Multiscale mutation clustering algorithm identifies pan-cancer mutational clusters associated with pathway-level changes in gene expression
Fig 1
Method Illustration on PIK3CA in Breast Cancer: From top to bottom.
A) Protein domains in PIK3CA. B) In gray: mutation histogram showing all mutations across the various cancer types; these data were used to generate the multiscale clusters. In blue: breast cancer mutation histogram showing all non-synonymous mutations; these data were used to assign mutation features to breast cancer tumor samples. C) Mutation clusters identified by the multiscale information-based clustering algorithm (M2C). Gray clusters have fewer than 5 mutations in breast cancer and are excluded from subsequent downstream analysis. Green clusters are assigned to breast cancer. D) LRP8 gene expression levels in breast cancer where the samples are grouped based on the mutation clusters. From left to right: “wild-type” (i.e., no non-synonymous mutations including tumors with no mutations at all), any non-synonymous mutation feature, and the seven mutation clusters assigned to breast cancer. E) Pathway association P-value heatmap showing differential pathway associations between clusters. L1S1 In Neuronal Migration and Development Pathway and Reelin Signaling Pathway do include LRP8.