Association Analyses Between Missense Variation at 70 Candidate Genes and Porcine Lipid Metabolism Traits

Non-synonymous polymorphisms can have important effects on protein structure and activity. In the current work, we wanted to find out if missense variation of porcine lipid metabolism genes is associated with the variation of fatness traits. We have sequenced the whole genomes of five founder Duroc boars that have sired 350 F₁ offspring with records for fatness and meat quality traits. Non-synonymous variation located at 760 lipid-related genes has been retrieved from these five genomic sequences and a total of 328 missense polymorphisms have been identified. In silico prediction tools (SIFT, cSNP Panther, MutPred etc.) have been used to select a panel of 108 SNPs (70 genes). Amongst these, mutations at the ACSM3, PCDH15, EHHADH, IRS1, DHCR24, DHCR7, HAO1 and CAV2 were identified by at least three softwares as highly deleterious. The level of prediction consistency for the remainings SNPs was weaker. All these polymorphisms were simultaneously genotyped in the 350 F₁Duroc pigs by real-time PCR in an Open Array platform. Analysis of genetic and phenotypic data with GEMMA, revealed highly significant associations between PARK2 (G80D) genotype and gluteus medius (GM) C18:0 (q-value = 0.019) and C18:3 (q-value =  0.038) content, FABP3 (G25S) and GM C17:0 (q-value = 0.029), LPIN3 (V580I) and GM fat content (q-value = 0.013), and DHCR7 (L142P and L149P) and C17:1 at the longissimus dorsi (q-value < 0.05). In general, the level of association between SNPs predicted to be functionally relevant and those displaying significant associations with lipid traits was quite modest.