Integrating whole-genome sequencing and metagenomics improves phenotype prediction of fat deposition traits in pigs

Background Fat deposition is a key economic trait affecting production efficiency and carcass quality in pigs and is influenced by both genetic and environmental factors. Although increasing evidence indicates that the gut microbiome contributes to host phenotypic variation, current genomic evaluation and prediction models largely rely on host genotype information and rarely account for microbiome effects. This study aimed to quantify the contribution of the gut microbiome to fat deposition traits and to evaluate whether incorporating microbiome information could improve phenotypic prediction in pigs. Results We collected phenotypic characteristics of six fat deposition traits including the shoulder backfat thickness, the 6th-7th rib backfat thickness, the last rib backfat thickness, waist backfat thickness, the 3rd-4th last rib backfat thickness, and the average backfat thickness from 478 Duroc × Landrace × Yorkshire pigs. Using whole-genome resequencing and metagenomic sequencing, we identified 21 million single nucleotide variants and quantified the abundances of 20,336 gut microbial species across 478 pigs. Distinct fecal microbial community structures were observed between pigs with high- and low-fat deposition, and 27 bacterial species were identified as potential biomarkers associated with fat deposition. Microbiability estimates for the six backfat thickness traits ranged from 0.17 to 0.44, indicating that the gut microbiome accounted for a considerable proportion of phenotypic variance beyond host genetic contributions. In contrast, the interaction effects between the host genome and gut microbiome accounted for only a negligible proportion of phenotypic variance across the six traits. Incorporating overall microbiome information into genomic prediction models resulted in mean phenotypic predictive abilities ranging from 0.082 to 0.292, representing increases in predictive ability of 0.023 to 0.084 across traits and reductions in the mean squared error of prediction of 0.001 to 0.009 compared with the additive genetic model. Conclusions Our findings demonstrate that the gut microbiome makes a substantial contribution to variation in fat deposition traits in pigs. Integrating host whole-genome resequencing and gut metagenomic information yielded modest absolute gains in phenotypic predictive ability, suggesting that the gut microbiome may provide complementary information for prediction models of complex traits in pigs.