A Cattle Pangenome Approach Reveals Novel Non-Reference Unique Insertions and Their Impact on Economic Traits

The current cattle reference genome assembly, which is derived from a singleHereford cow, does not encompass the entire range of genetic variationspresent within the species. Several megabases of sequences are still missingand these can contain genomic regions potentially associated with keyphenotypes.Structural variations (SVs ≥ 50 nucleotides) are difficult to detect using onlystandard approaches of either short or long-read sequence mapping to thecurrent bovine genome assembly. To overcome this, recent advances insequencing technologies to produce high-quality long-read sequencescombined with the development of pipelines and tools for sequence analysispave the way for the use of pangenomics in cattle. Those approaches providethorough comprehension of the extent of genetic diversity among cattlebreeds.In this study, we constructed a cattle pangenome graph using 16 high-qualityhaplotype-resolved genome assemblies originating from 9 French cattlebreeds, encompassing both major dairy and beef populations as well asregional breeds, and including the Yak (Bos grunniens) as a close species. Acombination of complementary methods led to the detection andcharacterization of 101,219 structural variations. Of these, 33,634 structuralvariants were classified as non-reference unique insertions (NRUIs), addingseveral megabases of new genomic sequences that are not present in thecurrent Hereford reference genome.These NRUIs were subsequently genotyped (presence/absence) in a largemulti-breed genotyped cattle population. With these NRUI data, all sampleswere clustered in distinct groups corresponding to the breed of origin. Thisresult is of particular interest as the NRUIs were able to provide a cleardescription of the breed structure.Our findings underscore the importance of pangenome-based approaches touncover the full spectrum of genomic variations and establish a potential linkwith cattle key phenotypes.