Variant annotation — the rich retrieval target

Annotation attaches meaning to each variant:

gene and transcript affected
consequence (missense, nonsense, splice)
HGVS notation (c./p. change)
population frequency (gnomAD — rare variants are more suspicious)
clinical significance from ClinVar
associated conditions/phenotypes (OMIM)
an ACMG classification

This annotated record is dense, factual, and citation-friendly — ideal RAG content.

A good genome RAG answer cites the annotation source ("ClinVar: Pathogenic; gnomAD AF 0.00001") rather than inventing significance.