Variant Interpretation Platforms and Evidence Traceability
Holding
A variant interpretation platform should not present a conclusion without its evidence trail. The clinically relevant object is not only the variant. It is the assertion, evidence basis, rule version, reviewer, and date.
Authority
Genomic data sharing policies, clinical laboratory practice, and medical software quality expectations all require controlled handling of evidence. Even when software is not itself making a final diagnosis, it may shape clinical interpretation and therefore must preserve context.
Issue
The failure mode is stale certainty. Literature, population frequency, disease association, and classification practice can change. A platform that stores only the latest label loses the history needed to understand why a prior decision was made.
Resolution
The system should version evidence sources, classification rules, transcript references, population datasets, assertion status, and reviewer action. It should distinguish automated prioritization from reviewed interpretation and expose conflicts rather than flatten them.
Evidence Package
The evidence should include source versions, query history, rule engine version, manual review notes, classification changes, notification logs, and export records. A defensible interpretation is one that can be re-examined under the evidence available at the time.