Variant Classification & Interpretation
Automated ACMG/AMP-compliant variant classification powered by a 4M+ connection knowledge graph. Evidence-backed, phenotype-aware, and verifiable.
Automated Tools Miss What Matters Most: the Patient
Existing automated classification tools apply ACMG criteria in a vacuum. They evaluate variants based on population frequency, computational predictions, and database lookups, but they ignore the patient's clinical presentation. Without phenotype context, critical codes like PP4 (phenotype specificity), PS4 (case-control data), and cosegregation evidence are left on the table.
The result is classifications that are technically compliant but clinically incomplete. Analysts must still manually layer in phenotype-specific evidence, cross-reference gene-disease associations, and adjust classifications case by case. The bottleneck hasn't moved, it's just shifted downstream.
How AIVA Automates Variant Classification
Purpose-built capabilities for variant classification & interpretation
4M+ AIVA-KG Connections
AIVA is powered by a 4M+ connection knowledge graph linking genes, diseases, drugs, and pathways. Each classification includes structured evidence tracking so analysts can verify every applied code and its supporting data.
Phenotype-Aware Classification
Unlike other automated tools, AIVA incorporates the patient's clinical presentation directly into classification. It applies phenotype-specific ACMG codes (PP4, PS4, PP1/BS4) by matching variant-disease associations to the patient's symptoms, producing classifications that reflect clinical context, not just database lookups.
Clinically Benchmarked Accuracy
AIVA achieves 90%+ sensitivity for pathogenic variant detection, benchmarked on 8,387 clinically curated variants from the FDA-approved ClinGen eRepo dataset. Classifications are validated against real-world clinical decisions, not just computational predictions.
Reclassification on Demand
Reclassify any variant for a specific phenotype through natural language. Ask AIVA to re-evaluate a variant with different clinical context and get updated classifications with citations in under a minute.
Detection rate for pathogenic variants on FDA-approved ClinGen eRepo data
Gene-disease-drug-pathway knowledge graph powering variant classification
Full coverage of all ACMG/AMP evidence codes including phenotype-specific criteria
Automate Your Classifications
See how AIVA transforms variant classification & interpretation for your lab. Request a demo or try AIVA live today.