Figure 1From: Dissecting systems-wide data using mixture models: application to identify affected cellular processesDistributions of p -values from the Hedenfalk data set. (A,C) Density histograms for the more than 3,000 genes comparing (A) BRCA1-mutated and sporadic tumors or (C) BRCA1- and BRCA2-mutated tumors. The dashed lines show the estimated distributions of the alternative (curves) and null (horizontal lines) genes, and solid lines show the sum the estimated distributions. (B,D) Scatterplots where the p-value of each gene is plotted against the percentile rank. A smooth function is fitted (solid line) and the local maxima of the curvature (dashed line) are used to split the null and alternative populations of genes for the comparisons (B) BRCA1-mutated and sporadic tumors or (D) BRCA1- and BRCA2-mutated tumors.Back to article page