Figure 5.

Mendelian randomisation analyses

Scatter plots of the SNP effects of childhood maltreatment on major depressive disorder (A), schizophrenia (B), and ADHD (C). Scatter plots of the SNP effects of major depressive disorder (D), schizophrenia (E), and ADHD (F) on childhood maltreatment. All units of associations are log-odds ratios. Slopes provided correspond to three different Mendelian randomisation methods used (inverse variance-weighted, weighted median, and Mendelian randomisation-Egger). The Mendelian randomisation-Egger intercept was significant only for the causal effect of childhood maltreatment on schizophrenia (B; p=0·012). SNP=single nucleotide polymorphism.