FAVOR is a portal for functional variant annotation, variant list analysis, search across biomedical entities, and graph-assisted exploration. This documentation is split into two parts: guides for using the product, and reference pages for how the system is built.
If you are using the web app, start with the portal guide below. If you need implementation details, data model notes, or system design, use the engineering reference pages under Under the hood.
Running offline?
8.9B
Variants
SNVs + indels across GRCh38
232
Annotations
Per variant, from 30+ sources
14.8M
Knowledge graph entities
Genes, diseases, drugs, and more
191M+
Graph edges
66 relationship types
These pages explain the user-facing parts of FAVOR: how to inspect annotation data, search and pivot between entities, upload a variant list, and work with the AI agent. Read them in any order.
232 annotation dimensions across 18 categories. Full catalog of scores, frequencies, and regulatory data.
Read guideQuery by variant, gene, region, or rsID across the knowledge graph.
Read guideUpload variant lists for large-scale annotation and analysis.
Read guideCommand-line tool for whole-genome functional analysis at scale.
Read guideAsk questions in natural language. Search, filter, run statistics, build visualizations.
Read guideEvery user-visible change across the portal, data, CLI, agent, and API. Newest first.
Read guideReference notes for the current implementation: architecture, graph structure, search internals, and batch processing. These pages are useful if you are integrating with FAVOR or working on the system itself. For user-visible changes, read the release notes.
The access patterns, the components, and the rules between them
Deep diveWhat lives in the graph, how queries work, what it powers
Deep diveHow the agent plans, runs tools, and self-corrects
Deep diveHow a variant upload becomes a queryable annotated variant list
Deep diveHow messy human queries become entities with confidence
Deep dive