Figure 3

Similar comparisons among 6 solid tumors. 21 comparisons of gene lists show similar with separation between signal and noise. In the plot, α is given on the x axis and the pAUC-score for the randomized and alternative scores on the y axis. The pAUC test detects the difference between the distributions of alternative scores and random scores to select an optimized α*, which reaches a highest value for each comparison. We iterate a sub-sample strategy C (= 500) times to obtain an estimation of the variability of the similarity score and the random score. Each time, by bootstrapping 80% the labels of patients, we obtain the alternative effect size (signals). And by shuffling these labels of patients, we calculate the background noise of the same size. The details of the similarities are given in Table 4.