Project 1

Project 1: Assessing the clinical relevance of variability in mutational signatures

Background

Somatic mutational processes are numerous, both due to endogenous (e.g., DNA repair defects) and exogenous processes (e.g., tobacco and UV light), generating characteristic mutational signatures (see review by Koh et al. 2021). Although the activity of each mutational signature has been shown to vary strongly among cancers, their variability have not been rigorously investigated. We recently developed summary and test statistics to assess the variability of stochastic matrices allowing to test hypotheses about the variability of a the proportions of feature among individuals (Morrison et al. 2021) that allows to tackle these questions.

Data

• Mutational signatures for cancers from 37 cancer types (https://dcc.icgc.org/api/v1/download?fn=/PCAWG/mutational_signatures/Signatures_in_Samples/SP_Signatures_in_Samples/PCAWG_sigProfiler_SBS_signatures_in_samples.csv , https://dcc.icgc.org/api/v1/download?fn=/PCAWG/mutational_signatures/Signatures_in_Samples/SP_Signatures_in_Samples/PCAWG_SigProfiler_ID_signatures_in_samples.csv , https://dcc.icgc.org/api/v1/download?fn=/PCAWG/mutational_signatures/Signatures_in_Samples/SP_Signatures_in_Samples/PCAWG_sigProfiler_DBS_signatures_in_samples.csv)

Requirements

Scripting in R, understanding of hypothesis testing and data exploration

Steps

1. download the data
2. Install the R package FSTruct https://github.com/MaikeMorrison/FSTruct
3. Compute the statistic

Expected difficulties

Resources

• COSMIC signatures (https://cancer.sanger.ac.uk/signatures/)