## ----init------------------------------------------------------------------ library(profileScoreDist) ## ----inr------------------------------------------------------------------- data(INR) INR ## ----regularize------------------------------------------------------------ inr.reg <- regularizeMatrix(INR) inr.reg ## ----dist params----------------------------------------------------------- granularity <- 0.05 gcgran <- 0.01 gcmin <- 0.01 gcmax <- 0.99 ## gc fractions to consider gcs <- seq(gcgran*round(gcmin/gcgran), gcgran*round(gcmax/gcgran), gcgran) ## ----dist------------------------------------------------------------------ ## compute probability distributions distlist <- lapply(gcs, function(x) computeScoreDist(inr.reg, x, granularity)) ## ----plot, fig.cap="Reproduction of Figure 1 in the article by Rahmann et al."---- distlist[[50]] plotDist(distlist[[50]]) ## ----cutoffs--------------------------------------------------------------- ab5 <- scoreDistCutoffs(distlist[[50]], 500, 1, c=1, 0.05) ## 5% FDR ab5$cutoffa ## 5% FNR ab5$cutoffb ## FDR = FNR ab5$cutoffopt