of the current analysis as quantitative and qualitative measures were combined. In the current study, a genetic association with the depression continuum may reflect an effect on broad depressive phenotypes but could also be accounted for by an association with low levels of general well-being (12–18% heritability) that co-occur with depressive symptoms (48). Second, we used a P-value based meta-analysis, as effect estimates could not be directly evaluated in a straightforward manner. Third, the heterogeneity of the imputation methods used in the PGC and CHARGE discovery samples might reduce the statistical power. However, different imputation references did not change the results in the published PGC MDD study (7).
