215-035111 674 Hahne, F., & Ivanek, R. (2016). Visualizing Genomic Data Using Gviz and Bioconductor (pp. 675 335–351). https://doi.org/10.1007/978-1-4939-3578-9_16 676 Hammoud, S. S., Nix, D. A., Zhang, H., Purwar, J., Carrell, D. T., & Cairns, B. R. (2009). 677 Distinctive chromatin in human sperm packages genes for embryo development. Nature, 678 460(7254), 473–478. https://doi.org/10.1038/nature08162 679 Hassold, T., & Sherman, S. (2000). Down syndrome: Genetic recombination and the origin of 680 the extra chromosome 21. In Clinical Genetics (Vol. 57, Issue 2, pp. 95 –100). 681 https://doi.org/10.1034/j.1399-0004.2000.570201.x 682 Hinch, A. G., Zhang, G., Becker, P. W., Moralli, D., Hinch, R., Davies, B., Bowden, R., & 683 Donnelly, P. (2019). Factors influencing meiotic recombination revealed by whole -684 genome sequencing of single sperm. Science, 363(6433), eaau8861. 685 https://doi.org/10.1126/science.aau8861 686 Hunter, N. (2015). Meiotic Recombination : The Essence of Heredity. Cold Spring Harbor 687 Perspectives in Biology, 7. https://doi.org/10.1101/cshperspect.a016618 688 Lam, I., & Keeney, S. (2015). Mechanism and Regulation of Meiotic Recombination Initiation. 689 Cold Spring Harbor Perspectives in Biology , 7(1), a016634. 690 https://doi.org/10.1101/cshperspect.a016634 691 Langmead, B., & Salzberg, S. L. (2012). Fast gapped -read alignment with Bowtie 2. Nature 692 Methods, 9(4), 357–359. https://doi.org/10.1038/nmeth.1923 693 Lawrence, M., Huber, W., Pagès, H., Aboyoun, P., Carlson, M., Gentleman, R., Morgan, M. 694 T., & Carey, V. J. (2013). Software for Computing and Annotating Genomic Ranges. 695 PLoS Computational Biology, 9(8). https://doi.org/10.1371/journal.pcbi.1003118 696 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 4 Leung, A. Q., Bell, A. D., Mello, C. J., Penzias, A. S., McCarroll, S. A., & Sakkas, D. (2021). 697 Single cell analysis of DNA in more than 10,000 individual sperm from men with 698 abnormal reproductive outcomes. Journal of Assisted Reproduction and Genetics , 699 0123456789. https://doi.org/10.1007/s10815-021-02300-3 700 Li, H. (2013). Aligning sequence reads, clone sequences and assembly contigs with BWA -701 MEM. http://github.com/lh3/bwa. 702 Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., 703 & Durbin, R. (2009). The Sequence Alignment/Map format and SAMtools. 704 Bioinformatics, 25(16), 2078–2079. https://doi.org/10.1093/bioinformatics/btp352 705 Li, R., Bitoun, E., Altemose, N., Davies, R. W., Davies, B., & Myers, S. (2019). A high -706 resolution map of non -crossover events reveals impacts of genetic diversity on 707 mammalian meiotic recombination. Nature Communications , 10(1), 3900. 708 https://doi.org/10.1038/s41467-019-11675-y 709 Lorenz, A., Osman, F., Sun, W., Nandi, S., Steinacher, R., & Whitby, M. C. (2012). The Fission 710 Yeast FANCM Ortholog Directs Non-Crossover Recombination During Meiosis. Science, 711 336(6088), 1585–1588. https://doi.org/10.1126/science.1220111 712 Lu, S., Zong, C., Fan, W., Yang, M., Li, J., Chapman, A. R., Zhu, P., Hu, X., Xu, L., Yan, L., 713 Bai, F., Qiao, J., Tang, F., Li, R., & Xie, X. S. (2012). Probing meiotic recombination and 714 aneuploidy of single sperm cells by whole-genome sequencing. Science (New York, N.Y.), 715 338(6114), 1627–1630. https://doi.org/10.1126/science.1229112 716 Luo, C., Li, X., Zhang, Q., & Yan, J. (2019). Single gametophyte sequencing reveals that 717 crossover events differ between sexes in maize. Nature Communications , 10(1). 718 https://doi.org/10.1038/s41467-019-08786-x 719 Lyu, R., Tsui, V., Crismani, W., Liu, R., Shim, H., & McCarthy, D. J. (2022). sgcocaller and 720 comapr: personalised haplotype assembly and comparative crossover map analysis using 721 single-gamete sequencing data. Nucleic Acids Research , 1 –13. 722 https://doi.org/10.1093/nar/gkac764 723 Martin, M. (2011). Cutadapt removes adapter sequences from high -throughput sequencing 724 reads. EMBnet.Journal, 17(1), 10. https://doi.org/10.14806/ej.17.1.200 725 McClintock, B. (1931). The Order of the Genes C, Sh and Wx in Zea Mays with Reference to 726 a Cytologically Known Point in the Chromosome. Proceedings of the National Academy 727 of Sciences, 17(8), 485–491. https://doi.org/10.1073/pnas.17.8.485 728 Mieulet, D., Aubert, G., Bres, C., Klein, A., Droc, G., Vieille, E., Rond-Coissieux, C., Sanchez, 729 M., Dalmais, M., Mauxion, J. P., Rothan, C., Guiderdoni, E., & Mercier, R. (2018). 730 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 5 Unleashing meiotic crossovers in crops. Nature Plants , 4(12), 1010 –1016. 731 https://doi.org/10.1038/s41477-018-0311-x 732 Morgan, T. H. (1910). Sex limited inheritance in drosophila. Science, 32(812), 120 –122. 733 https://doi.org/10.1126/science.32.812.120 734 Myers, S., Bowden, R., Tumian, A., Bontrop, R. E., Freeman, C., MacFie, T. S., McVean, G., 735 & Donnelly, P. (2010). Drive Against Hotspot Motifs in Primates Implicates the PRDM9 736 Gene in Meiotic Recombination. Science, 327(5967), 876 –879. 737 https://doi.org/10.1126/science.1182363 738 Novakovic, S., Tsui, V., Semple, T., Martelotto, L., McCarthy, D. J., & Crismani, W. (2022). 739 SSNIP-seq: A simple and rapid method for isolation of single-sperm nucleic acid for high-740 throughput sequencing. Plos One , 17(9), e0275168. 741 https://doi.org/10.1371/journal.pone.0275168 742 Oliva, R. (2006). Protamines and male infertility. Human Reproduction Update, 12(4), 417–743 435. https://doi.org/10.1093/humupd/dml009 744 Parvanov, E. D., Petkov, P. M., & Paigen, K. (2010). Prdm9 controls activation of mammalian 745 recombination hotspots. Science (New York, N.Y.) , 327(5967), 835. 746 https://doi.org/10.1126/science.1181495 747 Picelli, S., Björklund, Å. K., Reinius, B., Sagasser , S., Winberg, G., & Sandberg, R. (2014). 748 Tn5 transposase and tagmentation procedures for massively scaled sequencing projects. 749 Genome Research, 24(12), 2033–2040. https://doi.org/10.1101/gr.177881.114 750 Schwacha, A., & Kleckner, N. (1994). Identification of joint molecules that form frequently 751 between homologs but rarely between sister chromatids during yeast meiosis. Cell, 76(1), 752 51–63. https://doi.org/10.1016/0092-8674(94)90172-4 753 Séguéla-Arnaud, M., Choinard, S., Larchevêque, C., Girard, C., Froger, N., Crismani, W., & 754 Mercier, R. (2016). RMI1 and TOP3α limit meiotic CO formation through their C -755 terminal domains. Nucleic Acids Research , 45(4), 1860 –1871. 756 https://doi.org/10.1093/nar/gkw1210 757 Séguéla-Arnaud, M., Crismani, W., Mazel, J., Froger, N., Choinard, S., Lemhemdi, A., 758 Macaisne, N., Leene, J. Van, Gevaert, K., Jaeger, G. De, Chelysheva, L., & Mercier, R. 759 (2015). Multiple mechanisms limit meiotic crossovers : TOP3 α and two BLM homologs 760 antagonize crossovers in parallel to FANCM. Proceedings of the National Academy of 761 Sciences, Early edit. https://doi.org/10.1073/pnas.1423107112 762 Sherman, S. L., Takaesu, N., Freeman, S. B., Grantham, M., Phillips, C., Blackston, R. D., 763 Jacobs, P. A., Cockwell, A. E., Freeman, V., Uchida, I., Mikkelsen, M., Kurnit, D. M., 764 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 6 Buraczynska, M., Keats, B. J. B., & Hassold, T. J. (1991). Trisomy 21: Association 765 between reduced recombination and nondisjunction. American Journal of Human 766 Genetics, 49(3), 608–620. 767 Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W. M., Hao, Y., 768 Stoeckius, M., Smibert, P., & Satija, R. (2019). Comprehensive Integration of Single-Cell 769 Data. Cell, 177(7), 1888-1902.e21. https://doi.org/10.1016/j.cell.2019.05.031 770 Stuart, T., Srivastava, A., Madad, S., Lareau, C. A., & Satija, R. (2021). Single-cell chromatin 771 state analysis with Signac. Nature Methods , 18(11), 1333 –1341. 772 https://doi.org/10.1038/s41592-021-01282-5 773 Sturtevant, A. H. (1913). The linear arrangement of six sex‐linked factors in Drosophila, as 774 shown by their mode of association. Journal of Experimental Zoology , 14(1), 43 –59. 775 https://doi.org/10.1002/jez.1400140104 776 Sun, H., Rowan, B. A., Flood, P. J., Brandt, R., Fuss, J., Hancock, A. M., Michelmore, R. W., 777 Huettel, B., & Schneeberger, K. (2019). Linked -read sequencing of gametes allows 778 efficient genome-wide analysis of meiotic recombination. Nature Communications, 10(1), 779 1–9. https://doi.org/10.1038/s41467-019-12209-2 780 Szostak, J. W., Orr-Weaver, T. L., Rothstein, R. J., & Stahl, F. W. (1983). The double-strand-781 break repair model for recombination. Cell, 33(1), 25–35. https://doi.org/10.1016/0092-782 8674(83)90331-8 783 Tsui, V., Lyu, R., Novakovic, S., Stringer, J. M., Dunleavy, J. E. M., Granger, E., Semple, T., 784 Leichter, A., Martelotto, L. G., Merriner, D. J., Liu, R., McNeill, L., Zerafa, N., Hoffmann, 785 E. R., O’Bryan, M. K., Hutt, K., Deans, A. J., Heierhorst, J., McCarthy, D. J., & Crismani, 786 W. (2023). Fancm has dual roles in the limiting of meiotic crossovers and germ cell 787 maintenance in mammals. Cell Genomics , 100349. 788 https://doi.org/10.1016/j.xgen.2023.100349 789 Warren, A. C., Chakravarti, A., Wong, C., Slaugenhaupt, S. A., Halloran, S. L., Watkins, P. 790 C., Metaxotou, C., & Antonarakis, S. E. (1987). Evidence for reduced recombination on 791 the nondisjoined chromosomes 21 in down syndrome. Science, 237(4815), 652 –654. 792 https://doi.org/10.1126/science.2955519 793 Xie, H., Li, W., Guo, Y., Su, X., Chen, K., Wen, L., & Tang, F. (2023). Long-read-based single 794 sperm genome sequencing for chromosome -wide haplotype phasing of both SNPs and 795 SVs. Nucleic Acids Research, 51(15), 8020–8034. https://doi.org/10.1093/nar/gkad532 796 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 7 Xu, W., Wen, Y., Liang, Y., Xu, Q., Wang, X., Jin, W., & Chen, X. (2021). A plate -based 797 single-cell ATAC-seq workflow for fast and robust profiling of chromatin accessibility. 798 Nature Protocols, 16(8), 4084–4107. https://doi.org/10.1038/s41596-021-00583-5 799 Yelina, N. E., Ziolkowski, P. A., Miller, N., Zhao, X., Kelly, K. A., Muñoz, D. F., Mann, D. 800 J., Copenhaver, G. P., & Henderson, I. R. (2013). High -throughput analysis of meiotic 801 crossover frequency and interference via flow cytometry of fluorescent pollen in 802 Arabidopsis thaliana. Nature Protocols , 8(11), 2119 –2134. 803 https://doi.org/10.1038/nprot.2013.131 804 Zhang, Y., Liu, T., Meyer, C. A., Eeckhoute, J., Johnson, D. S., Bernstein, B. E., Nussbaum, 805 C., Myers, R. M., Brown, M., Li, W., & Shirley, X. S. (2008). Model -based analysis of 806 ChIP-Seq (MACS). Genome Biology, 9(9). https://doi.org/10.1186/gb-2008-9-9-r137 807 808 809 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 1 Supplementary information 810 Supplementary table 1: Summary of sequencing quality and crossovers detected in snATAC sequencing. 811 Sampl e ID Genotype Unfiltered cells High-quality fragments, >10k per barcode High-quality fragments per cell, >10k per barcode Total number of sequenced read pairs Average coverage per BAM file Barcodes passed filtering for COs - number of sperm Mean COs per sample Single COs Double COs Triple COs Quadruple COs 1826 Fancm+/+ 1,007 687 579 881,941,291 10.3488 173 10.7 1365 133 13 0 1919 Fancm+/+ 1,674 1794 1664 555,467,003 9.42015 1200 11.1 7796 875 31 4 1929 Fancm+/+ 889 1417 1362 608,787,524 5.05501 878 9.7 5617 394 14 1 1825 Fancm∆2/∆2 1,452 1083 923 1,015,085,96 0 18.5083 412 11.6 2678 366 26 7 1921 Fancm∆2/∆2 785 833 783 585,535,532 8.6143 539 12.2 3683 563 31 2 1928 Fancm∆2/∆2 889 940 886 1,078,035,28 3 9.96556 640 12.1 4135 688 36 2 812 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 2 Supplementary table 2: Genetic distances of each chromosome with a comparison between genotypes, with p corrected p-values. 813 Chromosome Genotype Sample Genetic distance per replicate (cM) Combined genetic distance (cM) Kruskal statistic Kruskal p-value Adjusted p-value (FDR) Significant Chr1 Fancm+/+ 1826 73.1 221.9 30.36485027 3.57956E-08 3.57957E-08 Yes 1919 81.2 1929 67.6 Chr1 Fancm∆2/∆2 1825 84.3 250.1 1921 80.5 1928 85.4 Chr2 Fancm+/+ 1826 64.4 198.2 149.0275109 2.82831E-34 2.82833E-34 Yes 1919 69.6 1929 64.3 Chr2 Fancm∆2/∆2 1825 81.1 244.9 1921 84.8 1928 79.0 Chr3 Fancm+/+ 1826 75.6 190.5 6.594019515 0.010232189 0.010232207 Yes 1919 61.7 1929 53.1 Chr3 Fancm∆2/∆2 1825 73.4 217.6 1921 68.2 1928 76.0 Chr4 Fancm+/+ 1826 66.9 199.9 2.81710079 0.093264724 0.093264845 No 1919 70.5 1929 62.6 Chr4 Fancm∆2/∆2 1825 64.4 215.9 1921 77.8 1928 73.7 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 3 Chr5 Fancm+/+ 1826 54.4 194.6 2.172197924 0.140525262 0.140525439 No 1919 76.6 1929 63.6 Chr5 Fancm∆2/∆2 1825 77.9 231.2 1921 76.6 1928 76.6 Chr6 Fancm+/+ 1826 52.5 166.3 120.7478494 4.33923E-28 4.33926E-28 Yes 1919 60.3 1929 53.5 Chr6 Fancm∆2/∆2 1825 69.2 207.4 1921 69.4 1928 68.8 Chr7 Fancm+/+ 1826 50.0 158.1 142.9302417 6.09E-33 6.08817E-33 Yes 1919 60.1 1929 48.1 Chr7 Fancm∆2/∆2 1825 61.2 197.5 1921 69.6 1928 66.7 Chr8 Fancm+/+ 1826 40.6 122.9 8.342491191 0.003872846 0.003872856 Yes 1919 45.9 1929 36.4 Chr8 Fancm∆2/∆2 1825 52.6 155.4 1921 52.8 1928 50.1 Chr9 Fancm+/+ 1826 65.0 176.3 1.760751996 0.184530746 0.184531039 No 1919 56.0 1929 55.2 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 4 Chr9 Fancm∆2/∆2 1825 56.7 190.2 1921 69.4 1928 64.1 Chr10 Fancm+/+ 1826 58.1 167.0 128.9050538 7.11414E-30 7.11E-30 Yes 1919 58.6 1929 50.2 Chr10 Fancm∆2/∆2 1825 59.9 199.2 1921 69.4 1928 69.9 Chr11 Fancm+/+ 1826 73.8 200.3 12.35816708 0.000439062 0.000439063 Yes 1919 67.1 1929 59.4 Chr11 Fancm∆2/∆2 1825 64.7 214.0 1921 73.0 1928 76.2 Chr12 Fancm+/+ 1826 51.3 148.6 22.41418947 2.19745E-06 2.19746E-06 Yes 1919 51.6 1929 45.8 Chr12 Fancm∆2/∆2 1825 55.1 158.2 1921 48.9 1928 54.1 Chr13 Fancm+/+ 1826 50.0 150.3 8.002043479 0.004672459 0.004672472 Yes 1919 53.4 1929 46.9 Chr13 Fancm∆2/∆2 1825 53.5 169.2 1921 56.4 1928 59.2 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 5 Chr14 Fancm+/+ 1826 48.8 132.3 6.702211243 0.009629342 0.009629366 Yes 1919 45.6 1929 37.9 Chr14 Fancm∆2/∆2 1825 49 144.7 1921 45.8 1928 49.9 Chr15 Fancm+/+ 1826 39.4 110.5 1.081911776 0.298270272 0.298270811 No 1919 40.4 1929 30.7 Chr15 Fancm∆2/∆2 1825 39.1 125.5 1921 43.1 1928 43.3 Chr16 Fancm+/+ 1826 38.8 111.4 188.5897274 6.45905E-43 6.45917E-43 Yes 1919 44.4 1929 28.3 Chr16 Fancm∆2/∆2 1825 37.2 131.0 1921 48.4 1928 45.4 Chr17 Fancm+/+ 1826 56.9 152.6 11.67365967 0.000633912 0.000633915 Yes 1919 50.7 1929 45.1 Chr17 Fancm∆2/∆2 1825 48.4 159.5 1921 50.1 1928 60.9 Chr18 Fancm+/+ 1826 37.5 118.5 30.1639027 3.97031E-08 3.97034E-08 Yes 1919 42.3 1929 38.7 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 6 Chr18 Fancm∆2/∆2 1825 51.3 150.5 1921 50.8 1928 48.4 Chr19 Fancm+/+ 1826 46.9 129.3 0.555288399 0.456164872 0.456166412 No 1919 44.1 1929 38.3 Chr19 Fancm∆2/∆2 1825 46.5 140.5 1921 48.0 1928 46.1 814 815 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 1 816 Supplementary Figure 1. SDS -PAGE analysis of Tn5 -intein construct expression and 817 purification. Samples were collected at various stages of expression at 23˚C and purification 818 process and assessed by SDS -PAGE to evaluate the presence and purity of the Tn5 -intein 819 protein. Analysed samples include: molecular weight marker; crude lysate at 0 hours (0 h) and 820 4 hours (4 h) post -induction with IPTG; cell pellet post -sonication (Sonicated); the soluble 821 (Soluble) and insoluble (Insoluble) fraction post -sonication; the soluble fraction treated with 822 polyethyleneimine (soluble+PEI); the soluble fraction treated with PEI and then centrifuged 823 (Soluble+PEI+spin); the Flow-through after loading lysate onto a chitin column (pass through); 824 the wash fraction from chitin column (Wash); the eluted fractions (E1 to E4) from the chitin 825 column. Molecular weight (MW) of the marker in kilodaltons (kDa) shown on the left. The 826 expected size of tagged and cleaved (untagged) Tn5 is indicated on the right.827 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 1 828 Supplementary Figure 2. Workflow of snATAC computational analysis. Nuclei were extracted and sequenced from F1.Fancm+/+ and 829 F1.Fancm∆2/∆2 mice, generated by crossing FVB/N and C57B L/6 inbred strains. Fastq files from ATAC sequencing were processed using Cell 830 Ranger ATAC (10x Genomics) to generate read counts, cell counts, and coverage metric. Peak files were utilised in Seurat for downstream analysis, 831 including filtering and processing to create UMAP plots. Crossover analysis was performed using BAM files output by Cell Ranger ATAC, which 832 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint 2 were filtered to retain cells over 10,000 barcodes. Crossover calling was conducted on the filtered samples using sgcocaller, followed by additional 833 filtering with comapr to ensure sufficient support for crossover identification. These filtered datasets were then used for downstream analysis. 834 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 16, 2024. ; https://doi.org/10.1101/2024.12.16.625960doi: bioRxiv preprint