tanaylohia/Ensembl_MCP

Ensembl MCP Server

Developed by Mandrake Bio — Tanay Lohia

A plug-and-play Model Context Protocol (MCP) server for interacting with Ensembl’s REST API across genes, variants, sequences, and species.
Supports advanced genomic research and tool chaining in environments like Claude Desktop.

🔍 Overview

The Ensembl MCP Server bridges the Ensembl REST API with an MCP-compatible interface. Use it for:

• Fast gene and transcript lookups
• Cross-species comparative genomics
• Variant analysis and regulatory annotation
• Full support for sequence translation, repeat masking, and batch processing

✨ Key Capabilities

1. Gene & Transcript Analysis

• lookup_gene — Retrieve detailed info via Ensembl ID or symbol
• get_transcripts — Structural info for all gene transcripts
• search_genes — Query by name, description, or identifier

2. Sequence Tools

• get_sequence — Genomic or feature-based DNA retrieval
• get_cds_sequence — Fetch CDS from transcripts
• translate_sequence — Convert DNA to protein
• Repeat masking (hard/soft) supported

3. Comparative Genomics

• get_homologs — Orthologs/paralogs across species
• get_gene_tree — Phylogenetic trees in multiple formats

4. Variant Exploration

• get_variants — Retrieve variants from genomic regions
• get_variant_consequences — Predict variant impact on genes

5. Regulatory Feature Mining

• get_regulatory_features — Promoters, enhancers, TFBS
• get_motif_features — Explore TF binding motifs

6. Annotation & Cross-References

• get_xrefs — PDB, RefSeq, EMBL and more
• map_coordinates — Convert across genome assemblies
• Access GO terms and functional metadata

7. Species & Assembly Metadata

• list_species, get_assembly_info, get_karyotype
• Chromosome banding and genome stats

8. Batch Processing

• batch_gene_lookup — Parallelize gene info extraction
• batch_sequence_fetch — Fetch multiple regions at once

⚙️ Installation

# Clone repo
git clone https://github.com/tanaylohia/Ensembl_MCP.git
cd Ensembl_MCP

# Install dependencies
npm install

# Build the server
npm run build

🧠 Using with Claude Desktop

Step-by-Step Integration

1. Build (if not already done):

   npm run build
   

2. Configure MCP in Claude Desktop:

   • Open Settings → MCP Servers
   • Add:
     • Name: ensembl
     • Command: node
     • Args: /absolute/path/to/ensembl-server/build/index.js

3. Restart Claude Desktop

Sample Prompts

• "Look up the BRCA2 gene and fetch its protein sequence"
• "What are the orthologs of TP53 in zebrafish?"
• "Get variants in chr17:43044295-43125364 with consequences"
• "Search for insulin-related genes"

🧬 Supported Inputs & Outputs

✅ Inputs

• Genomic Regions: chr1:1000000-2000000, 1:1000000-2000000
• Gene/Transcript IDs: Ensembl (ENSG...), RefSeq, gene symbols

📤 Outputs

🌍 Species Coverage

Supports all species available in Ensembl:

• Vertebrates: Human, Mouse, Zebrafish
• Plants: Rice, Arabidopsis, Wheat
• Fungi: Yeast
• Protists & Metazoa: Drosophila, C. elegans, and others

Default species: homo_sapiens unless specified

🧯 Error Handling

Robust, developer-friendly error messaging:

• Invalid inputs
• Unknown species
• REST API timeouts
• Coordinate format errors

📡 Data Sources

• Powered by the official Ensembl REST API
• Real-time querying of annotated genomic content

🤝 Contributing

Pull requests welcome — especially if you're:

• Adding new species or MCP tools
• Improving performance or modularity
• Reporting bugs or API changes

📜 License

MIT License. See LICENSE for details.

📬 Contact

Tanay Lohia
For issues, ideas, or collaborations, visit the GitHub repo

🙏 Acknowledgments

Thanks to:

• Ensembl Development Team
• Model Context Protocol (MCP) Contributors