A Simple Model-Based Approach to Inferring and Visualizing Cancer Mutation Signatures

Abstract

Recent advances in sequencing technologies have enabled the production of massive amounts of data on somatic mutations from cancer genomes. These data have led to the detection of characteristic patterns of somatic mutations or “mutation signatures” at an unprecedented resolution, with the potential for new insights into the causes and mechanisms of tumorigenesis. Here we present new methods for modelling, identifying and visualizing such mutation signatures. Our methods greatly simplify mutation signature models compared with existing approaches, reducing the number of parameters by orders of magnitude even while increasing the contextual factors (e.g. the number of flanking bases) that are accounted for. This improves both sensitivity and robustness of inferred signatures. We also provide a new intuitive way to visualize the signatures, analogous to the use of sequence logos to visualize transcription factor binding sites. We illustrate our new method on somatic mutation data from urothelial carcinoma of the upper urinary tract, and a larger dataset from 30 diverse cancer types. The results illustrate several important features of our methods, including the ability of our new visualization tool to clearly highlight the key features of each signature, the improved robustness of signature inferences from small sample sizes, and more detailed inference of signature characteristics such as strand biases and sequence context effects at the base two positions 5′ to the mutated site. The overall framework of our work is based on probabilistic models that are closely connected with “mixed-membership models” which are widely used in population genetic admixture analysis, and in machine learning for document clustering. We argue that recognizing these relationships should help improve understanding of mutation signature extraction problems, and suggests ways to further improve the statistical methods. Our methods are implemented in an R package pmsignature (https://github.com/friend1ws/pmsignature) and a web application available at https://friend1ws.shinyapps.io/pmsignature_shiny/.

Identification of mutational signatures active in individual tumors

“…We have developed a tool for dissecting mutational catalogs of individual tumor samples in terms of the simplified mutational signature model proposed by Shiraishi et al [3]. The tool is implemented in a fast, user-friendly R package, decompTumor2Sig, and shows a good performance.…”

“…Mutational processes can be represented by so called "mutational signatures" [1][2][3] which reflect the occurrences of base changes within their sequence contexts (i.e., in dependence on their flanking bases). The age-related mutations 1 initiated by spontaneous deamination of 5-methylcytosine, for example, regard C>T transitions in the context of CpGs (because those are methylated; see Fig.…”

“…The second model [3] describes probabilistic mutational signatures in analogy to the way we model transcription factor binding motifs, i.e., considering the single bases of the motif as independent. Using this approach, the most frequent changes due to spontaneous deamination of 5-methylcytosine can be described as C>T transitions followed by a base that with a very high probability is a G, and preceded by any of the four bases with approximately equal probability (see Fig.…”

“…1). Instead of 6 * 4 * 4 = 96 parameters, a "Shiraishi signature", as we call it, requires only 6 + 4 + 4 = 14 parameters when considering nucleotide triplets [3]. It can therefore be more easily extended to more flanking bases (e.g., two flanking bases in each direction as shown in Fig.…”

“…The initial discovery and construction of mutational signatures requires a large amount of tumor samples, such that regular patterns can be identified [1][2][3]. This is impractical in a clinical setting, where each cancer patient is diagnosed individually.…”