Summary statistics from large-scale genome-wide association studies (GWAS) show that MDD [11], loneliness [10], and coronary artery disease (CAD) [12], the most common type of heart disease, are genetically correlated, suggesting shared genetic risk factors (Fig. 1b). However, genetic correlations based on GWAS summary statistics can be difficult to interpret, as each study typically ascertains cases on only a single disease without adjusting for or matching on phenotypic comorbidities. While this approach is optimal in GWAS for many reasons, it introduces the possibility of “phenotypic hitchhiking” in which a comorbid trait (e.g., MDD) is unintentionally selected during the ascertainment of the index trait (e.g., CAD). Thus, two heritable phenotypes (e.g., MDD and CAD) that share common environmental risk factors (e.g., obesity) but no genetic risk factors can appear genetically correlated in an analysis based solely on GWAS summary statistics.
