collected during the first visit to predict remission status at the second visit. In a series of analyses, the groups were further divided according to ancestry (EA, AA) and sex (male, female). Stratified analysis by ancestry was done twice: once with ancestry identified by self-report and once identified by implementing SNPrelate23 to estimate principal components from GWAS data which was subsequently used to determine EA and AA. Sex, ancestry, and features’ missing values dictated a series of analyses that included different subsets of subjects. All groups were matched on age. A full description of each of the groups can be found in Supplementary Tables S1–S4.
