124
Genomic prediction and genomic variance partitioning of daily and residual feed intake in pigs using Bayesian Power Lasso models
Improvement of feed efficiency is essential in pig breeding and selection for reduced residual feed intake (RFI) is an option. The study applied Bayesian Power LASSO (BPL) models with different power parameter to investigate genetic architecture, to predict genomic breeding values, and to partition genomic variance for RFI and daily feed intake (DFI). A total of 1272 Duroc pigs had both genotypic and phenotypic records for these traits. Significant SNPs were detected on chromosome 1 (SSC 1) and SSC 14 for RFI and on SSC 1 for DFI. BPL had similar accuracy and bias as GBLUP but power parameters had no effect on predictive ability. Genomic variance partitioning showed that SNP groups either by position (intron, exon, downstream, upstream and 5’UTR) or by function (missense and protein-altering) had similar average explained variance per SNP, except that 3’UTR had a higher value.
Keywords:
Bayesian Power Lasso
genomic prediction
pigs
