(with total variance of liability of 1), then the lower heritability of MDD (∼0.37, although this may be higher in clinical samples)51 compared to schizophrenia (∼0.81), must be explained by fewer risk alleles, lower risk allele frequencies and/or smaller effect sizes. It seems most plausible that MDD might have smaller effect sizes compared with schizophrenia (except in the unlikely event that the number of loci for MDD is substantially fewer than for schizophrenia). We estimate that sample sizes 4–5 times those needed for schizophrenia studies may be needed for MDD to detect variants that explain an equal proportion of the known genetic variance (Supplementary File 1).
