Somatic mutations in cancer genomes are caused by mutational processes that involve components of DNA damage, DNA repair, and DNA modification [1]. Often, we observe specific patterns of mutational processes that drive tumor development [2], which are therefore named ‘mutational signatures’. Those signatures include base substitutions, small insertions, and deletions, as well as genome rearrangements and chromosome copy-number changes. Due to their characteristic mutational pattern in the genome, it is possible to link them to biological or chemical processes that take place in cancer. For instance, mutations driven by tobacco usage display a bias towards C to A substitutions, and mutations resulting from UV radiation exposure are biased towards C to T mutations [1].
Mutational signatures in cBioPortal
cBioPortal offers functionality to load and visualize mutational signatures data as generic assay data. On the study view interface, users can generate two types of figures visualizing mutation signatures: contribution score of a signature or a plot showing the counts per single base substitution, double base substitution, small insertions or deletions. These visualizations offer a high-level overview of the mutational landscape within a given study population, enabling initial assessment of the importance of a particular mutational signatures.
However, these plots fail to provide a comprehensive view of the specific mutations driving the observed contribution of an individual signature. Scientific publications often show mutational signatures as a histogram to provide insights on which recorded mutation classes (e.g., A[C>A]A for single-base substitution signatures) contribute to the signature, which cBioPortal could not provide.
Use case: mutational signatures in Pan-cancer analysis of whole genomes (ICGC/TCGA, Nature 2020) Â
For this use case, we use the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium dataset consisting of  2922 samples. An analysis of female subjects exhibiting a high mutation burden reveals a generally low contribution from most mutational signatures (Figure 1). Nevertheless, a distinct subpopulation of samples demonstrates a DBS2 contribution score of 1. Further investigation into the patient-specific mutational profiles within this subgroup indicates that a significant proportion of the double base substitutions are transformations of CC dinucleotides to AA, AG, AT, or GA (Figure 2).
Figure 1: Overview of the study view page for female subset with high mutation count (>170)
A closer analysis of the remaining signatures  reveals a potentially elevated contribution from ID3 (Indel Signature 3), with scores exceeding 0.6. This signature is most frequently observed in lung and non-small cell lung cancers, particularly in cases with high mutation counts, female patients, and strong contributions from both DBS2 (> 0.9) and ID3 (> 0.6). The predominant feature driving the elevated ID3 contribution is a single-base pair deletion involving either cytosine or thymine (Figure 3).
While refraining from making a statement on the biological relevance of these findings, this use case highlights the intuitive approach to view underlying mutational events contributing to mutational signatures with high observed contributions.
Loading of mutational signatures in cBioPortal                Â
Mutational signature contributions and their p-values can be calculated by various computational tools, including TempoSig [3], deconstructSigs [4], or SigProfiler [5]. Multiple versions of contributions and p-values matrices can be loaded to cBioPortal, such as SBS or DBS. To load these files in parallel, simply adjust the stable_id within the metadata file to reflect the specific version (for example, contributions_SBS or pvalues_SBS).
[1] The repertoire of mutational signatures in human cancer, Â Alexandrov et al, 2020
[2] The mutational signature comprehensive analysis toolkit (musicatk) for the discovery, prediction, and exploration of mutational signatures Chevalier A, Gurevich N, Guo T, Campbell JD (2025). musicatk: Mutational Signature Comprehensive Analysis Toolkit. doi:10.18129/B9.bioc.musicatk, R package version 2.2.0Â
[3] TempoSig, MSKCC (https://github.com/mskcc/tempoSig)Â
[4] deconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution, Rosenthal R, McGranahan N, Herrero J, Taylor B, Swanton C, Genome Biology, 2016Â
[5] SigProfiler Bioinformatic Tools, Welcome Sanger Institute (https://cancer.sanger.ac.uk/signatures/tools/)Â