Genome-wide association studies (GWAS) are currently the primary tool to identify genetic variants (GVs) underlying phenotypic variation. GWAS are generally univariate in nature, i.e., focus on a single phenotype. This means that researchers, prior to analyses, often reduce available, originally multivariate, phenotypic information (e.g., information on multiple questions from a diagnostic interview or questionnaire, or multiple items in a test) to a single phenotypic composite score, such as a continuous sum score or binary case-control status (the latter is often based on the number of endorsed symptoms, i.e., effectively a dichotomized sum score). Such univariate conceptualisations are consistent with the practical and diagnostic definitions employed in psychology and medicine of traits like depression, cognition, Type I diabetes, and asthma. However, whether they represent informative entities with respect to biological aetiology is questionable [1]. Many acknowledge the possible genetic heterogeneity of psychological and medical traits [2]–[3]. This heterogeneity implies that distinct GVs may give rise to the same univariate trait score, and that the same GV may have different behavioral manifestations, depending on genetic background and environmental exposure. It also implies
