Identification of Causal Variants Using Next Generation Sequencing Data

MBVs developed for one breed are not useful in others and genomic selection is restricted to numerically large breeds and the phenotypes recorded in these breeds. This has motivated a resurgence of interest in QTL hunting since, in the absence of epistasis, causal variants are expected to create V_A in all populations in which they segregate. The majority of quantitative trait variation is caused by small-effect variants but a few variants explain >1% of V_A in each trait. If the largest-effect variants explaining ~10% of V_A in each trait could be identified, low-cost genotyping assays could be deployed, regardless of breed, to predict MBVs with accuracies ~30%. About 3,000 cattle have now been sequenced worldwide and whole genome variant calls are publicly available for >1,000. These sequences can be used to impute 50K or 770K SNP genotypes to whole genome sequence (bialleleic loci with MAF > 0.05) with remarkable accuracies (>95%). Since North American beef breed associations now have large numbers of registered animals with 50K genotypes (60,000 Angus; 24,000 Simmental; 6,000 Hereford; 3,800 Red Angus; 1,100 Gelbvieh) an opportunity exists to impute these genotypes to full sequence and identify large-effect QTL. Imputed data can also be used to identify loci with moderate allele frequencies that differ from HWE within breeds because they are lethal with complete or incomplete penetrance. The approach is limited by quality of the reference assembly, analysis pipelines which may not accurately identify large indels or CNVs, and analysis software that ignores multi-allelic loci (~6.7%).