An integrative GWAS and RNA-seq study to identify SNPs and transcripts related to sperm quality traits in pigs

Resumen del póster presentado a la 37th International Conference on Animal Genetics (ISAG), celebrada en Lleida (España) del 7 al 12 de julio de 2019.
For the last decades, boars have been selected for their genetic merit on carcass and meat quality traits. However, breeders and researchers are now paying attention to additional phenotypes including sperm quality. We carried a GWAS and a semen RNA-seq experiment in pigs with the aim to identify SNP markers for 25 sperm quality traits. The GWAS included 288 boars and genotypes from the Affymetrix Axiom Porcine 660K Genotyping Array. Sperm from 40 of these pigs were subjected to total and to small RNA-seq. The GWAS resulted in 345 SNPs significantly associated mostly to sperm head abnormalities and motility. The RNA-seq evidenced 4,120 protein coding genes and 95 miRNAs. 3,053 genes showed significant correlations with at least one semen quality trait. Independently for each phenotype, we then searched for eQTLs using only the GWAS SNP hits and the genes that correlated with the given semen trait. This yielded 119 eQTLs. Several of these hits involve genes with known function related to semen quality. For example, we found a hit on ACTR2 and the percentage of head abnormalities. ACTR2 has been involved in the morphogenic modeling of the sperm head. We also detected a hit between NDUFS8 and sperm motility. NDUFS8 is a known mitochondrial NADH subunit involved in the respiratory chain which is important for sperm capacitation and motility. Using the RNA-seq data, we also searched for the subset of 10 RNAs that explained the largest percentage of the variability for the semen quality phenotypes. To achieve this objective, we retrieved the genes that (i) were co-associated with traits at the SNP level by building an Associated Weight Matrix and finding correlations with PCIT, and (ii) also displayed co-abundant RNA levels (by PCIT). The resulting list of genes, together with the information of their correlated phenotypes, was used for network analysis. We identified a subset of 10 genes, which included EFHC1, ATP9A and THADA, that explained between 16 to 67% of the different sperm quality parameters.