A Genome-Wide Association Study of Fetal Response to Type 2 PRRSV Challenge

It is reported that permutation technique provides an empirical test to control the overestimation behaviour of the least absolute shrinkage and selection operator (LASSO) in high-dimension regression, while keeping a relatively high power. We applied this method in a genome-wide association study (GWAS) of fetal response to type 2 porcine reproductive and respiratory syndrome virus (PRRSV) challenge. The analysis was conducted using data from 928 fetuses from pregnant gilts that had been experimentally challenged with type 2 PRRSV at gestation day 85 ±1. A generalized linear model was used to model environmental effects and estimate SNP effects. The LASSO was applied to fit the model, and the empirical significance criteria of association were determined using 1000 permutations. Two SNPs were significantly associated with viral load in fetal thymus (P < 0.05) and another showed suggestive association (P < 0.1). Sixteen SNPs were significantly associated with fetal death (P < 0.05) and four other SNPs showed suggestive association (P < 0.1). A single SNP was significantly associated with fetal viability (P < 0.05). Functional analyses showed that the detected candidate regions overlap with previously reported QTLs for pig health and reproduction, and the results were also supported by a related transcriptomic analyses of fetal response to PRRSV infection.