Figure 2
From: Predicting mortality among ischemic stroke patients using pathways-derived polygenic risk scores
Evaluation of predictive power of PRS derived from all common variants on Geisinger ischemic stroke. PRSs were derived from MEGASTROKE by PRSice-2. LD-clumping using the following PLINK command: –clump-p1 1, –clump-kb 1000, –clump-r2 0.3 was applied to all common variants, resulting in 231,307 SNPs for p-value thresholding in the MyCode IS (n = 1756) versus non-IS (n = 19,806) patients. The results were derived over a range of p value thresholds. Nagelkerke pseudo-R2 as shown in the Y-axis, represents how much variation is explained by the model. the X-axis represents the threshold for a base p value. P value on the top of each bar represents the probability of non-zero regression coefficient with the F statistic hypothesis testing of the fit of the intercept-only (PRS excluded) model and PRS included model are equal. The regression model was adjusted by the covariates including sex, and the five main principal components(PCs).
