Life Sciences

Overview

Life Sciences connects your AI client to four major public research APIs in one integration — no separate accounts for each database.

Use it for gene-centric biology and medicine: resolve a gene symbol across databases, inspect protein entries, read Ensembl annotations, and search or fetch PubMed papers. For chemical compounds use the separate PubChem integration; for species occurrence data use GBIF.

Data sources

Life Sciences fans out to read-only REST APIs maintained by NCBI, UniProt, and Ensembl. stackgate.ai handles authentication, rate limits, and response normalization — you activate once and call tools from any MCP client.

NCBI Entrez — Gene & PubMed

NCBI E-utilities power two databases in this integration:

NCBI requests identify stackgate.ai with a hub tool name and contact email (usage policy). Optional: add your own free NCBI API key for higher throughput (10 req/s vs 3 req/s without a key).

UniProt — proteins

The UniProt REST API returns curated protein knowledge: accessions (e.g. P38398), protein names, gene names, organism, and sequence length.

Tools: lookup_gene, get_uniprot_entry — no UniProt account or API key required.

Ensembl — genome annotation

The Ensembl REST API provides reference gene models: Ensembl IDs (e.g. ENSG00000012048), biotype, genomic coordinates, assembly, and descriptions.

Tools: lookup_gene, get_ensembl_gene — supports common model organisms (homo_sapiens, mus_musculus, rattus_norvegicus, danio_rerio, and aliases like human / mouse). Default is human.

Typical workflows

Gene research

1. lookup_gene with a symbol (e.g. BRCA1, TNNI2) — cross-source summary from NCBI Gene, UniProt, and Ensembl in one call.
2. Drill into one source with get_entrez_gene, get_uniprot_entry, or get_ensembl_gene.

Literature review

1. search_pubmed with a Boolean query (e.g. TNNI2 AND distal arthrogryposis).
2. get_pubmed_article with PMIDs from the results — metadata plus abstract text when available.

Combined example

Lookup TNNI2 → search PubMed for associated syndromes → fetch full abstracts for the most relevant PMIDs.

Prerequisites

• A stackgate.ai account

Getting credentials

1. Open Integrations → Life Sciences.
2. Click Activate — no NCBI API key is required to start.

For heavy research use, optionally add in integration settings:

• Your free NCBI API key (higher E-utilities rate limits; bypasses hub daily quota for NCBI calls)
• A contact email for NCBI fair-use identification

UniProt and Ensembl require no credentials.

Using with AI clients

• Dedicated endpoint: POST /mcp/life-sciences
• Private gateway: life-sciences__* tools (e.g. life-sciences__lookup_gene)

Connect through stackgate.ai like your other cloud integrations — HTTP MCP endpoint or private gateway with your API token. No local install.

Troubleshooting

• Hub quota exceeded (429) — daily fair-use limits apply on the shared hub NCBI key. Wait until tomorrow or add your own NCBI API key in integration settings.
• NCBI rate limit — reduce request frequency; lookup_gene counts as multiple upstream calls (Gene + UniProt + Ensembl).
• Partial gene results — lookup_gene returns data from sources that matched and lists per-source errors when one database has no match.
• Gene not found in any source — verify the symbol and organism; try granular tools on individual databases.
• Non-human organisms — pass organism (e.g. mus_musculus, mouse) on lookup and Ensembl tools; default is human (homo_sapiens).
• PubMed search syntax — supports Boolean operators (AND, OR, NOT), quoted phrases, and field tags (e.g. BRCA1[Gene], 2020:2024[Date - Publication]).
• Missing abstract — some PubMed records (letters, errata) have metadata only; get_pubmed_article returns what NCBI publishes.

Related integrations

• PubChem — chemical compounds, CIDs, properties, and bioassays
• GBIF — species taxonomy and biodiversity occurrence records

ClinVar, OMIM, and other Entrez databases are not included in this integration yet.