Fig. 2

a The clustering tree shows the process performed by the Model-based method. In the first level of clustering, all 12,270 samples were divided into two sub-groups. In the second round of clustering, the first sub-group with 9318 samples was divided into eight sub-groups (1.1–1.8), and the second sub-group was split into two new sub-groups (2.1 and 2.2). And finally, the third level of clustering sub-group 1.5, with 5630 samples, was divided into eight sub-groups (1.5.1–1.5.8). b Distribution of all samples in identified subtypes. Each color corresponds to a cancer type. The X-axis shows subtypes, and Y-axis indicates the number of samples. Subtype C7 is the most populated subtype and comprises many samples from all cancer types (Kidney and Breast are observed in the C7 subtype more than other cancer types). Subtype C16, the next most populated, comprises samples from all cancer types (Pancreas, Liver, and Breast cancers are observed in C16 more than other cancer types)