Next, we applied GPA to study pairwise pleiotropy of these five psychiatric disorders without using annotation data. As shown in Table 2, our results suggest that the pleiotropy effect was strong between BPD and SCZ (p-value is essentially zero), MDD and SCZ (p-value ), BDP and MDD (p-value ), ASD and SCZ (p-value ), and ASD-BPD (p-value ); moderate between ADHD and BPD (p-value ), ADHD and SCZ (); and non-significant for all other pairs. Our results largely agree with those reported in [12] and the disagreements mainly came from those between ADHD and other disorders. The pleiotropy between ADHD and MDD was reported to be moderate in [12], while GPA did not detect this moderate effect. From single GWAS analysis of ADHD, given in Table 1, the estimated parameters ((s.e. 0.006), ) indicate that its GWAS signals were very weak. For MDD, the estimated parameter also indicates the weak marginal signals for MDD. Consequently, the marginal GWAS signals of ADHD and MDD were too weak to allow GPA to detect the pleiotropic effect between them. Since the data analysis
