Fig. 3

Patient stratification and survival analysis. a Patient grouping by mutual exclusivity of a group of gene alterations. (1) As an example, we show a group of mutually exclusively altered genes (RORC, MDM2 and TP53) having a common downstream target (HIF1A) identified by Mutex [18]. Color intensity of each gene is proportional to the alteration ratio. Green and blue edges represent transcriptional relations and post-translational relations, respectively. The patient frequency of alteration in exclusive genes is indicated above the box (i.e. 57%). (2) Distribution of mutations and copy number changes shows the mutually exclusive pattern. (3) Division of patients into two groups of altered and unaltered, and the survival plot between two groups. b Patient grouping by gene expression signature. (1) Select the expression signature genes pre-defined for each cancer type (e.g. PAM50 for breast cancer [21]). (2) Decide the mathematical function to calculate the risk score from expression values of signature genes. Samples are sorted according to the risk score as shown in the waterfall plot. (3) Select the high risk and low risk groups by moving dotted vertical lines. The survival plot shows the difference of survival rates between two groups. (4) Clinical or molecular features of patients are mapped onto the waterfall plot. (5) Expression profile of signature genes in the TCGA patient cohort (LUAD in this example). The position of our patient under study is indicated with arrow icons