From genomes to hologenomes: integrating host and microbiome data for complex trait prediction in livestock and aquaculture

Over recent decades, global livestock and aquaculture systems have significantly increased protein production, largely driven by advancements in nutrition, health management, and selective breeding programs. The integration of genomic data, particularly dense SNP panels, into animal breeding has revolutionized trait prediction, enabling more accurate estimation of breeding values for complex traits such as growth, carcass yield, and disease resistance in animal farming. Currently, animal production faces new challenges, including production efficiency, environmental impact, and emerging and re-emerging diseases. There is broad evidence that variation in host-associated microbiomes is associated with host phenotypic diversity, allowing to predict complex traits in livestock and aquaculture. Additionally, the integration of host genomic and microbial metagenomic data has demonstrated potential to improve prediction accuracy for complex traits, accelerating the rate of genetic gain. These findings have led to the development of new concepts, including microbiability (the proportion of phenotypic variance explained by the microbiome) and holobiability (the joint contribution of host and microbial variance). This review discusses recent advances in incorporating microbiome information as an additional variation source into genomic selection methods, with applications for complex trait prediction in livestock and aquaculture, providing upcoming challenges and opportunities. We highlight the challenges of modeling host–microbiome interactions, the potential of intermediate and functional traits, and considerations when designing holobiont-driven breeding schemes. Integrating these dimensions into breeding programs requires methodological innovations in data collection, modeling, and computation. Advances in high-throughput sequencing, artificial intelligence, and multi-omics facilitate the analysis of both genomic and metagenomic datasets, and support targeted microbiome interventions, including microbiome engineering, diet modulation via prebiotics or probiotics, and microbiome breeding to select holobionts with improved performance for complex traits. Thus, transitioning from genomes to hologenomes and incorporating microbiome data into breeding programs represents a key step toward more precise, efficient, and sustainable animal breeding.