Reactome Database
Wire your agent to Reactome’s Analysis Service for pathway overrepresentation and expression enrichment without memorizing dozens of REST paths.
Overview
Reactome Database is an agent skill for the Build phase that documents the Reactome Analysis Service REST API for pathway and expression enrichment jobs.
Install
npx skills add https://github.com/google-deepmind/science-skills --skill reactome-databaseWhat is this skill?
- Documents 23 Analysis Service endpoints across database, identifier, batch, and token retrieval groups
- Supports single-ID GET, plaintext POST batches, multipart form uploads, and URL-sourced identifier lists
- Expression mode via TSV with #header row plus optional Homo sapiens projection variants
- Token-based result pagination and filtering by species or pathways after async analysis
- 23 documented Analysis Service endpoints grouped as database (2), identifier (2), batch identifiers (6), and token retri
Adoption & trust: 540 installs on skills.sh; 1.7k GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You need programmatic Reactome enrichment but the Analysis Service has many GET/POST routes, input formats, and token follow-ups that are not obvious from a single landing page.
Who is it for?
Indie bioinformatics or health-data builders wiring Reactome enrichment into APIs, notebooks, or agent tools.
Skip if: Builders who only need conceptual pathway biology without calling Reactome’s HTTP Analysis Service.
When should I use this skill?
Implementing or debugging HTTP calls to Reactome AnalysisService for identifier or expression enrichment.
What do I get? / Deliverables
Your agent can construct correct identifier batches, projection calls, and token retrievals so enrichment pipelines return filterable pathway results.
- Correct Analysis Service request bodies and URLs
- Token-based result fetch and filter calls
Recommended Skills
Journey fit
The skill is an API surface map for implementing enrichment calls in code or agent workflows during product build, not a one-off research brainstorm. Integrations is the canonical shelf because the payload is endpoint contracts (GET/POST, tokens, TSV formats) you embed in backends or automation.
How it compares
API integration reference for Reactome’s official service—not a local GSEA desktop substitute or a generic OpenAPI stub generator.
Common Questions / FAQ
Who is reactome-database for?
Solo and small-team developers, computational biologists, and agent users who implement Reactome overrepresentation or expression analysis in software.
When should I use reactome-database?
During Build integrations when you POST gene lists or TSV expression matrices, poll analysis tokens, or filter results by species and pathways.
Is reactome-database safe to install?
Review the Security Audits panel on this Prism page and treat outbound calls to reactome.org as network-exposed integration code you should scope in your agent policy.
SKILL.md
READMESKILL.md - Reactome Database
# Reactome Analysis Service API Reference Base URL: `https://reactome.org/AnalysisService` ## Database (2 endpoints) All endpoints use **GET**. - `/database/name` — Returns the database name - `/database/version` — Returns the database version number ## Identifier (2 endpoints) All endpoints use **GET**. - `/identifier/{id}` — Analyse a single identifier across species - `/identifier/{id}/projection` — Analyse with projection to Homo Sapiens **Parameters:** `interactors`, `species`, `includeDisease`, `pageSize`, `page`, `sortBy`, `order`, `resource` ## Identifiers - Batch Analysis (6 endpoints) All endpoints use **POST**. - `/identifiers/` — Analyse posted identifiers - `/identifiers/projection` — Analyse with projection to Homo Sapiens - `/identifiers/form` — Analyse identifiers from file upload - `/identifiers/form/projection` — File upload with projection - `/identifiers/url` — Analyse from URL - `/identifiers/url/projection` — URL analysis with projection **Content-Type:** `text/plain` for POST body, `multipart/form-data` for form. **Input format:** - One identifier per line for overrepresentation analysis - TSV with `#header` row for expression analysis (column 1: identifiers, columns 2+: numeric values) ## Token - Result Retrieval (13 endpoints) Endpoints use **GET** unless noted as POST. - `/token/{token}` — Retrieve full result by token - `/token/{token}/filter/species/{species}` — Filter result by species - `/token/{token}/filter/pathways` — Filter by posted pathway IDs (POST) - `/token/{token}/found/all` — Summary of found identifiers for posted pathways (POST) - `/token/{token}/found/all/{pathway}` — Found identifiers for one pathway - `/token/{token}/found/entities/{pathway}` — Found curated identifiers - `/token/{token}/found/interactors/{pathway}` — Found interactors - `/token/{token}/notFound` — List of not-found identifiers - `/token/{token}/page/{pathway}` — Page number for a pathway - `/token/{token}/pathways/binned` — Binned hit pathway sizes - `/token/{token}/reactions/{pathway}` — Reaction IDs for a pathway - `/token/{token}/reactions/pathways` — Reaction IDs for posted pathways (POST) - `/token/{token}/resources` — Resources summary ## Download (5 endpoints) All endpoints use **GET**. - `/download/{token}/result.json` — Full result as JSON - `/download/{token}/result.json.gz` — Full result as gzipped JSON - `/download/{token}/entities/found/{resource}/{filename}.csv` — Found identifiers CSV - `/download/{token}/entities/notfound/{filename}.csv` — Not-found identifiers CSV - `/download/{token}/pathways/{resource}/{filename}.csv` — Hit pathways CSV ## Mapping (6 endpoints) All endpoints use **POST**. - `/mapping/` — Map identifiers across species - `/mapping/projection` — Map with projection to Homo Sapiens - `/mapping/form` — Map from file upload - `/mapping/form/projection` — File upload with projection - `/mapping/url` — Map from URL - `/mapping/url/projection` — URL mapping with projection ## Import (3 endpoints) All endpoints use **POST**. - `/import/` — Import previously exported JSON - `/import/form` — Import JSON via file upload - `/import/url` — Import JSON from URL ## Report (1 endpoint) Uses **GET**. - `/report/{token}/{species}/{filename}.pdf` — Download PDF report ## Species Comparison (1 endpoint) Uses **GET**. - `/species/homoSapiens/{species}` — Compare Homo sapiens to another species ## Common Parameters - `pageSize` (int) — Results per page - `page` (int) — Page number (1-based) - `sortBy` (string) — Sort field (NAME, ENTITIES_PVALUE, ENTITIES_FDR, etc.) - `order` (string) — ASC or DESC - `resource` (string) — TOTAL, UNIPROT, ENSEMBL, etc. - `species` (string) — NCBI Taxon ID or species name - `interactors` (bool) — Include interactor data - `includeDisease` (bool) — Include disease pathways ## Supported Identifier Types UniProt, Gene Symbol, Ensembl, EntrezGene, ChEBI, OMIM, miRBase, GenBank/EMBL/DDBJ, RefPep, RefSeq, InterPro, Affymetrix, Agilent, Illu