A comparative genome landscape of mitochondrial DNA insertions into two cattle nuclear genome versions

mtDNA insertions have been detected in the nuclear genome of many eukaryotes, including mammals.1 These mtDNA-related regions (called nuclear DNA sequences of mitochondrial origin or NUMTs) are pseudogenes derived from mtDNA fragments that have been integrated into the nuclear genome through horizontal transfer mechanisms.2 NUMTs are considered sequence fossils that have contributed to shaping of the genome architecture and evolution of the nuclear genomes. Some of these regions have high homology with the original mtDNA genome as they derive from recent insertion events in terms of evolutionary time. Unintentionally amplified NUMTs can create biases in detecting true heteroplasmy and in phylogenetic analyses based on mtDNA, as also demonstrated by the large fraction of mtDNA sequences available in GenBank/EMBL databases that are estimated to contain errors.3, 4 A few studies have reported information on NUMTs in the genome of several livestock species,3-6 including cattle.7-9. In this study, we obtained a detailed genome map of NUMTs in the Bos taurus genome and compared their distribution between the latest assembled versions, UMD3.1 and ARS-UCD1.2. These two genome versions were aligned with the reference linearized B. taurus mtDNA sequence using LAST software,10 following the procedure already described in Schiavo et al.3 (Appendix S1). Fig. 1 reports the distribution of the identified NUMTs in the two genome versions over the modern cattle mtDNA genome and the distribution of the level of identity between NUMTs and the corresponding mitochondrial regions. The highest proportion of NUMTs had an identity of about 70–80% with the modern mtDNA sequence, suggesting that most integration events occurred about 10–30 million years before present in ancient genomes that evolved in the current B. taurus species.5, 8 It will be interesting to analyse the genomes of other ruminants to evaluate the impact of mtDNA integrations on the evolutionary processes of their lineages. Table S1 reports comparative information on the NUMTs identified in the different BTA chromosomes in the two genome versions. For most chromosomes, these two assemblies added or eliminated some NUMTs in several positions (Table S1 and Fig. S1). Data S1 reports detailed information on all identified NUMTs. A total of 422 and 441 NUMTs (for a total of 242 409 and 238 345 bp, which covered the whole mtDNA sequence, Fig. 1) were identified in the UMD3.1 and ARS UCD1.2 nuclear genome versions, respectively. A few of these NUMTs were amplified and sequenced in 50 cattle of different breeds (Appendix S1 and Table S2). Dot-plot analysis of NUMT-containing regions (Appendix S1) indicated that some of them might be derived from single insertional events that subsequently underwent additional mutations over evolutionary time that split the integrated mtDNA in separated but co-linear regions with a few exceptions that included more complex rearrangements with duplications and inversions. Some examples are reported in Fig. S2. Those regions including more than one NUMT sequence were defined as NUMT regions (Appendix S1). A total of 171 and 178 NUMT regions were identified in the UMD3.1 and ARS UCD1.2 nuclear genome versions, respectively (Data S1). Two GFF files were produced to support the annotation of the two cattle genome versions (Data S2 and S3). These data will be helpful to plan studies based on mtDNA sequences in cattle and to interpret previous mtDNA sequence results obtained in this species.