Actionable genotypes beyond the coding sequence and their association with lifespan in the UK Biobank.

Pathogenic coding variants in the 81 ACMG SF v3.2 genes and core promoter variants functionally mapped by saturation mutagenesis coupled to a massively parallel reporter assay.

The authors examined associations between genetic variants and all-cause mortality in 490,086 UK Biobank participants with whole-genome sequencing, interpreting pathogenic and likely pathogenic coding variants across the 81 genes of the ACMG secondary findings v3.2 list. In parallel, saturation mutagenesis combined with a massively parallel reporter assay was applied to the core promoters of these genes to build a functional map of every possible variant and interpret the promoter variants observed in the cohort. Coding actionable genotypes were present in 3.40% of participants and associated with increased all-cause mortality (HR 1.42; 95% CI 1.34 to 1.52 in females; HR 1.31; 95% CI 1.24 to 1.39 in males). All rare variants within promoter regions showed no association with survival, whereas functionally validated promoter variants downregulating cancer genes were associated with elevated mortality risk (HR 1.15; 95% CI 1.01 to 1.31 in females; HR 1.12; 95% CI 1.00 to 1.26 in males).

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

The first part of this work is essentially a very large-scale validation of what we tell patients: a coding actionable genotype translates into measurable excess mortality, which remains the strongest argument for reporting secondary findings. The second part is far more fragile than it looks: the association emerges only in a subgroup defined post hoc by in vitro reporter activity, with hazard ratios around 1.1 whose confidence intervals touch 1.00. There is therefore no immediate practical consequence, but the annotation message matters: our genomes already contain these promoter variants, and it is our inability to interpret them, not their absence from the data, that makes them invisible.

Analysis by Dr Thibaut Benquey