SigMA (Mutational Signatures)

Overview

SigMA is a computational tool for detecting mutational signatures — in particular COSMIC SBS3 (associated with homologous recombination deficiency, HRD) — from targeted panel sequencing data such as MSK-IMPACT. Standard mutational-signature decomposition requires whole-genome or whole-exome sequencing for reliable signal; SigMA adapts this analysis to the reduced mutation counts available from targeted panels by applying likelihood-ratio tests and machine-learning classifiers trained on WGS-calibrated data. The tool reports high-confidence and low-confidence SBS3 calls, with high-confidence calls used as HRD assignment surrogates when WGS is unavailable.

Used by

  • PMID:35764743 — SigMA applied to MSK-IMPACT targeted sequencing data from 130 MSKCC HGSOC patients with research consent to detect COSMIC SBS3 (HRD signature); high-confidence SBS3 in 48 patients, low-confidence in 30; SigMA-derived HRD status combined with BRCA1/BRCA2/DDR variant evidence and CDK12/CCNE1 subtype rules for final genomic-subtype assignment; genomic HRD status on its own had test concordance c = 0.52 in late-stage HGSOC PMID:35764743.
  • PMID:39746944 — mutational signature analysis (COSMIC v3.4, SigProfilerAssignment v0.1.8) applied to the MiMSI prospective cohort (n=5,037); MMR-D samples showed median MMR-signature contribution of 0.62 vs 0.036 in MMR-P; MSH6-loss tumors had the lowest MMR-signature contribution (median 0.42) and MLH1-loss the highest (median 0.67) among the four MMR proteins PMID:39746944.
  • PMID:39975212 — mutational signature decomposition (SigProfilerAssignment v0.1.8 against COSMIC v3.4) applied to 297 clonally expanded single melanocytes; UV-attributable SBS7 (sum of SBS7a–d) dominated HighMut cells while clock-like SBS1/SBS5 dominated LowMut cells; signature fractions were compared by Wilcoxon rank-sum (p<0.0001) PMID:39975212.

Notes

  • SBS3 in COSMIC corresponds to exposure to defective homologous recombination-based DNA repair; it is elevated in BRCA1/BRCA2-mutant tumors but also in a subset of BRCA-wildtype HRD tumors.
  • SigMA’s panel-adapted calling introduces uncertainty relative to WGS; the authors of the HGSOC multimodal study note that patients sequenced only with germline HRD-DDR panels (not MSK-IMPACT) had looser HRD assignments.
  • The tool is publicly available and was designed specifically for clinical panel data where mutation burden is low.
  • SigMA outputs are distinct from the Myriad myChoice HRD genomic instability score (GIS) and BRCA-centered germline tests; all three approaches classify HRD but with different sensitivities and specificities.

Sources

This page was processed by crosslinker on 2026-04-30.