Distinct mutational landscapes for germline and somatic cancer variants in forty tumor suppressor genes.

Germline and somatic tumour suppressor gene variants share a loss-of-function mechanism but are distributed differently by molecular consequence and cDNA position, shaped by distinct selection pressures and mutational mechanisms.

The authors compared 32,941 high-quality pathogenic or likely pathogenic germline variants from ClinVar with 12,907 oncogenic or likely oncogenic somatic tumour variants from cBioPortal across 40 tumour suppressor genes. Only 3,863 variants (9.2%) were shared between the two sets, and 18 genes showed significantly different distributions by molecular consequence, replicated using non-overlapping somatic data from COSMIC (chi-squared tests, false discovery rate 5%). DICER1, TP53 and SMAD4 displayed excess somatic missense events, while nine genes including RB1 and APC showed excess somatic stop-gain events throughout the coding sequence, enriched in tissues exposed to environmental mutagens with corresponding mutation signatures. The authors identified 103 regions of preferential clustering in 39 genes (78 somatic and 25 germline), with 20 somatic clusters containing recurring frameshifts in homopolymer runs, many in microsatellite-unstable tumours. For WT1, germline variants predispose to tumours (Wilms' tumour) distinct from the majority source of somatic data (myeloid leukaemia).

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The 9.2% figure for shared variants is enough to disqualify a common practice: mining somatic databases for a frequency argument to classify a germline variant. Clustering regions overlap no better — 78 somatic against 25 germline — which immediately limits repurposing a tumour hotspot as a germline PM1-type argument. This is annotation work rather than discovery, but it targets a shortcut taken daily in variant interpretation.

Analysis by Dr Thibaut Benquey