2, see Supplementary Table S18A for differences in effect sizes between depression-subtypes). Likewise, individuals with depression who had transitioned to bipolar disorder had a significantly increased PRS for 3 out of 9 psychiatric phenotypes tested compared to those with depression without a later bipolar disorder diagnosis, including BP-PRS (P=6.1 x 10−17), SZ-PRS (P=3.7 x 10−12) and DEP-PRS (P=8.6 x 10−5) (Extended Data Figure 3, see Supplementary Table S18D for differences in effect sizes between depression subtypes). Individuals with depression and a co-diagnosis of schizophrenia showed increased PRS load for all 9 psychiatric phenotypes except autism (Extended Data Figure 4, see Supplementary Table S12G for differences in effect sizes between depression subtypes), most significantly for SZ-PRS (P=1.1 x 10−14), BP-PRS (P=3.0 x 10−8), substance use (SU) PRS (P=1.8 x 10−5) and DEP-PRS (P=1.9 x 10−5). SUD comorbidities showed highly increased PRS loads for 8 out of 9 psychiatric phenotypes (Extended Data Figure 5, see Supplementary Table S18J for differences in effect sizes between depression subtypes and Supplementary Figure S19 for sex-stratified analysis), most significantly for SU-PRS (P=7.5 x 10−86), ADHD-PRS (P=3.9 x 10−32) and SUD-PRS (P=1.8 x 10−30).
