Figure 3.

Scale of Genetic Effects and Power Estimates

(A and B) Estimates of scale of genetic effects $\left(\tau \right)$ from the multivariate polygenic mixture model (MVPMM) for genome-wide association studies (GWAS) summary statistics generated using cases ascertained from hospital records and verbal questionnaire (combined) versus summary statistics generated using only cases ascertained from (A) hospital records or (B) verbal questionnaires; the dots represent the means of the posterior distributions and the error bars are the 95% highest posterior density (HPD). Phenotypes whose 95% HPD size for $\tau$ was less than 0.1 for both hospital record and verbal questionnaire comparisons are plotted.

(C and D) Statistical power to detect associations between rare genetic variants at different minor allele frequencies for (C) asthma and (D) psoriasis in the UK Biobank. Dot-dash lines show power for GWAS performed using only cases ascertained from hospital records, and dotted lines show power for GWAS performed using cases ascertained from both hospital records and verbal questionnaire data. Top panel shows power for rare risk variants and bottom panel shows power for rare protective variants. Different colors indicate power for different association effect sizes. The only parameters that differ between the dot-dash lines and dotted lines of a given color are the numbers of cases and controls; data represented by the dot-dash lines include cases that were identified from hospital record data, and data represented by the dotted lines include cases identified from either hospital record or verbal questionnaire data.