Uncovering the architecture of selection in two Bos taurus cattle breeds

Selection alters the genome via hard sweeps, soft sweeps, and polygenic selection. However, mapping polygenic selection is difficult because it does not leave clear signatures on the genome like a selective sweep. In populations with temporally-stratified genotypes, the Generation Proxy Selection Mapping (GPSM) method identifies variants associated with generation number (or appropriate proxy) and thus variants undergoing directional allele frequency changes. Here, we use GPSM on two large datasets of beef cattle to detect associations between an animals generation and 11 million imputed SNPs. Using these datasets with high power and dense mapping resolution, GPSM detected a total of 294 unique loci actively under selection in two cattle breeds. We observed that GPSM has a high power to detect selection in the very recent past (< 10 years), even when allele frequency changes are small. Variants identified by GPSM reside in genomic regions associated with known breed characteristics, such as fertility and maternal ability in Red Angus and carcass merit and coat color in Simmental. Over 60% of the selected loci reside in or near (<50 kb) annotated genes. Additionally, 36% of selected loci overlap known epigenetic marks or putative functional genomic regions. Using RAiSD and nSL, we identify hundreds of putative selective sweeps; however, these sweeps have little overlap with polygenic selected loci. This makes GPSM a complementary approach to sweep detection methods when temporal genotype data are available. The selected loci that we identify across methods demonstrate the complex architecture of selection in domesticated cattle.