Table 3. Polygenic Scoring.
P-value threshold |
Major depressive disorder |
Bipolar disorder |
Schizophrenia |
||||||
---|---|---|---|---|---|---|---|---|---|
NSNPs | PDiagnosis | PSeverity | NSNPs | PDiagnosis | PSeverity | NSNPs | PDiagnosis | PSeverity | |
0.05 | 103 | 0.908 | 0.907 | 176 | 0.226 | 0.077 | 140 | 0.981 | 0.631 |
0.1 | 172 | 0.339 | 0.559 | 282 | 0.429 | 0.197 | 244 | 0.831 | 0.623 |
0.2 | 309 | 0.067 | 0.127 | 446 | 0.360 | 0.119 | 381 | 0.555 | 0.285 |
0.3 | 423 | 0.093 | 0.109 | 611 | 0.479 | 0.091 | 514 | 0.675 | 0.485 |
0.4 | 530 | 0.086 | 0.073 | 736 | 0.401 | 0.041 | 621 | 0.663 | 0.383 |
0.5 | 632 | 0.067 | 0.044 | 849 | 0.397 | 0.045 | 715 | 0.433 | 0.257 |
Polygenic scores were derived for NHS participants using weights from PGC for major depressive disorder, bipolar disorder, and schizophrenia at varying p-value thresholds in the Psychiatric GWAS Consortium. For each disorder and p-value threshold, we report the number of SNPs used to derive the polygenic score (NSNPs) and the p-value testing the association between the polygenic score and the PTSD phenotypes (PDiagnosis, PSeverity) in NHS. The association between PTSD severity and the major depressive disorder and bipolar disorder polygenic scores becomes stronger at more liberal p-value thresholds. The most significant model for bipolar occurs at a p-value threshold of 0.4 (p=0.041, R2=0.3%).