Tobacco mutational signature SBS4 and mutation placement in predicted disordered protein regions: a context-adjusted equivalence analysis in 373 lung tumours.
Intrinsically disordered protein regions lack a stable fold and are often functionally important, but it is unknown whether tobacco-smoke mutagenesis places mutations preferentially in such regions. I analysed 373 lung tumours: 143 small-cell lung carcinomas (SCLC; published cohorts, essentially all smokers) and 230 lung adenocarcinomas (LUAD; TCGA), including 32 never-smokers. Mutations were mapped to UniProt residues and classed as disordered where AlphaFold pLDDT < 50. For each tumour I compared the observed number of mutations in disordered regions (O) with the number expected from its own 96-context mutation spectrum (E), and tested association with the smoke signature SBS4 by Poisson/negative-binomial regression with log(E) as offset. Equivalence bounds (0.833–1.2 per SD) and negative controls (SBS1 as predictor; silent mutations as outcome) were set in the analysis plan. No association was found: O/E = 1.008 per SD of SBS4 in SCLC (95% CI 0.985–1.033) and 0.9998 in LUAD (0.975–1.025), with pack-years (0.9991) and never- versus current smokers (1.017) agreeing. Effects of 1.2 per SD or larger are excluded, and the intervals bound effects to about ±3% per SD on the measured-predictor scale. The strength of the tobacco signature does not detectably track where mutations fall relative to predicted disordered protein regions.
Authors
- Arsen Turagabulov
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-16
- DOI
- https://doi.org/10.5281/zenodo.24288108
- Primary Topic
- Cancer Genomics and Diagnostics
- Type
- preprint