Integrative analysis of metabolomic, proteomic and genomic data to reveal functional pathways and candidate genes for drip loss in pigs

In genetic analyses new types of functional traits can be used to increase the information density between genes and phenotypes. The consideration of different omics levels, like protein and metabolite abundance allows us to elucidate the black box of phenotype expression and to identify potential candidate genes. Aim of this study was to integrate the data of the genome, proteome and metabolome to characterize underlying functional pathways, biochemical processes and corresponding candidate genes of the meat quality parameter drip loss in pigs. Generally, this trait is strongly influenced by environmental effects and therefore our hypothesis was that metabolites and proteins, significantly associated with drip loss, are more accurate and reliable phenotypes than drip loss itself. We applied an untargeted (holistic) metabolomics approach and a targeted protein profiling to determine and quantify the metabolite and protein profiles in Musculus longissimus dorsi (LD). Therefore, LD samples of 97 Duroc × Pietrain pigs were collected and phenotypes as well as genotypes were recorded. In total, 126 and 40 KEGG annotated metabolites and proteins were identified in the tissue samples. An enrichment analysis resulted in 206 detected pathways and 10 pathways showed a significant meaning for drip loss. Besides drip loss itself, 18 metabolites and four proteins that belong to the significant metabolic pathways, namely sphingolipid metabolism, pyruvate metabolism, glycolysis/ gluconeogenesis and methane metabolism were analyzed as phenotypes within a genome-wide association study (GWAS). For drip loss, two proteins and 11 metabolites we detected in total 430 significantly associated genetic markers (SNPs) located on Sus scrofa chromosomes (SSC) 1, 2, 3, 4, 6, 7, 8, 10, 13, 14, 16, 17, and 18. The functional annotation of these SNPs allowed discovering promising genomic regions as well as potential candidate genes. As one interesting example, on SSC 18 the genes PTN, CREB3L2 and LRGUK were identified based on significant SNPs for muscle physiology associated with drip loss, protein phosphoglycerate mutase 2 (PGAM2) and metabolite glycine. These genes are involved in the homoeostasis of muscle energy metabolism and metabolic disorders. Because of the increased information density due to the consideration of proteome and metabolome, our results lead to the conclusion that a combined omics analysis has an advantage compared to 'classical' GWAS. Therefore, we expect that G