Comparative architectures of direct and social genetic effects from the genome-wide association study of 170 phenotypes in outbred laboratory mice

Social genetic effects (SGE, also called indirect genetic effects) are associations between genotypes of one individual and phenotype of another. SGE arise when two individuals interact and heritable traits of one influence the phenotype of the other. Recent studies have shown that SGE substantially contribute to phenotypic variation in humans and laboratory mice, which suggests that SGE, like direct genetic effects (DGE, effects of an individual9s genes on their own phenotype), are amenable to mapping. Using 170 phenotypes including behavioural, physiological and morphological traits measured in outbred laboratory mice, we empirically explored the potential and challenges of genome-wide association study of SGE (sgeGWAS) as a tool to discover novel mechanisms of social effects between unrelated individuals. For each phenotype we performed sgeGWAS, identifying 21 genome-wide significant SGE associations for 17 phenotypes, and dgeGWAS for comparison. Our results provide three main insights: first, SGE and DGE arise from partially different loci and/or loci with different effect sizes, which implies that the widely-studied mechanism of phenotypic 9contagion9 is not sufficient to explain all social effects. Secondly, several DGE associations but no SGE associations had large effects, suggesting sgeGWAS is unlikely to uncover 9low hanging fruits9. Finally, a similar number of variants likely contribute to SGE and DGE. The analytical framework we developed in this study and the insights we gained from our analyses will inform the design, implementation and interpretation of sgeGWAS in this and other populations and species.