Evolutionary persistence of a highly prevalent multicopy mitochondrial-derived nuclear insertion (Mega-NUMT) in Neotropical Drosophila flies

39

Although strict maternal transmission of mitochondria is a general feature of animals 40 and humans for ensuring homogeneity in mitochondrial DNA (mtDNA) across generations, 41 exceptions were reported in the recent past. For example, some extremely rare but spectacular cases 42 of heteroplasmy and paternal transmission in humans have questioned the universal evolutionary 43 principle. Hence, as an alternative, the Mega-NUMT concept was coined to explain this discovery 44 and was thereafter partly proven to exist. This co ncept expands on the quite common transfer of 45 mtDNA fragments to the nucleus (NUMTs) by considering the existence of multicopy 46 mitochondrial nuclear insertions. Mega -NUMT reports are currently restricted to a few cases in 47 animals, including humans. However, even in humans, their detailed genomic organization, natural 48 prevalence, and potential biological functions remain unclear. 49 50 Methodology/Principal Findings: Here, we discovered that up to 60 full -sized mitochondrial 51 genomes are integrated into the nuclear genome of the neotropical fruit fly Drosophila paulistorum 52 using long-read sequencing and confirmed their presence by in situ hybridization. The copies are 53 organized in one cluster on chromosome 3, which we, due to its similarity with the Mega -NUMT 54 concept, designated the “Dpau Mega-NUMT”. Contrary to the rarity in humans, this Mega-NUMT 55 is found at high prevalence (40%) in both long -term laboratory lines and natural D. paulistorum 56 populations of different semispecies. Additionally, the mitochondrial cop ies in the Mega -NUMT 57 cluster are phylogenetically separated from the current mitotypes of D. paulistorum . Together, 58 these observations suggest long -term maintenance of the Mega -NUMT in nature. Hence, we 59 propose that the Dpau Mega-NUMT may have been transfe rred to the nuclear genome before D. 60 paulistorum semispecies radiation and maintained at relatively high prevalence in nature by 61 balancing selection due to yet undetermined functions. 62 Conclusions/Significance: To our knowledge, this is the first verified e xistence and detailed 63 dissection of a Mega -NUMT outside cats and humans. We show that Mega -NUMTs can be 64 persistent in nature, even at high prevalence, potentially due to balancing selection. Our findings 65 strengthen the importance of high -quality long -read sequencing technologies for deciphering 66 complex repeat -rich genomic regions to deepen our understanding of the dynamics of genome 67 evolution within genomic “dark matter”. 68 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 3

69 Mitochondrial DNA (mtDNA) has been widely used as a genetic marker for molecular systematics 70 of animals for more than 30 years (Harrison, 1989). The high copy number of mtDNA per cell, as 71 well as its rapid mutation rate, makes it pa rticularly informative for investigating phylogenetic 72 relationships and evolutionary history (Brown et al., 1979) . Mitochondrial markers have been 73 extensively utilized to resolve taxonomic uncertainties and to identify cryptic species (Hebert et 74 al., 2003; Elías -Gutiérrez e t al., 2008) . Additionally, the use of mitochondrial data allows the 75 reconstruction of phylogeographic patterns, providing insights into historical population dynamics 76 and speciation events (Avise et al., 1983; Baião et al., 2023). 77 Studies from the pre-genomic era using Southern blots and PCR have revealed the existence 78 of nuclear mitochondrial DNA segments (NUMTs), described a s fragments of mtDNA that have 79 become integrated into the nuclear genome (Lopez et al., 1994; Hazkani -Covo & Graur, 2007) . 80 Although the mechanisms of integration are not yet fully described, a common explanation outlines 81 the insertion of mitochondrial sequences during the repair of double -strand chromosomal breaks 82 (Blanchard & Schmidt, 1995, 1996; Bensasson et al., 2001). NUMTs have primarily been found in 83 genomic regions with high levels of repetitive elements (Behura, 2007; Tsuji et al., 2012; Dayama 84 et al., 2014; Schiavo et al., 2017; Wang et al., 2020) and distinctive GC content (Tsuji et al., 2012; 85 Dayama et al., 2014; Schiavo et al., 2017). After their integration, NUMTs often degrade quickly, 86 as the different genetic codes in the nuclear and mitochondrial genomes prevent proper translation 87 of the genes (Bensasson et al., 2001). Nonetheless, some functions associated with NUMTs have 88 been found, including the formation of new exons (Noutsos et al., 2007) and regulatory elements 89 (Vendrami et al., 2022). 90 The presence of NUMTs can lead to misinterpretations when studying mitochondrial 91 genomes (Calvignac et al., 2011), as it can result in erroneous inferences of relationships between 92 lineages (Kress et al., 2015; DeSalle & Goldstein, 2019) . Furthermore, the presence of NUMTs 93 may lead to wrongful inference of heteroplasmy (Parr et al., 2006; Balciuniene & Balciunas, 2019), 94 which is the coexistence of more than one mtDNA sequence variant within an individual. The 95 misinterpretations arise from the difficulty in distinguishing NUMTs from real heteroplasmic 96 variants when using techniques such as PCR. However, the high copy number of mtDNA compared 97 to nuclear loci containing NUMTs may commonly suffice to mask the NUMT signal. Even so, one 98 must be cautious (Parfait et al., 1998) , as several studies have reviewed investigations in which 99 high copy numbers of NUMTs in the genome had been interpreted as heteroplasmy (Hirano et al., 100 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 4 1997; Salas et al., 2005; Yao et al., 2006; Lutz-Bonengel et al., 2021). To describe such multicopy 101 NUMTs, the term Mega -NUMT was coined as a theoretical concept in response to a claim of 102 paternal transmission of mitochondria in seemingly heteropla smic humans (Balciuniene & 103 Balciunas, 2019; Lutz -Bonengel et al., 2021) . However, large multicopy NUMTs, covering 104 approximately half of the mt-genome and with an estimated 38-76 copies, had also previously been 105 found in the domestic cat and its close relatives (Lopez et al., 1994) and later also identified in 106 several cat sequences using long -read sequencing, albeit in somewhat fewer copies (Patterson et 107 al., 2023). 108 In this study, we focus on the Neotropical fly Drosophila paulistorum spp., commonly used 109 as a study system for the evolutionary process of sympatric speciation (Burla e t al., 1949; 110 Dobzhansky & Spassky, 1959; Dobzhansky et al., 1964). The D. paulistorum superspecies belongs 111 to the D. willistoni group and is composed of six semispecies, named Amazonian (AM), Andean 112 Brazilian (AB), Centr o-American (CA), Interior (IN), Orinocan (OR), and Transitional (TR), 113 presenting different levels of pre - and post -mating incompatibilities (Dobzhansky & Spassky, 114 1959). Phylogenetic incongruences between the nuclear and mitochondrial DNA of D. paulistorum 115 spp. have been described (Gleason et al., 1998; Robe et al., 2010; Baião et al., 2023) , suggesting 116 past introgression events. Additionally, a recent phylogenomic analysis of D. paulistorum spp. 117 revealed that their mitochondrial genomes are polyphyletic and split into two major clades, α and 118 β (Baião et al., 2023). 119 Here, we identify a D. paulistorum line from the OR semispecies, O11, with an apparently 120 fixed presence of the two different mitotypes, α and β. However, our analysis revealed that in 121 addition to the presence of a β mitochondrion, O11 has a NUMT consisting of around 60 copies of 122 nearly com plete mt -genomes with high similarity to the α mitotype on chromosome 3. We 123 designate this nuclear insertion as the Dpau Mega-NUMT. Furthermore, we found that this Mega-124 NUMT is present in at least one line from three of the six described D. paulistorum semispecies 125 and has a relatively high prevalence of around 40% in fly populations from French Guiana. Taken 126 together, these data suggest that the Dpau Mega-NUMT was present before or just after the split 127 of the semispecies and may be maintained in nature by balancing selection. 128 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 5

129 The two mitotypes, α and β, are fixed in individual flies of the D. paulistorum line O11 130 We sequenced PCR products of the mitochondrial Cytochrome Oxidase I (COI) gene from several 131 D. paulistorum lines for mt -barcoding. We foun d multiple robust double peaks in the resulting 132 chromatograms derived from single flies of the OR line O11, suggesting the presence of both α and 133 β mitotypes within the same individual. The presence of a HindIII restriction site for the α-mitotype 134 but not ß (Fig. 1A) was used to perform diagnostic RFLP/ HindIII Southern blots on DNA from 135 single O11 flies (Junakovic, 2004) probed with the COI cloned plasmid (Fig. 1B and Table S1, 136 Fig. S1A). We observed that the O11 line again showed a clear heteroplasmic pattern, with both α 137 and β mitotypes represented in individual flies at approximately similar abundances, suggesting 138 almost equal copy numbers ( Fig. 1B). In c ontrast, single flies of line C2 belonging to the CA 139 semispecies and line A28 from the AM semispecies were homoplasmic and carried either the α or 140 the β mitotype, respectively ( Fig. 1B). Furthermore, we performed CO1-specific RFLP/HindIII-141 PCR on multiple single females and males (Table S1, Fig. S1) which showed a consistency on the 142 presence of both α and β mitotypes in all tested O11 flies ( Fig. 1C), which suggested fixation in 143 this long-term lab line collected in the 1960s. 144 145 146 147 148 149 150 151 152 153 154 155 Figure 1. Presence of two mitotypes in the D. paulistorum Orinocan line O11. (A) Chromatogram 156 derived from COI direct Sanger sequencing of one O11 female covering the α-diagnostic HindIII restriction 157 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 6 site (underlined). The asterisk highlights the double peak with Guanine (G) for the α- and Adenine (A) for 158 the β-mitotype. (B) Single-fly CO1 Southern blot digested with HindIII and probed with cloned CO1. While 159 the Amazonian females of th e A28 line (n=3) present only the β -mitotype (4.8kb) and the females of the 160 Centro-American line C2 (n=2) only the α -mitotype (2.4kb), all randomly picked Orinocan O11 females 161 (n=3) present both the α- and β-mitotypes (4.8 and 2.4kb). (C) α-diagnostic CO1 RFLP/HindIII-PCR screen 162 (Fig. S1) on O11 single flies (n=20 for each sex). Undigested (U) and digested (D) CO1-PCR fragments per 163 fly are shown exemplary for two indi vidual specimens (magnification at the bottom right in C). All tested 164 single flies (40/40) exhibited clear double bands after HindIII digestion. 165 Subsequently, we measured the copy numbers of both α and β mitochondrial (mt) genomes, relative 166 to the nuclear genome, using discriminant qPCR on the CO3 gene (see Materials and Methods). 167 The quantification of both the α and β mt genomes was in multiple tissues of both females and 168 males and resulted in similar average values of 27.2±13.9 α and 31.6±22.7 β mt genome copies per 169 nuclear genome (Fig. 2). The β mt-genome copy numbers were strongly affected by tissue (Table 170 S2; F-value=27.694, p<0.001, F-test, Fig. 2A, Table S2) and a more subtly, yet significantly, by 171 age and sex combined ( Table S2; F-value=6.66, p<0.05, F-test), and by sex alone ( Table S2; F-172 value=4.714, p<0.05, F-test), as expected based on mitochondrial function. In contrast, the number 173 of copies of the α mt -genome was not significantly affected by any of the tested parameters ( Fig. 174 2B, Table S2), suggesting that the α mitotype is not a typical mitochondrion. 175 176 Figure 2. qPCR quantification of α and β mitotypes in D. paulistorum O11 flies. 120 flies were analyzed 177 to assess the number of α and β mitotype copies relative to the nuclear genome copies using discriminant 178 qPCR. Boxplots show the distribution of values, mean, and standard deviation for the ( A) β mitotype and 179 the (B) α mitotype. 180 181 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 7 Discovery of a Mega-NUMT with long reads 182 Given the seemingly fixed high abundance of two mitotypes in one individuum and unresponsive 183 behaviour of the α mitotype in different tissues of the D. paulistorum O11 line, we sequenced the 184 O11 genome using Oxford Nanopore technology (ONT). The assembly of the nuclear genome of 185 O11 was 231 Mb (111 contigs, N50 = 18.6 Mb) (Table S3), with a BUSCO completeness of 99% 186 (diptera_odb, Table S4). Using the mitochondrial genome from the O11 line (β mitotype, Baião et 187 al., 2023), we searched the nuclear assembly and found 12 mt-genome copies clustered at the end 188 of one contig (ctg000940). From here on, we refer to this cluster as the Dpau Mega-NUMT, given 189 its similarity to the theoretical Mega -NUMT concept. Further analysis of the 12 NUMT copies 190 revealed a conserved synteny of the mitochondrial genes across most copies ( Fig. 3). Four of the 191 NUMT copies have lost genes compared to the mt-genome, and in all NUMT copies, many of the 192 genes have suffered frameshift mutations (Fig. 3, Table S5). When comparing the gene length to 193 the fraction of intact gene copies per protein -coding gene, we found a significant inverse 194 correlation, suggesting that the pseudogenization process is random (Fig. S2). However, given the 195 complexity of the Mega-NUMT, it is possible that the sequence of each gene copy in the assembly 196 is not 100% accurate. Nevertheless, at least one copy of each of the 13 mitochondrial protein -197 coding genes is intact in the Mega-NUMT, albeit with the mitochondrial genetic code (Table S5). 198 A phylogeny based on the c omplete set of genes and pseudogenes in the Dpau Mega-199 NUMT, together with the published mitochondrial genomes of D. paulistorum spp. and other 200 willistoni group species (Baião et al., 2023), shows that all individual NUMT copies in the Mega-201 NUMT are found in one clade, which is sister to the α mitotype of the D. paulistorum C2 and MS 202 lines (Fig. 3) belonging to the CA and AB semispecies, respectively ( Table 1, Table S6). Hence, 203 we refer to these NUMT copies as α-NUMTs. Importantly, though, the separation of the α mitotype 204 and α-NUMTs into different clades, and the monophyly of within these clades, suggests that none 205 of the α -NUMT copies in the Dpau Mega-NUMT are rec ent insertions from an extant α mt -206 genome. This is further supported by the fact that the O11 line carries the β mitotype, as does the 207 majority of D. paulistorum semispecies and lines. 208 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 8 209 Figure 3. Phylogenetic position and synteny of NUMT copies of D. paulistorum O11 and closely related 210 mitochondrial genomes. Maximum likelihood phylogeny based on the complete set of genes and 211 pseudogenes of the NUMT copies found in the Mega -NUMT of D. paulistorum O11, together with the 212 mitochondrial genes in D. paulistorum and closely related species. Black dots indicate nodes with 100 213 bootstrap supports. The α and β mitotype clades are annotated on the tree. Mitochondrial genomes of D. 214 paulistorum O11 (green) and D. paulistorum MS (blue) were used as references for the β and α mitotypes. 215 The grey ribbons between sequences represent homologous regions of the same direction, while red ribbons 216 represent inversions. The 13 mitochondrial protein -coding genes are shown as colored arrows, and rRNA 217 and tRNA genes as blocks. Intact genes are filled, and pseudogenes only have a colored outline. 218 Further, we analyzed the genomic region where the Dpau Mega-NUMT is located. The Mega -219 NUMT is directly flanked by an AT-rich 36 bp satellite sequence (Fig. 4 and Fig. S3) and the gene 220 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 9 FOXO encoding the Forkhead box O transcription factor (Jünger et al., 2003) , located in Muller 221 element E on chromosome 3, as deduced by comparison to D. melanogaster. The same 36 bp 222 satellite DNA is also found inside the Mega -NUMT, downstream of α-NUMT 3 (Fig. 4 and Fig. 223 S3). The positions of FOXO and the Mega-NUMT next to each other were also confirmed by DNA-224 hybridizations on polytene chromosomes of larval salivary glands of O11, and the bright 225 polytenized signal of the Mega -NUMT ( Fig. S4 ) implies its euchromatic state since it is 226 polytenized and in close genomic vicinity to the essential host gene FOXO. 227 228 Figure 4. Graphic representation of the end of contig000940 in the assembly of D. paulistorum O11. 229 The region is divided into three rows to facilitate visualization. The position of the nuclear gene FOXO and 230 each α-NUMT is represented with a white box, which also contains colored lines that represent the position 231 of each protein-coding and RNA gene. The positions of the satellite and repeat elements are shown below 232 in the order: The AT-rich 36-bp satellite, Gypsy, TcMariner, Helitron, Copia-like transposons, and simple 233 repeats. 234 All α-NUMT copies except α-NUMT 3 have the same orientation in the Dpau Mega-NUMT (Fig. 235 4). The sequence of α-NUMT 3 also contains a rearrangement not seen in any of the other copies 236 of the contig ( Fig. 3 and 4). Additionally, the 12 α -NUMT copies in the O11 nuclear genome 237 assembly are surrounded by repeat elements, and in some cases, insertions of transposable elements 238 are present within an α -NUMT sequence ( Fig. 4). Specifically, we found remnants o f the LTR 239 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 10 retrotransposon Gypsy within the α-NUMT copies 6, 9, and 11 and of the LTR element Copia in 240 α-NUMT copies 5 and 9 ( Fig. 4). In addition, we found short regions with sequence similarity to 241 the DNA transposon Tc Mariner in six of the α-NUMTs and to the rolling-circle DNA transposable 242 element, Helitron, in nine of the α -NUMTs, as well as between α -NUMT copies when a non -243 NUMT sequence was found between them (Fig. 4). 244 Number of α-NUMT copies in the genome 245 Since the discriminant qPCR results indicated that more than 12 α-NUMT copies could be present, 246 we investigated the coverage of ONT reads in the Mega -NUMT region of O11 to assess the true 247 number of α-NUMT copies. When mapping the reads to the assembly, we also included the β mt -248 genome from Baião et al., (2023) and the genome of the bacterial symbiont Wolbachia of O11 from 249 Papachristos et al., (2025) to prevent mapping of reads from the real mitochondrion or Wolbachia 250 to the Mega -NUMT. Additionally, to be as stringent as possible in our assessment, only one 251 mapping per read was allowed. Th e read coverage over the Dpau Mega-NUMT shows a clear 252 discordance with the average coverage across the genome (Fig. 5). The Dpau Mega-NUMT has an 253 average coverage of 402±159, while the genome assembly as a whole has an average coverage of 254 80.8±77.4 (Fig. 5). Using these numbers, we thus infer that the genome contains around 60 whole 255 mt-genome α-NUMTs (Fig. 5). We note that the coverage over the different α-NUMT copies varies 256 substantially. Especially high coverage was found on α -NUMT 8, with 17 times higher coverage 257 than the genome average, followed by α -NUMTs 7, 9, and 10, with ten, seven -, and six -times 258 higher coverage than the genome average, respectively. These results indicate that the unassembled 259 α-NUMT copies resemble α-NUMT copies 7-10 the most. Specifically, we note that both α-NUMT 260 7 and 8 contain the full mt-genome and have no intervening sequence, suggesting that many of the 261 unresolved copies are also full mt-genomes with no intervening sequences (Figs. 4 and 5). 262 We note that the 60 α-NUMT copies estimated from long read mapping are more than the 263 ca. 30 we estimated by our qPCR results. As α-NUMT 3 lacks CO3 completely, only 11 of the 12 264 copies in our assembly have a binding site for the primer CO3_SNP_1 used in qP CR for 265 determining the copy numbers (Fig. 2, Table S1). Hence, although we do not know exactly what 266 the remaining copies look like, this indicates that the discrepancy in total copy numbers of the α -267 NUMT between the coverage and qPCR results might be explained by a lack of primer binding in 268 some of the α-NUMT copies. 269 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 11 270 Figure 5. Coverage quantification of α-NUMTs in the nuclear genome of D. paulistorum O11. The x-271 axis represents the end region of contig000940 of the genome assembly of D. paulistorum O11, which 272 includes the nuclear gene FOXO, and the 12 assembled α-NUMTs. Coverage was obtained from mapping 273 Nanopore reads to the genome assembly. The average and standard deviation of the coverage over each α-274 NUMT and the FOXO gene are plotted in the same color as used for the box depicting each feature. The 275 total number of α-NUMT copies in the Mega-NUMT is estimated from the average over all the assembled 276 α-NUMT copies in relation to the average coverage of the whole genome assembly. 277 Mega-NUMT visualization on D. paulistorum chromosome. 278 To validate the genomic findings of the Mega -NUMT localization on the 3 rd chromosome of D. 279 paulistorum O11, we performed DNA fluorescent in situ hybridization (FISH) on mitotic 280 chromosomes of larval brains us ing mitochondria -specific probes and a probe for the Mega -281 NUMT-linked satellite (see Fig. 4 and the Materials and Methods section). In O11, we observed 282 two satellite signals (red) on the 3rd chromosome—one in the (peri)centromeric region and a second 283 in the subtelomeric region—as well as a Mega-NUMT signal (green) adjacent to the subtelomeric 284 satellite signal ( Fig. 6A ), in agreement with the assembly data ( Fig. 4 ). In contrast, the D. 285 paulistorum line FG295 (Table S6), which also belongs to the OR semispecies ( Table 1), shows 286 only a single satellite signal in the (peri)centromeric region of the third chromosome, but no Mega-287 NUMT signal (Figure 6B). 288 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 12 289 Figure 6. DNA -FISH on mitotic chromosomes from larval brains of the D. paulistorum O11 and 290 FG295 lines. Probes targeting the Mega -NUMT (green) and the Mega -NUMT–linked satellite (red) were 291 used. The scale bar represents 5 μm. The inset highlights the third chromosome, showing two satellite 292 signals and one Mega -NUMT signal in O11 ( A), but only a single satellite signal in FG295 ( B). The 293 schematic on the right summarizes the localization of the satellite (red) and the Mega -NUMT (green) in 294 each strain. 295 Additionally, to follow Mega -NUMT dynamics in interphase and dividing cell s, we 296 investigated the localization of the Dpau Mega-NUMT in early embryonic mitosis by mt -DNA 297 FISH ( Fig. S5 ). Contrary to larval brains that were generated by squash fixation and thereby 298 washing off the cytoplasm ( Fig. 6), in our whole mount confocal mt -DNA-FISH assays on O11 299 embryos, we observed staining of mitochondria in the cytoplasm and the chromosome -associated 300 Mega-NUMT signal as two dots, with some cases of a fused, larger dot (arrowheads and arrow, 301 respectively, Fig. S5A -D, F ). Each dot consists o f two parts with a small bridge in between, 302 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 13 suggesting somatic pairing (inset of Fig. S5B). Based on our results on metaphase chromosomes 303 from 3rd instar larvae (Fig. 6), we conclude that each dot represents a Mega-NUMT signal located 304 on one of the duplicated arms of chromosome 3, with two smaller parts representing sister 305 chromatids. At the anaphase stage, we could clearly observe the separation of sister chromatid 306 signals, creating a 4-dot pattern (Fig. S5E). Altogether, systematic DNA-FISH at different stages 307 of early embryo mitosis reconfirmed the presence of only one Mega-NUMT signal on chromosome 308 3 of the O11 line. 309 310 Prevalence of the Mega-NUMT across D. paulistorum semispecies and populations 311 To test if the Dpau Mega-NUMT is restricted to only the O11 line, we analyzed α -NUMT 312 prevalence across different D. paulistorum lines. We used RFLP-PCR from adult flies to assess the 313 prevalence and abundance of α -NUMTs (Table 1). In addition to O11, another OR line, FG111, 314 presented α-NUMT, but it appeared not to be fixed in the population, as it was only found in 10 315 out of 40 flies that originated from an isofemale ( Table 1). In yet another OR line, FG103, the 316 Mega-NUMT was lost between the first test (2014-15) and the last one (2018) (Table 1). Two AM 317 lines, FG16 and FG572, also presented a non-fixed Mega-NUMT (Table 1), while two lines from 318 the AB semispecies, MS and Yellow, both carried the Mega -NUMT in all tested flies ( Table 1). 319 To sum up, we found the Mega -NUMT in several OR lines, as well as in the AM and AB 320 semispecies, but not in the few representatives of the other three semispecies of D. paulistorum 321 (CA, IN and TR). Possibly, this is caused by a sampling bias, since none of these three semispecies 322 were found in our collections from French Guiana between 2014 and 2018 ( Table 1 and Table 323 S7). In addition to O11, we only found fixation of the Dpau Mega-NUMT in the old long -term 324 stocks of MS and Yellow that were maintained at small population sizes in the laboratory since the 325 1960s or earlier. Our finding that the Mega-NUMT is not fixed in isofemale lines from our recent 326 collections in French Guiana suggests that it is frequently hemizygous in wild populations. 327 Table 1. α-NUMT abundance and prevalence in D. paulistorum lines based on CO1 RFLP-PCR from adult flies. 328 Abundance is given in percentage of α-NUMTs compared to the mitochondrial copies per fly (Fig S1) and is presented 329 separately for females (left) and males (right). Prevalence of the Mega-NUMT among tested individuals is presented 330 both as a percentage and in absolute numbers (in parentheses). Abbreviations: Spp: Semispecies, AB: Andean -331 Brazilian, AM: Amazonian, CA: Centro -American, IN: Interior, OR: Orinocan, TR: Transitional, Coll: Collection 332 year, ND: Not determined. Further details about the fly lines are provided in Table S6. 333 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 14 Line Coll. Spp. Mitotype α-NUMT detected α-NUMT abundance (%)# α-NUMT prevalence O11 1957 OR ß Yes 30.7±3.7/36.8±3.6 100% (39/39) FG103 2014 OR ß Yes ND§ 0% (0/40)§ FG111 2014 OR ß Yes 13.4±1.1/18.3±2.9 25% (10/40) FG295 2014 OR ß - 0 0% (0/40) TP37 2009 OR ß - 0 0% (0/20) POA1 2003 OR ß - 0 0% (0/20) RP 1995 OR ß - 0 0% (0/20) A28 1952 AM ß - 0 0% (0/40) FG16 2018 AM ß Yes 57.2±8.6/66.5±7.2 55% (22/40) FG18 2018 AM ß - 0 0% (0/40) FG572 2015 AM ß Yes 45.3±8.1/63.7±7.0 90% (36/40) MS* 1962 AB α Yes 21.8±3.1/36.0±4.1 100% (40/40) Yellow* 1960 AB α Yes 20.4±4.7/27.4±4.3 100% (40/40) C2 1954 CA α - 0 0% (0/20) White 1960 CA α - 0 0% (0/20) L1 1958 IN ß - 0 0% (0/20) SM 1956 TR ß - 0 0% (0/20) #Abundance of the α-NUMT is given in percentage of females and males determined by CO1 RFLP-PCR. 334 §The line FG103 presented α-NUMT in 2014/15 (determined by Sanger and Illumina sequencing) but has lost it 335 afterwards (verified by CO1 RFLP-PCR and Sanger sequencing in 2018/19). 336 *To distinguish αNUMTs from the α-mitotype of AB semispecies, the α-diagnostic EcoRV site of CO1 was used. 337 To evaluate the prevalence of the Dpau Mega-NUMT in nature, we screened a total of 80 D. 338 paulistorum spp. isofemale lines at generation F1 post-collection that were sampled between 2014 339 and 2024 from five different locations in French Guiana (Fig. 7; Table S7). 340 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 15 341 Fig. 7. Natural prevalence of Mega-NUMTs in D. paulistorum spp. populations in French Guiana. All 342 80 fly samples were collected between 2014 and 2024 at five different locations (red dots, and Table S7). 343 The individual pie charts are in reference to locality and collection date and show the actual number of 344 Mega-NUMT-positive (green) and negative ( gray) specimens. Species designation and Mega -NUMT-345 typing per established D. paulistorum spp isoline (n=80) was performed on multi-fly DNA at generation F1 346 post-collection by CO1 PCR and direct sequencing. The presence of robust double peaks in their 347 chromatograms at α-NUMT-diagnostic positions was used in combination with cloning and sequencing. 348 349 Sampled flies were established as isofemale lines and barcoded at F1 post -collection as 350 members of the D. paulistorum species complex by CO1 PCR followed by direc t Sanger 351 sequencing and/or cloning of CO1 (Table S7). Similar to O11 ( Fig 1A), the presence of robust 352 double peaks in the chromatograms from direct sequencing was used for Mega -NUMT typing, 353 suggesting that 34 of the 80 isofemale lines harbor α-NUMTs (Table S7). The high prevalence of 354 α-NUMT of 43% in our recent collections from D. paulistorum populations, together with its 355 presence in at least three of the six semispecies ( Table 1 ), strongly suggests their ancestral 356 evolutionary state in this neotropical species complex and its stable maintenance both in nature and 357 over long evolutionary time. 358 359 360 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 16

361 In animal species, mtDNA is normally transmitted strictly uniparentally from mother to progeny 362 (Ladoukakis & Zouros, 2017) , with rare e xceptions (Breton et al., 2007) . However, reports of 363 mitochondrial heteroplasmy and biparental transmission in humans challenged this view (Schwartz 364 & Vissing, 2002; Luo et al., 2018) and prompted skepticism by the scientific community 365 (McWilliams & Suomalaine n, 2019; Salas et al., 2020) . Consequently, the theoretical Mega -366 NUMT concept, proposing nuclear integration of multiple copies of whole mitochondrial genomes, 367 was coined as an alternative explanation for such unorthodox findings (Balciuniene & Balciunas, 368 2019). Although this concept has been verified in humans (Wei et al., 2020; Lutz-Bonengel et al., 369 2021; Pagnamenta et al., 2021; Wei et al., 2022) , similar phenomena have so far only been 370 described in a limited number of animal systems where only partial regions of the mt-genome were 371 found at multicopy (Lopez et al., 1994; Patterson et al., 2023; Biró et al., 2024) , but not in 372 Drosophila (Rogers & Griffiths -Jones, 2012), despite a relatively high prevalence of NUMTs in 373 insects (Hebert et al., 2023) . On the other hand, nuclear insertions of whole bacterial genomes 374 larger than 1 Mb, like the α -proteobacterium Wolbachia, have been repeatedly detected in 375 invertebrate genomes, including insects a nd nematodes (Kondo et al., 2002; Hotopp et al., 2007; 376 Nikoh et al., 2008; Klasson et al., 2014; and recently reviewed in Keeling, 2024). 377 The Mega-NUMT of D. paulistorum spp. 378 In this study, we found that multiple whole mitochondrial genomes are integrated in the nuclear 379 genome of the O11 line from the D. paulistorum species complex, designated α-NUMTs because 380 of their close phylogenetic relationship with the cytop lasmic α-mitotype of the Centro -American 381 and Andean-Brazilian semispecies of D. paulistorum (Baião et al., 2023). These α-NUMTs localize 382 in a nuclear cluster in the Müller element E on chromosome 3 with up to 60 nearly complete 15kb 383 mitochondrial genomes, potentially with a total size of 900 kb. Given the agreement between this 384 finding and the Mega-NUMT concept, we named this cluster the “Dpau Mega-NUMT”. Similar to 385 earlier NUMT reports (Behura, 2007; Tsuji et al., 2012; Dayama et al., 2014; Schiavo et al., 2017; 386 Wang et al., 2020; Biró et al., 2024), the Dpau Mega-NUMT of the O11 line is located in a genomic 387 region containing other repetitive elements, i.e., a chromosome 3 -specific satellite and several 388 different TEs ( Fig. 6A, B ), which suggests it might have served as a “s afe haven” for such 389 sequences (Werren, 2011). Such repetitive sequences may have enlarged the Dpau Mega-NUMT 390 beyond 1 Mb in size, which is larger than the 641 kb NUMT recently discovered in the genome of 391 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 17 Arabidopsis thaliana, consisting of a partially duplicated large mitochondrial genome (Fields et 392 al., 2022) and curren tly one of the largest described fully sequenced NUMTs. Additionally, 393 contrary to currently described human Mega -NUMTs with prevalence below 0,5% (Wei et al., 394 2020; Bai et al., 2021; Lutz -Bonengel et al., 2021) , the Dpau Mega-NUMT is found at high 395 prevalence, around 40% in long -term laboratory lines (3/8; Table 1 ) and 42.5% (34/80) in D. 396 paulistorum spp. flies collected between 2014 and 2024 at different locations in French Guiana 397 (Fig. 7; Table S7). Furthermore, the Dpau Mega-NUMT is present in members of at least three of 398 the six semispecies of D. paulistorum, either in homo- or hemizygosity (Table 1). Finally, the α-399 NUMTs form a separate monophyletic clade to th e exclusion of the current α mitotype ( Fig. 3), 400 indicating that the insertion did not involve any of the extant mitotypes but possibly a more 401 ancestral one. Therefore, we proposed that the Dpau Mega-NUMT might have been inserted in the 402 nuclear genome before or just after the semispecies radiation, estimated to have occurred between 403 1.2-2.1 MYs ago (Zanini et al., 2018; Kim et al., 2022). 404 Integration and chromosomal position 405 The overall well-conserved head-to-tail structure of the D. paulistorum α-NUMT cluster suggests 406 that the Dpau Mega-NUMT might have integrated as a concatemer of multiple mt -genomes. As 407 earlier proposed by Balciuniene & Balciunas, 2019, the mechanism of generating Mega -NUMTs 408 of complete mt -genomes may be explained by replication via a rolli ng circle intermediate. Such 409 mt-rolling circles have, for example, been reported in yeast, nematodes, and humans (Pohjoismäki 410 et al., 2009; Ling et al., 2016; Ling & Yoshida, 2020) , but not in Drosophila until now. The 411 switching between conventional circular to rolling circle replication might be quite feasible because 412 key components of the mitochondrial transcription and replication apparatus are derived from the 413 T-odd lineage of bacteriophage rather than from an α -Proteobacterium, as the endosymbiont 414 hypothesis would predict (Filée et al., 2002; Shutt & Gray, 2006) . In addition, rolling circle 415 replication (RCR) has been described in Drosophila for histone and the Stellate and Suppressor of 416 Stellate (Su(Ste)) genes (Cohen et al., 2005; Cohen & Segal, 2009). Moreover, Helitron transposons 417 like DINE-1 elements that can reach extremely high copy numbers in D. melanogaster (Thomas et 418 al., 2014) also propagate via RCR (reviewed in Barro-Trastoy & Köhler, 2024). 419 The nuclear integration site of the mt -concatemer in the chromosome 3 of D. paulistorum 420 spp. is interesting, since it is a fusion between Muller elements E and F, orthologous to 421 chromosome arm 3R and chromosome 4 of the melanogaster group, respectively (Muller, 1940; 422 Spassky & Dobzhansky, 1950). The location of the Dpau Mega-NUMT in the O11 line next to the 423 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 18 satellite DNA ( Fig 6A ), also present in the acrocenter of the fusion autosome E/F ( Fig. 6B ), 424 potentially suggests a chromosomal inversion event splitting the satellite in two, but not affecting 425 the upstream gene FOXO. Such paracentric inversions are quite commonly found in the D. 426 paulistorum species complex (Kastritsis, 1967) and can even play adaptive roles (Krimbas & 427 Powell, 1992), like for the successful recent colonization of urban environments (Valiati & Valente, 428 1997). In addition, inversions are enriched on chromosome 3 and often maintained in 429 heterozygosity, potentially under balancing selection (Gromko & Richmond, 1978). These earlier 430 findings suggest that chromosome 3 of D. paulistorum spp. may be an evolutionary hotspot for 431 chromosomal rearrangements lik e inversions and translocations, but also mitochondrial -derived 432 insertions like the Dpau Mega-NUMT as reported here. 433 Why is the Dpau Mega-NUMT maintained at high prevalence and over evolutionary time? 434 Contrary to currently described human Mega-NUMTs, which are extremely rare in prevalence, the 435 Dpau Mega-NUMT appears to be evolutionarily persistent since it is present in lines from at least 436 three different D. paulistorum semispecies, i.e., Amazon ian, Andean Brazilian and Orinocan. 437 Additionally, it is present at a relatively high prevalence of potentially more than 40% in natural 438 populations ( Figure 7 ; Table S7 ). Even so, the fact that the Mega -NUMT is not fixed in D. 439 paulistorum except for old laboratory stocks ( Fig. 1; Table 1), and that natural populations are 440 mainly hemizygous, suggests balancing selection rather than an essential function. 441 Although almost half of the mitochondrial genes in Mega-NUMT present in the 12 resolved 442 α-NUMT copies in our assembly (Table S5) still appear to encode intact ORFs (albeit with the mt 443 genetic code), we believe that they are unlikely to be under selection for having the same function 444 as when they were present in the mt -genome. Mainly because the mt genetic code will cause 445 problems during translation, but also since the pseudogenizing mutations appear to have occurred 446 randomly throughout the genes ( Fig. S2), suggesting neutrality. However, we can’t exclude that 447 some of the DNA in the Mega-NUMT might be transcribed and serve yet undetermined functions. 448 Potentially, the TEs within and between α -NUMT units, especially remnants of LTR 449 retrotransposons like Copia and Gypsy (Fig. 4), may contribute to expression, as LTRs are well 450 known to harbor cell - and tissue -specific cis-regulatory elem ents that affect expression of 451 neighboring host genes (McDonald et al., 1997) . Alternatively, TEs can also create de novo 452 regulatory regions acting in trans , so -called piRNA clusters (recently reviewed in Pritam and 453 Signor, 2025 , which, if expressed in the antisense orientation of the transposon, can serve as 454 efficient epigenetic silencers of homologous TEs via the piRNA pathway (Brennecke et al., 2008). 455 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 19 Indeed, de novo creation of such a repressive piRNA cluster has recently been demonstrated in D. 456 simulans (Rafanel et al., 2025). 457 Another possibility is that the Mega -NUMT provides a structural function. Other studies 458 have seen that NUMTs are located close to centromeric or telomeric regions (Puertas & González-459 Sánchez, 2020) , suggesting that they might be involved in their function. However, the Dpau 460 Mega-NUMT of O11 is not found in either the centromeric or telomeric chromosomal region of 461 chromosome 3, according to our FISH. Even so, a yet undetermined structural role for th e Dpau 462 Mega-NUMT is still possible. 463 Further experiments will be necessary to elucidate any potential function and fitness 464 benefits of the Mega-NUMT in the D. paulistorum species complex. 465 466

467 To our knowledge, this is the first verified existence a nd detailed dissection of a Mega -NUMT 468 outside cats and humans. Moreover, we show that Mega-NUMTs can persist at high prevalence in 469 nature and over relatively long time periods, suggesting balancing selection. Our findings 470 strengthen the importance of high -quality long -read sequencing technologies for deciphering 471 complex repeat-rich genomic regions and mobile DNAs to deepen our understanding of genomic 472 “dark matter”. Finally, it is expected that with the rapidly increasing number of high -quality 473 genomes, the Dpau Mega-NUMTs will not remain a unicorn. Instead, similar Mega -NUMTs will 474 very likely be found in further eukaryotic genomes. 475 476 Data availability 477 All data will be available upon publication. 478 479

480 We thank Tom Martin for performing the DNA e xtraction for Nanopore sequencing of D. 481 paulistorum O11. ONT sequencing of O11 was performed by the Uppsala Genome Center (UGC) 482 in Uppsala, Sweden. The facility is part of the National Genomics Infrastructure (NGI) Sweden 483 and the Science for Life Laboratory. Some of the data handling was enabled by resources provided 484 by the Swedish National Infrastructure for Computing (SNIC) at UPPMAX. UGC and UPPMAX 485 are supported by the Swedish Research Council. This work was supported by the Swedish Research 486 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 20 Council VR grant 2014–4353 to LK and by the Austrian Science Fund FWF grants P28255B22, 487 FW613A0501 and FW613A0502 to WJM. We also thank the Nouragues research field station 488 (managed by CNRS), which benefits from “Investissement d'Avenir” grants managed by Agence 489 Nationale de la Recherhe (AnaEE France ANR-11-INBS-0001; Labex CEBA ANR-10-LABX-25-490 01) to WJM. 491 492 Contributions 493 LK and WJM conceiv ed and designed the study. MMN performed assemblies, annotations and 494 phylogenetic analyses. EH performed qPCRs. JT and DIS performed line screening and RFLP 495 analyses. AS and CC performed FISH. AH -V and WJM collected flies. LK performed 496 bioinformatic analyses. MMN, WJM and LK wrote the paper with contributions from CC and AS. 497 All authors read and approved the manuscript. 498 499

500 Fly strains 501 All fly lines used in the present study are listed in Table S6, including their collection date and 502 semispecies designation. All flies were kept at 25±1°C on Formula 4-24 Drosophila instant food 503 (Carolina, USA) with a 12 hrs. light-dark cycle. 504 Drosophila collections from French Guiana 2014-2024 505 Field samples of neotropical Drosophila specimens were collected in French Guiana at five 506 different geographic locations between 2014 and 2024 ( Table S7 ) by trapping live flies with 507 banana/yeast baits in plastic PET bottles over a 48 hrs. time period. Taxonomic grouping - at least 508 at the species-group level - was performed by eye under the stereo loupe by using FlyNap. Trapped 509 inseminated single females were isolated into individual small single fly culture vials supplemented 510 with Carolina Instant Fly -Food, shipped back to the Medical University of Vienna to establish 511 consecutive isofemale lines in WM´s lab. Upon egg deposition and the emergence of F1 larvae, 512 DNA was isolated from the G0 mother for single fly DNA extraction with the Gentra Puregene 513 Tissue kit (Qiagen, Germany) and mitochondrial CO1-barcoding. PCR was performed with the 514 primer pair CO1-F and CO1-univR (Table S1), followed by direct Sanger sequencing and cloning 515 (for further details, see Madi-Ravazzi et al., 2021). 516 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 21 This study was conducted under ABS Permits No. ABSCH -IRCC-FR-245930-1 and ABSCH -517 IRCC-FR-257164-1 issued by the French Ministry of Ecological Transition on 25 Jun 2018 and 16 518 Sep 2021 respectively, in compliance with the Nagoya Protocol and French Biodiversity Law 519 (2016). 520 Single fly Southern blots 521 Single fly Southern blots were performed following the protocol by Junakovic, 2004. In brief, DNA 522 was extracted from single individuals, digested with the restriction enzyme HindIII and processed 523 for electrophoresis on a vertical 0.8% agarose gel. After vacuum-blotting onto a positively charged 524 nylon membrane, fragments were hybridized with a P32-labelled mitochondrial CO1 probe derived 525 from a plasmid. Detection was performed by exposing the membranes to X-ray films at -80°C. 526 Restriction Fragment Length Polymorphism-PCR (RFLP-PCR) 527 Total DNA was extracted from single or multiple flies using the Gentra Puregene Tissue kit 528 (Qiagen, Germany). Flies were homogenized using a TissueLyser LT (Qiagen, Germany) at 50 Hz 529 for 50 sec and DNA was consequently isolated following the manufacturer’s protocol. DNA 530 concentration was determined with a NanoDrop OneC Spectrophotometer (Thermo Scientific, 531 USA). RFLP-PCR for the mitochondrial Cytochrome c oxidase genes CO1 and CO2 (Figure S1) 532 was set up in 20 µl reactions containing 1x Promega reaction buffer, 2.5 mM MgCl 2, 0.5 μM of 533 each primer ( Table S1), 200 μM of each dNTP, and 0.025 U of GoTaq G2 DNA Polymerase. 534 CO1/CO2 reactions were performed using the following thermal profile: initial denaturation for 2 535 min at 95°C followed by 30 cycle s consisting of 45 sec denaturation at 95°C, 45 sec annealing at 536 the appropriate primer conditions Tm°C (Table S1) and 30 sec extension at 72°C. Final extension 537 was carried out at 72°C for 10 min. PCR products were consequently used for RFLP (digest plus 538 undigested control) using the restriction enzymes HindIII and EcoRV-HF for CO1, and MspI for 539 CO2 (Figure S1). Reactions were performed in 20 µl volume containing 10 or 20 U of enzyme, 1x 540 reaction buffer and 10 µl PCR product. After overnight incubation at 37°C, enzymes were 541 inactivated by either heat (80°C for 20 min) or by adding EDTA-containing gel loading dye. 10 μl 542 of undigested and digested samples were then run on a 1.5% agarose gel and post -stained with 10 543 mg/ml ethidium bromide. Imaging was performed using a Molecular Imager ChemiDocTM XRS 544 Imaging System and the intensities of obtained PCR bands were quantified using Image Lab 5.2.1 545 software (Bio-Rad). 546 Quantifying Mega-NUMT abundance within a fly 547 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 22 Undigested samples representing the sum of both mitotypes were set as a reference band with 548 relative quantity 1. The relative levels of mitotypes were then quantified b y summing up the two 549 digested fragments plus the undigested fragment. Due to technical differences in band intensity, 550 however, the sum of all fragments did not result in a relative quantity of 1. Hence, the actual sum 551 value was set to 100% to determine the percentage of mitotype levels. Smaller fragments with a 552 length up to approximately 100 bp were not detectable after ethidium bromide staining and we 553 hence used the correction factor C f = undigested [bp] / digested [bp] to estimate their relative 554 quantities. 555 Quantitative PCR (qPCR) 556 DNA was extracted by homogenizing whole flies in 200µl Chelex solution at 5% concentration in 557 distilled water. A total of 10µl of proteinase K (10mg/ml) was added per sample and the vials were 558 incubated for 6 hours at 56°C. After digestion, vials were centrifuged at 16,200 x g for 3 minutes, 559 and the supernatant was transferred to a new vial. DNA was stored at 4°C and freeze -thaw cycles 560 were avoided when possible. Samples were run in duplicate using SYBR Green reagent on a 561 StepOnePlus Real Time PCR machine (Applied Biosystems, USA). Technical repeats with >1 Ct 562 difference were rerun or discarded. Each reaction contained 10µl Maxima SYBR Green/ROX 563 qPCR Master Mix (2x), 1µl of forward primer, 1µl of reverse primer, 6µl nuclease-free water and 564 2µl of the DNA sample. Discriminant mitochondrial primers cycling conditions were 95°C for 20 565 seconds, followed by 40 cycles of: 95°C for 3 seconds and 55°C for 30 seconds ( Table S1 ). 566 Discriminant primers used linked base technology to increase specificity (Exiqon, Denmark). Melt 567 curves and the fly control gene rps17 were run on each plate to control for successful single -peak 568 amplification and to control for DNA quality. All primers had >90% efficiency. For absolute 569 quantification, each 96 -well qPCR plate was analyzed using StepOne Software v2.2.2 (Applied 570 Biosystems, USA) and Ct values were obtained by comparing each primer sample to a single 571 standard curve of known concentration and using identical threshold and baseline levels for each 572 primer target across plates. Standard curves were created by amplifying positive control samples 573 using PCR, calculating DNA concentrations using QBIT, and then serially diluting the sample 1:10 574 with distilled water to create a 5-sample curve comprising known concentrations decreasing from 575 10pmol/ ml. Samples with Ct values over 30 were classed as negative, confirmed by our negative 576 controls. As a result, the detection limit was 50 copies per sample for rps17, 115 for CO3 577 (coxSNP_1) and 735 for CO3 (coxSNP_2). DNA concentration was calculated for each sample by 578 comparing the sample Ct values back to the equation of the standard curve. DNA copy numbers 579 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 23 were then calculated by using the DNA product length to determine DNA concentration. To control 580 for fly size and extraction efficiency, copy numbers for each sample were normalized using the fly 581 housekeeping gene as a control, giving a final value in terms of mitochondrial genome copies to 582 fly nuclear genome copy ratio. Since rps17 is located on Muller element D, which in D. 583 paulistorum is fused to Muller element A and part of the X chromosome, the copy numbers differ 584 between males and females. Hence, male rps17 numbers were doubled. We used ANOVA in R 585 version 3.3.2 (R Core Team, 2016) when testing for significant changes in copy numbers of the α-586 NUMT and the real β mitochondrial genome between ages and sexes of different tissues. The data 587 was log-transformed to increase normality. 588 DNA extraction, genome sequencing and assembly 589 DNA extractions and Nanopore sequencing 590 Total DNA for Oxford Nanopore (ONT) sequencing of O11 was extracted from many adult flies 591 of both sexes using the MagAttract HMW DNA Kit (Qiagen, Germany) with only minor 592 modifications of the manufacturer’s instructions. Several similar independent DNA ext ractions 593 were then pooled, cleaned and concentrated with the Ampure XP beads clean -up protocol 594 (Beckman Coulter, USA). The Circulomics SRE XS kit (Pacific Biosciences, USA) was then used 595 to eliminate short DNA fragments before sequencing. A DNA library was prepared using the 596 Ligation Sequencing kit (Oxford Nanopore Technologies, UK) and sequenced on an R9.4 597 PromethION flowcell (FLO -PRO002) at the Uppsala Genome Center, Uppsala, Sweden. 598 Basecalling was performed with Guppy 4.3.4 and the HAC model. A total of 1.4M reads were 599 generated, with 19.54 Gb passed bases (Table S3). 600 D. paulistorum O11 genome assembly 601 Reads were subsampled with Filtlong v.0.2.1 (https://github.com/rrwick/Filtlong, last visited 18th 602 March 2022) to lower the overall coverage of the reads to facilitate the assembly process. After 603 testing different settings (Table S3), the O11 Nanopore reads were filtered to a coverage of 50X, 604 giving priority to the quality of the reads ( --target_bases 12500000000 –min_length 1000 –605 mean_q_weight 10). The whole genome assembly was built from the dataset of subsampled reads 606 using NextDenovo v.2.5.0 (Hu et al., 2024; https://github.com/Nextomics/NextDenovo, last visited 607 18th March 2022) with a read cutoff of 1 Kb, and setting the genome size to 250 Mb. The genome 608 assembly was then polished with both the same Nanopore reads and Illumina reads from Baião et 609 al., 2023 (available via NCBI Bioproject PRJNA643793). First, Nanopore reads were mapped to 610 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 24 the genome assembly with Minimap2 (Li, 2018) using the map-ont mode and bam files were sorted 611 and indexed with samtools (Danecek et al., 2021) . Then, the pipeline PEPPER -Margin-612 DeepVariant r0.4 (Shafin et al., 2021) was used to call variants. The variants were filtered with 613 bcftools v1.15.1 -21 (Danecek et al., 2021) using the following threshol ds: QUAL>=30 && 614 FMT/DP>=10 && FMT/GQ>=30 && FMT/VAF>=0.8, and remaining SNPs were corrected in 615 the genome with bcftools v1.15.1 -21 (Danecek et al., 2021) . Illumina reads were aligned to the 616 genome assembly using BWA mem v0.7.17-r1198-dirty (Li, 2013) and sorted with samtools v1.14 617 (Danecek et al., 2021), and two runs of Pilon v1.24 (Walker et al., 2014) were used for correction. 618 BUSCO v5.2.2 with diptera_odb10 and augustus species fly was run to check genome assembly 619 completeness, and Quast v5.0.2 (Mikheenko et al., 2018) was run with defau lt settings to 620 quantitatively assess the genome assembly stats. Scripts can be found in 621 https://github.com/mmontonerin/Dpaulistorum_Mega-NUMT. 622 Detection and annotation of α-NUMT copies in the genome assembly 623 The O11 β mt-genome (Baião et al., 2023) was used as a query when running BLASTn v2.12.0+ 624 (Camacho et al., 2009) against the genome assembly. Hits over 500 bp long and with 90% identity 625 to the mitochondrial sequence were inspected, and the 12 mt -genome copies in the Mega-NUMT 626 region were identified on contig ctg000940. The position of each gene in the α-NUMT copy in the 627 Mega-NUMT was obtained by using gene sequences from the D. willistoni mt-genome (GD-H4-1 628 line) as queries in a BLASTn search. Each inferred gene was subsequently checked manually in 629 Artemis v.16 (Rutherford et al., 2000) to identify pseudogenes and adjust the position if needed. 630 Scripts can be found in https://github.com/mmontonerin/Dpaulistorum_Mega-NUMT. 631 Phylogenomic analysis of the α-NUMT copies 632 The 13 mitochondrial protein-coding genes from all α -NUMT copies and the genes from the real 633 mt-genomes from several species of the willistoni group (see Fig. 3), previously published in Baião 634 et al., 2023 (available via NCBI Bioproject PRJNA643793), were aligned using MAFFT v.7.407 635 (Katoh & Standley, 2013) and subsequently trimmed with trimAl v1.4.1 (Capella-Gutiérrez et al., 636 2009) using the setting -gt 0.1. The python script gen e_stitcher.py 637 (https://github.com/ballesterus/Utensils/blob/master/geneStitcher.py, last visited 12 March 2026 ) 638 was used to concatenate the trimmed alignments. The concatenated alignment was used to infer a 639 maximum likelihood phylogeny in IQTree2 v2.2.0 (Minh et al., 2020) with one partition per gene 640 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 25 and the parameter settings -m MFP -b 100 -T AUTO. Scripts can be found in 641 https://github.com/mmontonerin/Dpaulistorum_Mega-NUMT. 642 Analyzing insertion sequences in α-NUMTs 643 Finally, we used a snakemake pipeline for repeat discovery and annotation in the genome assembly, 644 which includes prediction of repeat content and transposable elements with RepeatModeler v1.0.8 645 (https://github.com/Dfam-consortium/RepeatModeler, last visited 12 March 2026). The pipeline 646 also creates a database based on UniProt/Swissprot (The UniProt Consortium, 2025) protein dataset 647 without transposable elements that initial predictions can be blasted against and removed if 648 matching. By this, protein-coding genes in multiple copies are avoided in our final repeat library. 649 The final repeat library was then used to mask the genome assembly with RepeatMasker v4.0.7 650 (https://github.com/Dfam-consortium/RepeatModeler, last visited 12 March 2026). Scripts can be 651 found in https://github.com/mmontonerin/Dpaulistorum_Mega-NUMT. 652 Collection of staged fly embryos for DNA extraction and DNA FISH 653 D. paulistorum embryos at early (0 -30 min AEL), mid (3 -4 hrs. AEL) and late (18 -20 hrs. AEL) 654 stages of development were collected on homemade medium (agar, molasses, cornmeal and yeast 655 medium) in plastic Petri dishes (3 x 3 cm) using collection chambers at 25°C. Prior t o collecting 656 embryos, 1 –2-week-old flies were kept in these chambers for 2 -3 days to adapt to the new 657 environment. Embryos were then collected by gently washing them off the plate onto a mesh with 658 water. A subset of embryos was further processed for DNA FI SH (see below). The rest was 659 manually picked with a dissection needle and transferred immediately to an ice -cold lysis buffer 660 (Qiagen, Germany) and processed for DNA extraction or stored at -20°C for no more than one 661 week. In total, we collected 20 embryos for each sample with at least five biological replicates. 662 Fluorescent in situ hybridization (FISH) for Mega-NUMT visualization 663 We performed DNA FISH on various fly tissues using both satellite and Mega-NUMT probes. The 664 Mega-NUMT probe was generated by direct labeling via PCR with modified dNTPs (Alexa Fluor 665 488 dUTP; Jena Bioscience, Germany). We used a mixture of five mitochondrial loci (CO1, CO2, 666 ND1, ND2 and ND4) of D. paulistorum, cloned beforehand in pTZ57R/5 vector (Thermo 667 Scientific, USA). For targeting the satellite sequence adjacent to the Mega-NUMT, we used a 668 primary oligo probe coupled with a sec6 adaptor ( satO11- 669 CACACGCTCTCCGTCTTGGCCGTGGTCGATCAttttttttttTTTATAAAAATTAATACTGAG670 GACAAACTGAGGACT) and the secondary probe Sec6 coupled with Cy3 (Sec6 - 671 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 26 /5Cy3/aTGATCGACCACGGCCAAGACGGAGAGCGTGTGaa). The probe for the nuclear 672 single copy gene FOXO gene was generated in the same way as the Mega -NUMT probe, but by 673 only using three probes c overing the D. paulistorum FOXO gene and labelled with Alexa Fluor 674 594 dUTP (Jena Bioscience, Germany). 675 To visualize Mega -NUMT in mitotically active neuroblasts of third instar larvae, we 676 dissected their brains in PBS and incubated for 15 min in 0.5% sodiu m citrate. Then we fixed the 677 brains for 15 min in 4% formaldehyde and 45% acetic acid before squashing them between the 678 slide and coverslip. Squashed samples were immediately immersed in liquid nitrogen and 679 transferred afterwards in 100% ethanol for 5 min. Then we air dried slides for at least 1 hour before 680 proceeding to the hybridization. For the hybridization, we used 20 pmol of satellite probes coupled 681 with 80 pmol of the sec6 probe and approximately 150 ng of Mega -NUMT probe in 50 μL of 682 hybridization buffer (50% formamide, 10% dextran sulfate, 2xSSC). We heated slides for 5 min at 683 95°C to denature and incubated them overnight at 37°C in a humid chamber. We then washed the 684 slides 3 times for 5 min with 4xSSCT and 3 times for 5 min with 0.1xSSC before moun ting in 685 SlowFade with DAPI and sealing with nail polish. 686 To visualize Mega-NUMT in mitotically active early embryos (0-3 h), staged embryos were 687 collected as stated above, dechorionated in 50% bleach and fixed for 20 min in 3.7% 688 paraformaldehyde in PBX (PBS and 0.15% Triton X-100) mixed with heptane (1:1 v/v). Then the 689 embryos were postfixed in ice-cold methanol and further processed for FISH following a protocol 690 by Gemkow et al., 1996 with slight modifications. Additionally, salivary glands and polytene 691 chromosomes were used for Mega -NUMT visualization. The glands were dissected in PBS and 692 fixed for 20 min in 3.7% paraformaldehyde in PBX (PBS and 0.15% Triton X-100). After washing 693 three times in PBX, the tissues were kept in 70% ethanol until being used for DNA FISH. Polytene 694 chromosomes were prepared according to (Pimpinelli et al., 2010) and processed for DNA FISH 695 as described above. 696 Images were taken with an Olympus FW3000 laser scanning confocal microscope 697 (Olympus Corporation, Japan) using a 60x objective and further processed with Photoshop CS6 698 (Adobe CS, USA), adjusting levels for each channel. 699 700

701 Avise, J. C., Shapira, J. F., Daniel, S. W., Aquadro, C. F., & Lansman, R. A. (1983). 702 Mitochondrial DNA differentiation during the speciation process in Peromyscus. 703 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 27 Molecular Biology and Evolution, 1(1), 38–56. 704 https://doi.org/10.1093/oxfordjournals.molbev.a040301 705 Bai, R., Cui, H., Devaney, J. M., Allis, K. M., Balog, A. M., Liu, X., Schnur, R. E., 706 Shapiro, F. L., Brautbar, A., Estrada-Veras, J. I., Hochstetler, L., McConkie-Rosell, A., 707 McDonald, M. T., Solomon, B. D., Hofherr, S., Richard, G., & Suchy, S. F. (2021). 708 Interference of nuclear mitochondrial DNA segments in mitochondrial DNA testing 709 resembles biparental transmission of mitochondrial DNA in humans. Genetics in 710 Medicine, 23(8), 1514–1521. https://doi.org/10.1038/s41436-021-01166-1 711 Baião, G. C., Schneider, D. I., Miller, W. J., & Klasson, L. (2023). Multiple 712 introgressions shape mitochondrial evolutionary history in Drosophila paulistorum and 713 the Drosophila willistoni group. Molecular Phylogenetics and Evolution, 180, 107683. 714 https://doi.org/10.1016/j.ympev.2022.107683 715 Balciuniene, J., & Balciunas, D. (2019). A Nuclear mtDNA Concatemer (Mega-NUMT) 716 Could Mimic Paternal Inheritance of Mitochondrial Genome. Frontiers in Genetics, 10. 717 https://www.frontiersin.org/articles/10.3389/fgene.2019.00518 718 Barro-Trastoy, D., & Köhler, C. (2024). Helitrons: Genomic parasites that generate 719 developmental novelties. Trends in Genetics, 40(5), 437–448. 720 https://doi.org/10.1016/j.tig.2024.02.002 721 Behura, S. K. (2007). Analysis of Nuclear Copies of Mitochondrial Sequences in 722 Honeybee (Apis mellifera) Genome. Molecular Biology and Evolution, 24(7), 1492–723 1505. https://doi.org/10.1093/molbev/msm068 724 Bensasson, D., Zhang, D.-X., Hartl, D. L., & Hewitt, G. M. (2001). Mitochondrial 725 pseudogenes: Evolution’s misplaced witnesses. Trends in Ecology & Evolution, 16(6), 726 314–321. https://doi.org/10.1016/S0169-5347(01)02151-6 727 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 28 Biró, B., Gál, Z., Fekete, Z., Klecska, E., & Hoffmann, O. I. (2024). Mitochondrial 728 genome plasticity of mammalian species. BMC Genomics, 25(1), 278. 729 https://doi.org/10.1186/s12864-024-10201-9 730 Blanchard, J. L., & Schmidt, G. W. (1995). Pervasive migration of organellar DNA to the 731 nucleus in plants. Journal of Molecular Evolution, 41(4), 397–406. 732 https://doi.org/10.1007/BF00160310 733 Blanchard, J. L., & Schmidt, G. W. (1996). Mitochondrial DNA migration events in yeast 734 and humans: Integration by a common end-joining mechanism and alternative 735 perspectives on nucleotide substitution patterns. Molecular Biology and Evolution, 13(3), 736 537–548. https://doi.org/10.1093/oxfordjournals.molbev.a025614 737 Brennecke, J., Malone, C. D., Aravin, A. A., Sachidanandam, R., Stark, A., & Hannon, G. 738 J. (2008). An Epigenetic Role for Maternally Inherited piRNAs in Transposon Silencing. 739 Science, 322(5906), 1387–1392. https://doi.org/10.1126/science.1165171 740 Breton, S., Beaupré, H. D., Stewart, D. T., Hoeh, W. R., & Blier, P. U. (2007). The 741 unusual system of doubly uniparental inheritance of mtDNA: Isn’t one enough? Trends in 742 Genetics, 23(9), 465–474. https://doi.org/10.1016/j.tig.2007.05.011 743 Brown, W. M., George, M., & Wilson, A. C. (1979). Rapid evolution of animal 744 mitochondrial DNA. Proceedings of the National Academy of Sciences, 76(4), 1967–745 1971. https://doi.org/10.1073/pnas.76.4.1967 746 Burla, H., da Cunha, A. B., Cordeiro, A. R., Dobzhansky, Th., Malogolowkin, C., & 747 Pavan, C. (1949). The Willistoni Group of Sibling Species of Drosophila. Evolution, 3(4), 748 300–314. https://doi.org/10.2307/2405716 749 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 29 Calvignac, S., Konecny, L., Malard, F., & Douady, C. J. (2011). Preventing the pollution 750 of mitochondrial datasets with nuclear mitochondrial paralogs (numts). Mitochondrion, 751 11(2), 246–254. https://doi.org/10.1016/j.mito.2010.10.004 752 Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer, K., & 753 Madden, T. L. (2009). BLAST+: Architecture and applications. BMC Bioinformatics, 754 10(1), 421. https://doi.org/10.1186/1471-2105-10-421 755 Capella-Gutiérrez, S., Silla-Martínez, J. M., & Gabaldón, T. (2009). trimAl: A tool for 756 automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics, 757 25(15), 1972–1973. https://doi.org/10.1093/bioinformatics/btp348 758 Cohen, S., Agmon, N., Yacobi, K., Mislovati, M., & Segal, D. (2005). Evidence for 759 rolling circle replication of tandem genes in Drosophila. Nucleic Acids Research, 33(14), 760 4519–4526. https://doi.org/10.1093/nar/gki764 761 Cohen, S., & Segal, D. (2009). Extrachromosomal Circular DNA in Eukaryotes: Possible 762 Involvement in the Plasticity of Tandem Repeats. Cytogenetic and Genome Research, 763 124(3–4), 327–338. https://doi.org/10.1159/000218136 764 Danecek, P., Bonfield, J. K., Liddle, J., Marshall, J., Ohan, V., Pollard, M. O., 765 Whitwham, A., Keane, T., McCarthy, S. A., Davies, R. M., & Li, H. (2021). Twelve years 766 of SAMtools and BCFtools. GigaScience, 10(2), giab008. 767 https://doi.org/10.1093/gigascience/giab008 768 Dayama, G., Emery, S. B., Kidd, J. M., & Mills, R. E. (2014). The genomic landscape of 769 polymorphic human nuclear mitochondrial insertions. Nucleic Acids Research, 42(20), 770 12640–12649. https://doi.org/10.1093/nar/gku1038 771 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 30 DeSalle, R., & Goldstein, P. (2019). Review and Interpretation of Trends in DNA 772 Barcoding. Frontiers in Ecology and Evolution, 7. 773 https://www.frontiersin.org/articles/10.3389/fevo.2019.00302 774 Dobzhansky, T., & Spassky, B. (1959). Drosophila paulistorum, a cluster of species in 775 statu nascendi*†. Proceedings of the National Academy of Sciences, 45(3), 419–428. 776 https://doi.org/10.1073/pnas.45.3.419 777 Dobzhansky, Th., Ehrman, L., Pavlovsky, O., & Spassky, B. (1964). The superspecies 778 drosophila paulistorum*. Proceedings of the National Academy of Sciences, 51(1), 3–9. 779 https://doi.org/10.1073/pnas.51.1.3 780 Elías-Gutiérrez, M., Jerónimo, F. M., Ivanova, N. V., Valdez-Moreno, M., & Hebert, P. 781 D. N. (2008). DNA barcodes for Cladocera and Copepoda from Mexico and 782 Guatemala, highlights and new discoveries. Zootaxa, 1839(1), Article 1. 783 https://doi.org/10.11646/zootaxa.1839.1.1 784 Fields, P. D., Waneka, G., Naish, M., Schatz, M. C., Henderson, I. R., & Sloan, D. B. 785 (2022). Complete Sequence of a 641-kb Insertion of Mitochondrial DNA in the 786 Arabidopsis thaliana Nuclear Genome. Genome Biology and Evolution, 14(5), evac059. 787 https://doi.org/10.1093/gbe/evac059 788 Filée, J., Forterre, P., Sen-Lin, T., & Laurent, J. (2002). Evolution of DNA Polymerase 789 Families: Evidences for Multiple Gene Exchange Between Cellular and Viral Proteins. 790 Journal of Molecular Evolution, 54(6), 763–773. https://doi.org/10.1007/s00239-001-791 0078-x 792 Gemkow, M. J., Buchenau, P., & Arndt-Jovin, D. J. (1996). FISH in whole-mount 793 Drosophila embryos. RNA: Activation of a transcriptional locus, DNA: gene architecture 794 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 31 and expression. Bioimaging, 4(2), 107–120. https://doi.org/10.1002/1361-795 6374(199606)4:2%3C107::AID-BIO8%3E3.0.CO;2-Q 796 Gleason, J. M., Griffith, E. C., & Powell, J. R. (1998). A MOLECULAR PHYLOGENY 797 OF THE DROSOPHILA WILLISTONI GROUP: CONFLICTS BETWEEN SPECIES 798 CONCEPTS? Evolution, 52(4), 1093–1103. https://doi.org/10.1111/j.1558-799 5646.1998.tb01836.x 800 Gromko, M. H., & Richmond, R. C. (1978). MODES OF SELECTION MAINTAINING 801 AN INVERSION POLYMORPHISM IN DROSOPHILA PAULISTORUM. Genetics, 802 88(2), 357–366. https://doi.org/10.1093/genetics/88.2.357 803 Harrison, R. G. (1989). Animal mitochondrial DNA as a genetic marker in population and 804 evolutionary biology. Trends in Ecology & Evolution, 4(1), 6–11. 805 https://doi.org/10.1016/0169-5347(89)90006-2 806 Hazkani-Covo, E., & Graur, D. (2007). A Comparative Analysis of numt Evolution in 807 Human and Chimpanzee. Molecular Biology and Evolution, 24(1), 13–18. 808 https://doi.org/10.1093/molbev/msl149 809 Hebert, P. D. N., Bock, D. G., & Prosser, S. W. J. (2023). Interrogating 1000 insect 810 genomes for NUMTs: A risk assessment for estimates of species richness. PLOS ONE, 811 18(6), e0286620. https://doi.org/10.1371/journal.pone.0286620 812 Hebert, P. D. N., Cywinska, A., Ball, S. L., & deWaard, J. R. (2003). Biological 813 identifications through DNA barcodes. Proceedings of the Royal Society of London. 814 Series B: Biological Sciences, 270(1512), 313–321. 815 https://doi.org/10.1098/rspb.2002.2218 816 Hirano, M., Shtilbans, A., Mayeux, R., Davidson, M. M., DiMauro, S., Knowles, J. A., & 817 Schon, E. A. (1997). Apparent mtDNA heteroplasmy in Alzheimer’s disease patients and 818 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 32 in normals due to PCR amplification of nucleus-embedded mtDNA pseudogenes. 819 Proceedings of the National Academy of Sciences, 94(26), 14894–14899. 820 https://doi.org/10.1073/pnas.94.26.14894 821 Hotopp, J. C. D., Clark, M. E., Oliveira, D. C. S. G., Foster, J. M., Fischer, P., Torres, M. 822 C. M., Giebel, J. D., Kumar, N., Ishmael, N., Wang, S., Ingram, J., Nene, R. V., Shepard, 823 J., Tomkins, J., Richards, S., Spiro, D. J., Ghedin, E., Slatko, B. E., Tettelin, H., & 824 Werren, J. H. (2007). Widespread Lateral Gene Transfer from Intracellular Bacteria to 825 Multicellular Eukaryotes. Science, 317(5845), 1753–1756. 826 https://doi.org/10.1126/science.1142490 827 Hu, J., Wang, Z., Sun, Z., Hu, B., Ayoola, A. O., Liang, F., Li, J., Sandoval, J. R., 828 Cooper, D. N., Ye, K., Ruan, J., Xiao, C.-L., Wang, D., Wu, D.-D., & Wang, S. (2024). 829 NextDenovo: An efficient error correction and accurate assembly tool for noisy long 830 reads. Genome Biology, 25(1), 107. https://doi.org/10.1186/s13059-024-03252-4 831 Junakovic, N. (2004). Southern Blot Analysis of Individual Drosophila Flies. In W. J. 832 Miller & P. Capy (Eds), Mobile Genetic Elements: Protocols and Genomic Applications 833 (pp. 41–57). Humana Press. https://doi.org/10.1385/1-59259-755-6:041 834 Jünger, M. A., Rintelen, F., Stocker, H., Wasserman, J. D., Végh, M., Radimerski, T., 835 Greenberg, M. E., & Hafen, E. (2003). The Drosophila Forkhead transcription factor 836 FOXO mediates the reduction in cell number associated with reduced insulin signaling. 837 Journal of Biology, 2(3), 20. https://doi.org/10.1186/1475-4924-2-20 838 Kastritsis, C. D. (1967). A comparative study of the chromosomal polymorphs in the 839 incipient species of the Drosophila paulistorum complex. Chromosoma, 23(2), 180–202. 840 https://doi.org/10.1007/BF00331112 841 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 33 Katoh, K., & Standley, D. M. (2013). MAFFT Multiple Sequence Alignment Software 842 Version 7: Improvements in Performance and Usability. Molecular Biology and 843 Evolution, 30(4), 772–780. https://doi.org/10.1093/molbev/mst010 844 Keeling, P. J. (2024). Horizontal gene transfer in eukaryotes: Aligning theory with data. 845 Nature Reviews Genetics, 25(6), 416–430. https://doi.org/10.1038/s41576-023-00688-5 846 Kim, B. Y., Wang, J. R., Miller, D. E., Barmina, O., Delaney, E., Thompson, A., 847 Comeault, A. A., Peede, D., D’Agostino, E. R., Pelaez, J., Aguilar, J. M., Haji, D., 848 Matsunaga, T., Armstrong, E., Zych, M., Ogawa, Y., Stamenković-Radak, M., Jelić, M., 849 Veselinović, M. S., … Petrov, D. A. (2022). Correction: Highly contiguous assemblies of 850 101 drosophilid genomes. eLife, 11, e78579. https://doi.org/10.7554/eLife.78579 851 Klasson, L., Kumar, N., Bromley, R., Sieber, K., Flowers, M., Ott, S. H., Tallon, L. J., 852 Andersson, S. G. E., & Dunning Hotopp, J. C. (2014). Extensive duplication of the 853 Wolbachia DNA in chromosome four of Drosophila ananassae. BMC Genomics, 15(1), 854 1097. https://doi.org/10.1186/1471-2164-15-1097 855 Kondo, N., Nikoh, N., Ijichi, N., Shimada, M., & Fukatsu, T. (2002). Genome fragment 856 of Wolbachia endosymbiont transferred to X chromosome of host insect. Proceedings of 857 the National Academy of Sciences, 99(22), 14280–14285. 858 https://doi.org/10.1073/pnas.222228199 859 Kress, W. J., García-Robledo, C., Uriarte, M., & Erickson, D. L. (2015). DNA barcodes 860 for ecology, evolution, and conservation. Trends in Ecology & Evolution, 30(1), 25–35. 861 https://doi.org/10.1016/j.tree.2014.10.008 862 Krimbas, C. B., & Powell, J. R. (1992). Drosophila Inversion Polymorphism. CRC Press. 863 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 34 Ladoukakis, E. D., & Zouros, E. (2017). Evolution and inheritance of animal 864 mitochondrial DNA: Rules and exceptions. Journal of Biological Research-Thessaloniki, 865 24(1), 2. https://doi.org/10.1186/s40709-017-0060-4 866 Li, H. (2013). Aligning sequence reads, clone sequences and assembly contigs with BWA-867 MEM (arXiv:1303.3997). arXiv. https://doi.org/10.48550/arXiv.1303.3997 868 Li, H. (2018). Minimap2: Pairwise alignment for nucleotide sequences. Bioinformatics, 869 34(18), 3094–3100. https://doi.org/10.1093/bioinformatics/bty191 870 Ling, F., Niu, R., Hatakeyama, H., Goto, Y., Shibata, T., & Yoshida, M. (2016). Reactive 871 oxygen species stimulate mitochondrial allele segregation toward homoplasmy in human 872 cells. Molecular Biology of the Cell, 27(10), 1684–1693. 873 https://doi.org/10.1091/mbc.E15-10-0690 874 Ling, F., & Yoshida, M. (2020). Rolling-Circle Replication in Mitochondrial DNA 875 Inheritance: Scientific Evidence and Significance from Yeast to Human Cells. Genes, 876 11(5). https://doi.org/10.3390/genes11050514 877 Lopez, J. V., Yuhki, N., Masuda, R., Modi, W., & O’Brien, S. J. (1994). Numt, a recent 878 transfer and tandem amplification of mitochondrial DNA to the nuclear genome of the 879 domestic cat. Journal of Molecular Evolution, 39(2), 174–190. 880 https://doi.org/10.1007/BF00163806 881 Luo, S., Valencia, C. A., Zhang, J., Lee, N.-C., Slone, J., Gui, B., Wang, X., Li, Z., Dell, 882 S., Brown, J., Chen, S. M., Chien, Y.-H., Hwu, W.-L., Fan, P.-C., Wong, L.-J., Atwal, P. 883 S., & Huang, T. (2018). Biparental Inheritance of Mitochondrial DNA in Humans. 884 Proceedings of the National Academy of Sciences, 115(51), 13039–13044. 885 https://doi.org/10.1073/pnas.1810946115 886 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 35 Lutz-Bonengel, S., Niederstätter, H., Naue, J., Koziel, R., Yang, F., Sänger, T., Huber, G., 887 Berger, C., Pflugradt, R., Strobl, C., Xavier, C., Volleth, M., Weiß, S. C., Irwin, J. A., 888 Romsos, E. L., Vallone, P. M., Ratzinger, G., Schmuth, M., Jansen-Dürr, P., … Parson, 889 W. (2021). Evidence for multi-copy Mega-NUMTs in the human genome. Nucleic Acids 890 Research, 49(3), 1517–1531. https://doi.org/10.1093/nar/gkaa1271 891 McDonald, J. F., Matyunina, L. V., Wilson, S., King Jordan, I., Bowen, N. J., & Miller, 892 W. J. (1997). Ltr retrotransposons and the evolution of eukaryotic enhancers. Genetica, 893 100(1), 3–13. https://doi.org/10.1023/A:1018392117410 894 McWilliams, T. G., & Suomalainen, A. (2019). Mitochondrial DNA can be inherited 895 from fathers, not just mothers. Nature, 565(7739), 296–297. 896 https://doi.org/10.1038/d41586-019-00093-1 897 Mikheenko, A., Prjibelski, A., Saveliev, V., Antipov, D., & Gurevich, A. (2018). 898 Versatile genome assembly evaluation with QUAST-LG. Bioinformatics, 34(13), i142–899 i150. https://doi.org/10.1093/bioinformatics/bty266 900 Minh, B. Q., Schmidt, H. A., Chernomor, O., Schrempf, D., Woodhams, M. D., von 901 Haeseler, A., & Lanfear, R. (2020). IQ-TREE 2: New Models and Efficient Methods for 902 Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution, 37(5), 903 1530–1534. https://doi.org/10.1093/molbev/msaa015 904 Muller, H. J. (1940). An analysis of the process of structural change in chromosomes 905 ofDrosophila. Journal of Genetics, 40(1), 1–66. https://doi.org/10.1007/BF02982481 906 Nikoh, N., Tanaka, K., Shibata, F., Kondo, N., Hizume, M., Shimada, M., & Fukatsu, T. 907 (2008). Wolbachia genome integrated in an insect chromosome: Evolution and fate of 908 laterally transferred endosymbiont genes. Genome Research, 18(2), 272–280. 909 https://doi.org/10.1101/gr.7144908 910 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 36 Noutsos, C., Kleine, T., Armbruster, U., DalCorso, G., & Leister, D. (2007). Nuclear 911 insertions of organellar DNA can create novel patches of functional exon sequences. 912 Trends in Genetics, 23(12), 597–601. https://doi.org/10.1016/j.tig.2007.08.016 913 Pagnamenta, A. T., Wei, W., Rahman, S., & Chinnery, P. F. (2021). Biparental 914 inheritance of mitochondrial DNA revisited. Nature Reviews Genetics, 22(8), 477–478. 915 https://doi.org/10.1038/s41576-021-00380-6 916 Papachristos, K., Miller, W. J., & Klasson, L. (2025). A co-speciation dilemma and a 917 lifestyle transition with genomic consequences in Wolbachia of Neotropical Drosophila. 918 BMC Genomics, 27(1), 41. https://doi.org/10.1186/s12864-025-12340-z 919 Parfait, B., Rustin, P., Munnich, A., & Rötig, A. (1998). Coamplification of Nuclear 920 Pseudogenes and Assessment of Heteroplasmy of Mitochondrial DNA Mutations. 921 Biochemical and Biophysical Research Communications, 247(1), 57–59. 922 https://doi.org/10.1006/bbrc.1998.8666 923 Parr, R. L., Maki, J., Reguly, B., Dakubo, G. D., Aguirre, A., Wittock, R., Robinson, K., 924 Jakupciak, J. P., & Thayer, R. E. (2006). The pseudo-mitochondrial genome influences 925 mistakes in heteroplasmy interpretation. BMC Genomics, 7(1), 185. 926 https://doi.org/10.1186/1471-2164-7-185 927 Patterson, E. C., Lall, G. M., Neumann, R., Ottolini, B., Sacchini, F., Foster, A. P., 928 Jobling, M. A., & Wetton, J. H. (2023). Defining cat mitogenome variation and 929 accounting for numts via multiplex amplification and Nanopore sequencing. Forensic 930 Science International: Genetics, 67, 102944. https://doi.org/10.1016/j.fsigen.2023.102944 931 Pimpinelli, S., Bonaccorsi, S., Fanti, L., & Gatti, M. (2010). Fluorescent In Situ 932 Hybridization (FISH) of Mitotic Chromosomes from Drosophila Larval Brain. Cold 933 Spring Harbor Protocols, 2010(3), pdb.prot5391. https://doi.org/10.1101/pdb.prot5391 934 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 37 Pohjoismäki, J. L. O., Goffart, S., Tyynismaa, H., Willcox, S., Ide, T., Kang, D., 935 Suomalainen, A., Karhunen, P. J., Griffith, J. D., Holt, I. J., & Jacobs, H. T. (2009). 936 Human Heart Mitochondrial DNA Is Organized in Complex Catenated Networks 937 Containing Abundant Four-way Junctions and Replication Forks*. Journal of Biological 938 Chemistry, 284(32), 21446–21457. https://doi.org/10.1074/jbc.M109.016600 939 Pritam, S., & Signor, S. (2025). Evolution of piRNA-guided defense against transposable 940 elements. Trends in Genetics, 41(5), 390–401. https://doi.org/10.1016/j.tig.2024.11.011 941 Puertas, M. J., & González-Sánchez, M. (2020). Insertions of mitochondrial DNA into the 942 nucleus—Effects and role in cell evolution. Genome, 63(8), 365–374. 943 https://doi.org/10.1139/gen-2019-0151 944 R Core Team. (2016). R: A language and environment for statistical computing 945 [Computer software]. R Foundation for Statistical Computing, Vienna, Austria. 946 https://www.R-project.org/ 947 Rafanel, B., Protsenko, L., Handler, D., Brennecke, J., & Senti, K.-A. (2025). Antisense 948 transposon insertions into host genes trigger piRNA mediated immunity (p. 949 2025.07.28.667215). bioRxiv. https://doi.org/10.1101/2025.07.28.667215 950 Robe, L. J., Cordeiro, J., Loreto, E. L. S., & Valente, V. L. S. (2010). Taxonomic 951 boundaries, phylogenetic relationships and biogeography of the Drosophila willistoni 952 subgroup (Diptera: Drosophilidae). Genetica, 138(6), 601–617. 953 https://doi.org/10.1007/s10709-009-9432-5 954 Rogers, H. H., & Griffiths-Jones, S. (2012). Mitochondrial Pseudogenes in the Nuclear 955 Genomes of Drosophila. PLOS ONE, 7(3), e32593. 956 https://doi.org/10.1371/journal.pone.0032593 957 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 38 Rutherford, K., Parkhill, J., Crook, J., Horsnell, T., Rice, P., Rajandream, M.-A., & 958 Barrell, B. (2000). Artemis: Sequence visualization and annotation. Bioinformatics, 959 16(10), 944–945. https://doi.org/10.1093/bioinformatics/16.10.944 960 Salas, A., Schönherr, S., Bandelt, H.-J., Gómez-Carballa, A., & Weissensteiner, H. 961 (2020). Extraordinary claims require extraordinary evidence in asserted mtDNA 962 biparental inheritance. Forensic Science International: Genetics, 47, 102274. 963 https://doi.org/10.1016/j.fsigen.2020.102274 964 Salas, A., Yao, Y.-G., Macaulay, V., Vega, A., Carracedo, Á., & Bandelt, H.-J. (2005). A 965 Critical Reassessment of the Role of Mitochondria in Tumorigenesis. PLOS Medicine, 966 2(11), e296. https://doi.org/10.1371/journal.pmed.0020296 967 Schiavo, G., Hoffmann, O. I., Ribani, A., Utzeri, V. J., Ghionda, M. C., Bertolini, F., 968 Geraci, C., Bovo, S., & Fontanesi, L. (2017). A genomic landscape of mitochondrial 969 DNA insertions in the pig nuclear genome provides evolutionary signatures of 970 interspecies admixture. DNA Research, 24(5), 487–498. 971 https://doi.org/10.1093/dnares/dsx019 972 Schwartz, M., & Vissing, J. (2002). Paternal Inheritance of Mitochondrial DNA. New 973 England Journal of Medicine, 347(8), 576–580. https://doi.org/10.1056/NEJMoa020350 974 Shafin, K., Pesout, T., Chang, P.-C., Nattestad, M., Kolesnikov, A., Goel, S., Baid, G., 975 Kolmogorov, M., Eizenga, J. M., Miga, K. H., Carnevali, P., Jain, M., Carroll, A., & 976 Paten, B. (2021). Haplotype-aware variant calling with PEPPER-Margin-DeepVariant 977 enables high accuracy in nanopore long-reads. Nature Methods, 18(11), 1322–1332. 978 https://doi.org/10.1038/s41592-021-01299-w 979 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 39 Shutt, T. E., & Gray, M. W. (2006). Bacteriophage origins of mitochondrial replication 980 and transcription proteins. Trends in Genetics, 22(2), 90–95. 981 https://doi.org/10.1016/j.tig.2005.11.007 982 Spassky, B., & Dobzhansky, T. (1950). Comparative genetics of Drosophila willistoni. 983 Heredity, 4(2), 201–215. https://doi.org/10.1038/hdy.1950.15 984 The UniProt Consortium. (2025). UniProt: The Universal Protein Knowledgebase in 985 2025. Nucleic Acids Research, 53(D1), D609–D617. 986 https://doi.org/10.1093/nar/gkae1010 987 Thomas, J., Vadnagara, K., & Pritham, E. J. (2014). DINE-1, the highest copy number 988 repeats in Drosophila melanogaster are non-autonomous endonuclease-encoding rolling-989 circle transposable elements (Helentrons). Mobile DNA, 5(1), 18. 990 https://doi.org/10.1186/1759-8753-5-18 991 Tsuji, J., Frith, M. C., Tomii, K., & Horton, P. (2012). Mammalian NUMT insertion is 992 non-random. Nucleic Acids Research, 40(18), 9073–9088. 993 https://doi.org/10.1093/nar/gks424 994 Valiati, V. H., & Valente, V. L. S. (1997). Chromosomal polymorphism in urban 995 populations of Drosophila paulistorum. Brazilian Journal of Genetics, 20, 567–581. 996 https://doi.org/https://doi.org/10.1590/S0100-84551997000400004 997 Vendrami, D. L. J., Gossmann, T. I., Chakarov, N., Paijmans, A. J., Litzke, V., Eyre-998 Walker, A., Forcada, J., & Hoffman, J. I. (2022). Signatures of Selection on Mitonuclear 999 Integrated Genes Uncover Hidden Mitogenomic Variation in Fur Seals. Genome Biology 1000 and Evolution, 14(7), evac104. https://doi.org/10.1093/gbe/evac104 1001 Walker, B. J., Abeel, T., Shea, T., Priest, M., Abouelliel, A., Sakthikumar, S., Cuomo, C. 1002 A., Zeng, Q., Wortman, J., Young, S. K., & Earl, A. M. (2014). Pilon: An Integrated Tool 1003 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint 40 for Comprehensive Microbial Variant Detection and Genome Assembly Improvement. 1004 PLOS ONE, 9(11), e112963. https://doi.org/10.1371/journal.pone.0112963 1005 Wang, J.-X., Liu, J., Miao, Y.-H., Huang, D.-W., & Xiao, J.-H. (2020). Tracking the 1006 Distribution and Burst of Nuclear Mitochondrial DNA Sequences (NUMTs) in Fig Wasp 1007 Genomes. Insects, 11(10), Article 10. https://doi.org/10.3390/insects11100680 1008 Wei, W., Pagnamenta, A. T., Gleadall, N., Sanchis-Juan, A., Stephens, J., Broxholme, J., 1009 Tuna, S., Odhams, C. A., Fratter, C., Turro, E., Caulfield, M. J., Taylor, J. C., Rahman, S., 1010 & Chinnery, P. F. (2020). Nuclear-mitochondrial DNA segments resemble paternally 1011 inherited mitochondrial DNA in humans. Nature Communications, 11(1), 1740. 1012 https://doi.org/10.1038/s41467-020-15336-3 1013 Wei, W., Schon, K. R., Elgar, G., Orioli, A., Tanguy, M., Giess, A., Tischkowitz, M., 1014 Caulfield, M. J., & Chinnery, P. F. (2022). Nuclear-embedded mitochondrial DNA 1015 sequences in 66,083 human genomes. Nature, 611(7934), 105–114. 1016 https://doi.org/10.1038/s41586-022-05288-7 1017 Werren, J. H. (2011). Selfish genetic elements, genetic conflict, and evolutionary 1018 innovation. Proceedings of the National Academy of Sciences, 108(supplement_2), 1019 10863–10870. https://doi.org/10.1073/pnas.1102343108 1020 Yao, Y.-G., Salas, A., Bravi, C. M., & Bandelt, H.-J. (2006). A reappraisal of complete 1021 mtDNA variation in East Asian families with hearing impairment. Human Genetics, 1022 119(5), 505–515. https://doi.org/10.1007/s00439-006-0154-9 1023 Zanini, R., Müller, M. J., Vieira, G. C., Valiati, V. H., Deprá, M., & Valente, V. L. da S. 1024 (2018). Combining morphology and molecular data to improve Drosophila paulistorum 1025 (Diptera, Drosophilidae) taxonomic status. Fly, 12(2), 81–94. 1026 https://doi.org/10.1080/19336934.2018.1429859 1027 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 1, 2026. ; https://doi.org/10.64898/2026.03.31.715258doi: bioRxiv preprint