Figure 1

Methodology Overview. The cumulative distribution function (CDF) of individual genes in the tumor tissue is calculated in a fitted Gaussian, and the mean and variance are estimated from the pooling of normal samples. Then the genes of individual cancer patients are ranked based on their p-values that indicate how far their expression values are away from the normal. The 200 differentially expressed genes (DEGs) are selected based on their ranking of p-values (smaller ones). The distance (difference) between any two patients is calculated by using the average gene set enrichment analysis (GSEA) scores of the DEGs of the two patients. Consequently, the subtyping analysis is applied on the sample distance matrix to discover the drug response subtypes and mutation modules.