Posterior p-values of fitted model

Using the test data glmfit_top100genes_tcga_skcm.data.Rds and glmfit_top100genes_tcga_skcm.Rds, the posterior p-values for each observation as calculated by function auc() span a range of ~0.25-0.6. The function currently uses the range of the produced p-values as is for ROC curve cutoffs. We might want to re-consider the approach, possibly with better fitted dataset(?)