Integration of Collaborative Cross and Hybrid Mouse Diversity Panel studies does not significantly improve genome wide association study results

Genome-wide association studies rely on maximizing cohort size and minimizing environmental noise to increase their efficacy. Model organism Genetic Reference Populations offer significantly reduced environmental noise and deep phenotyping, yet rarely are multiple populations merged together. We evaluated whether combining the Collaborative Cross (CC) and Hybrid Mouse Diversity Panel (HMDP), two well-studied murine Genetic Reference Populations would improve GWAS outcomes. Utilizing two prior studies by our group that had performed the same study for heart failure in each cohort, we examined each cohort individually and then together using the pyLMM GWAS algorithm. Surprisingly, the combined cohort of 175 strains produced only 123 loci, a modest gain over the 104-strain HMDP (114), and fewer than in the 71-strain CC (149) and did not systematically improve resolution. Locus overlap between the CC and HMDP cohorts was very sparse, with only a single locus (near Eprs1) found to be significant in each cohort, although nearly half of all combined loci were also found in the CC cohort. These results caution that merging divergent GRPs may act to dilute signal and complicate interpretation, despite larger sample sizes.