vus·lens variant confidence auditor

It reads a genetic variant's evidence, and flags where that evidence can't be trusted for this patient.

For one variant, vus-lens gathers the public evidence (population frequency, computational predictors, clinical databases), maps it to the ACMG rules the field uses, and then reads how far that evidence actually holds up, especially for ancestries that reference databases barely cover.

Variant under analysis

…

Deterministic ACMG classification

Why this matters

Raw shows the aggregated evidence, no judgement. Confidence surfaces where that evidence has worn thin for this patient.
The evidence, source by source
Population frequency how often this exact change appears in large population databases
In-silico predictions computational models estimating whether the change harms the protein
ClinVar what clinical labs have previously reported for this variant
Claude reads all of the evidence together
claude-opus-4-8 · streamed live · audits, never decides
0 /1,277
known-pathogenic ATM / PALB2 variants tested were ever called benign: a safe-direction result (never a false benign), not an accuracy score.
Analyze another variant
Try