Design and Application of the Cattle GGP-F250 Array

As many as 3,000 bovine genomes have now been sequenced worldwide and much of the variation within the cattle genomes has been discovered. Many variants likely to have functional effects due to their locations within annotated coding regions have been identified. However, the lack of complete annotation leads to our inability to identify variants that lie within regulatory regions. Current livestock genotyping assays are based on evenly spaced, common variants randomly sampled from the genome without respect to their functionality.  These may be imputed to whole genome sequence (WGS) with accuracies ~90% and used in GWAS to fine-map QTL, however, imputation accuracy is probably limited by the allele frequency spectrum disparity between WGS and chip-based variants. To address this, and other opportunities, we have designed the first generation of a bovine functional assay, the GGP-F250, to contain 34K common variants present on many of the genotyping assays currently used by the cattle industry and 199K predicted genic functional variants. These variants were discovered by analyzing WGS data for 297 cattle from 17 breeds and RNA-seq data for 159 animals and were confirmed using data from the 1000 Bull Genomes Project and dbSNP. Here we describe the design of the assay and provide a characterization of its performance in 4,000 genotyped individuals from 3 breeds. We also discuss its applications in WGS imputation, QTL discovery and identification of loss-of function variants responsible for embryonic mortality or heterosis and inbreeding depression. The assay is publicly available from GeneSeek.