Mutational Signatures (v3.4 - October 2023)
RNA Single Base Substitution (RNA SBS) signatures are defined using the 192-channel for the tri-nucleotide context of every possible point nucleotide change on an RNA molecule. Because RNA molecules are single-stranded, we can infer the exact nucleotide change, thus resulting in the 192 channel, unlike DNA single base substitutions, where the strandedness of the nucleotide change cannot always be inferred. The signatures displayed here were identified from 333 non-small cell lung cancer tumours from the TRACERx (TRAcking Cancer Evolution through therapy (Rx)) cohort.
Signature extraction methods
The current set of reference signatures were extracted using a Hierarchical Dirichlet process (as introduced in Li et al, 2020, with the parameters described in Martinez-Ruiz et al, 2023), using RNA-specific single nucleotide variants from 333 non-small cell lung cancer tumours. RNA variants were called from bulk RNA-seq data using Mutect2, variants that were already present in whole-exome DNA sequencing from the same samples were removed. Strandedness of the variants was inferred from the f1r2 field from the Mutect2 output, owing to a stranded library preparation for the RNA-seq data. Only variants that were present in RNA molecules exclusively were used for calling signatures. Details on further filtering of RNA variants can be found in Martinez-Ruiz et al, 2023.
The table with the RNA variants used in the analysis and the scripts used for downstream analyses can be found here.