Variant Classification Pipeline
End-to-end workflow from variant input to report generation. Integrating annotation, classification, and literature-based evidence synthesis.
Variant Classification Pipeline
End-to-end workflow from variant input to report generation.
Pipeline Overview
Understanding the data flow through each stage of the variant classification process.
Stage 1: Annotation
Enriches variants with gnomAD frequencies, computational scores (AlphaMissense, SpliceAI, REVEL, CADD), and reference database information from RefSeq, MANE, ClinVar, and ClinGen.
Stage 2: Classification
Analyzes HPO, ClinVar, UniProt, and patient data to assign pathogenic/benign/VUS labels with detailed reasoning and confidence metrics based on multi-source evidence.
Stage 3: Reporting
Uses LLM for literature retrieval and generates comprehensive report with classification label, supporting evidence, clinical recommendations, and data source citations.