Fig 1.
Mapping mutations detected from different cancers to domain instances.
Rectangles represent protein domain instances in a given gene. Colored dots represent mutations detected in different cancer types.
Table 1.
Prevalence of predicted damaging mutations in domain instances among cancer types.
Table 2.
Patient tissue samples from selected cancer genome studies across 21 cancer types.
Fig 2.
Mutation Densities for Domain Instances across Cancers.
Box plots display mutation densities for mutated domain instances in different cancers. Outliers are shown as dots. Only predicted- damaging mutations predicted by IntOGen were used for this analysis (S1 Table).
Table 3.
Significantly mutated domain instances and corresponding genes in each cancer type.
Table 4.
Genes that encode cancer-type-specific significantly mutated domain instance and overlap with the Sleeping Beauty dataset.
Fig 3.
Clustering of significantly-mutated domain instances across 21 cancer types.
The heatmap reflects the significance of cancer-type-specific mutation density of each domain instance in different cancers. Side bars in the same color indicate domain instances encoded by the same gene, and domain instances belonging to the same domain type.
Fig 4.
Mutations in EGFR across 5 different cancers with protein structure context.
(A) The histogram displays the proportions of mutation counts detected at each residue to the total number of mutations that fall in the four different domains encoded by the gene EGFR, in five different cancers. The x-axis indicates the position of mutant residues. Mutations in different domains are shown in different colors. (B) shows the structure of the EGFR protein with epidermal growth factors colored in orange. The arrows point to enlargments of portions of the protein. The tails of the kinase domain are not shown in this structure. The structure visualization was based on Protein Data Bank structure models 1nql, 1ivo, 2jwa, 1m17 and 2gs6[47–50]. Significantly-mutated domain instances (SMDs) were shown as thicker boxes.
Fig 5.
Cancer-type-specific mutational hotspots and mutational hotspots shared by several cancer types.
A. shows the distribution of mutational hotspots for different cancer types within a given domain instance. B. shows mutational hotspot distribution patterns of different domain instances (encoded by different genes) that each correspond to the same protein domain type. Mutational hotspots are shown as balls and sticks, domain instances are shown as boxes. Mutational hotspots in different colors represent mutations in different cancer types.
Fig 6.
Distribution of mutated residues within a single gene.
(A) compares the prevalence with which mutations from a specific cancer type fall within significantly mutated domain instances (SMDs) to the prevalence of mutations in other domain instances. Genes with at least one SMD are represented on x-axis in descending order by the number of mutated residues. The length of each blue bar shows the number of the mutated residues falling in SMDs for each cancer type, the length of red bars shown the number of mutated residues falling in other domain instances within the same gene. (B) compares the fraction of mutated residues in SMDs that are hotspots in oncogenes (yellow) and tumor suppressors (green).
Fig 7.
Distribution of mutated residues in FGFR.
Sequence positions and frequencies of mutated residues in the FGFR protein are shown. Mutational hotspots for each cancer type are displayed as red dots. SMDs are shown as thicker boxes.
Table 5.
Genes that encode more than one cancer-type-specific significantly mutated domain instance.
Fig 8.
Structural context of p53 protein (PDB 3q05[59]) mutational hotspots.
Mutational hotspots shared by eight cancers are displayed as blue sticks. Liver-cancer-specific mutational hotspots are displayed as magenta sticks. The p53 protein structure is colored according to amino acid chain.
Fig 9.
Mutation distributions of different Ras domain instances and the structure of Ras domain.
(A) bar graph shows Ras domains encoded by different genes have different mutation rates across cancer types. (B) heat map shows fraction of mutations observed at each residue of a given gene in a given cancer. (C) the structure of the Ras domain encoded by the KRAS gene (PDB structural model 4lpk[63]). GTP/GDP binding sites are displayed as magenta sticks, GDP binding sites are colored in cyan.
Table 6.
Mutational hotspots observed at functional sites.
Table 7.
Domain position-based mutational hotspots shared by at least three cancers with functional annotations.