AlphaGenome

AlphaGenome MCP Server provides a natural language interface to Google DeepMind's AlphaGenome variant effect predictions. You can explore genomic variant effects, pathogenicity, tissue-specific impacts, and more from MCP clients using plain language queries, without writing code. This server wraps the AlphaGenome API behind easy-to-use tools and a consistent interface for rapid prototyping and research.

How to use

You interact with the AlphaGenome MCP Server through an MCP client such as Claude Desktop, Claude Code, Gemini CLI, Cursor, or Windsurf. Start by ensuring your MCP client is configured to reach the AlphaGenome MCP server, then ask natural language questions like “Use alphagenome to analyze chr19:44908684T>C” or more detailed prompts across the available modalities. The server routes your natural language query to a single AlphaGenome API endpoint through a Python bridge and returns a structured report with pathogenicity scores, expression impacts, splicing effects, and more. Use these patterns to explore multiple variants, compare effects across tissues, or generate clinical reports.

How to install

Prerequisites you need before installation are Node.js version 18 or newer, Python version 3.8 or newer, and an AlphaGenome API key from Google DeepMind. You will also install Python packages for the API integration.

pip install alphagenome numpy

# The MCP client integration is done via an MCP command shown in the setup steps.

Add the MCP server to your client

Run the MCP integration in your preferred client. The following command registers the Alphagenome MCP server with your client and uses the latest published MCP wrapper to connect to AlphaGenome. Replace YOUR_API_KEY with your actual API key.

claude mcp add alphagenome -- npx -y @jolab/alphagenome-mcp@latest --api-key YOUR_API_KEY

# Alternative configuration snippets are provided for commonly used MCP clients. Use the exact block that matches your client and copy the snippet into your config.

Verification

After configuring, restart your MCP client and run a test query such as the following to verify the setup succeeds and you receive a detailed regulatory impact report within 30–60 seconds.

"Use alphagenome to analyze chr19:44908684T>C"

Notes on usage across clients

The AlphaGenome MCP Server supports multiple MCP client integrations. You typically provide the API key either via a flag or secure config in your MCP client. The server presents 20 specialized tools as wrappers around a single API endpoint, enabling functions like pathogenicity assessment, tissue-specific analysis, variant comparison, and clinical reporting through natural language prompts.

Additional setup for common MCP clients

The following client-specific configuration examples show how to add alphagenome as an MCP server. Use the exact JSON structure or CLI invocations as shown for your client.

<details>
<summary><b>Claude Desktop</b></summary>

**Recommended method:**

claude mcp add alphagenome -- npx -y @jolab/alphagenome-mcp@latest --api-key YOUR_API_KEY

**Or manually add to `~/.config/claude/claude_desktop_config.json`**:

{ "mcpServers": { "alphagenome": { "command": "npx", "args": ["-y", "@jolab/alphagenome-mcp@latest", "--api-key", "YOUR_API_KEY"] } } }

**Test:**

"Use alphagenome to analyze chr19:44908684T>C"

</details>

<details>
<summary><b>Claude Code</b></summary>

Add to `~/.config/claude/claude_code_config.json`:

{ "mcpServers": { "alphagenome": { "command": "npx", "args": ["-y", "@jolab/alphagenome-mcp@latest", "--api-key", "YOUR_API_KEY"] } } }

**Test:**

"Use alphagenome to analyze chr19:44908684T>C"

</details>

<details>
<summary><b>Gemini CLI</b></summary>

Add to `~/.gemini/settings.json`:

{ "mcpServers": { "alphagenome": { "command": "npx", "args": ["-y", "@jolab/alphagenome-mcp@latest", "--api-key", "YOUR_API_KEY"] } } }

**Test:**

"Use alphagenome to analyze chr19:44908684T>C"

</details>

<details>
<summary><b>Cursor</b></summary>

Add to `.cursor/mcp.json` in your project root:

{ "mcpServers": { "alphagenome": { "command": "npx", "args": ["-y", "@jolab/alphagenome-mcp@latest", "--api-key", "YOUR_API_KEY"] } } }

**Test:**

"Use alphagenome to analyze chr19:44908684T>C"

</details>

<details>
<summary><b>Windsurf</b></summary>

Add to your Windsurf settings JSON:

{ "mcpServers": { "alphagenome": { "command": "npx", "args": ["-y", "@jolab/alphagenome-mcp@latest", "--api-key", "YOUR_API_KEY"] } } }

**Test:**

"Use alphagenome to analyze chr19:44908684T>C"

</details>

Security and best practices

Keep your API key secure. Do not hard-code credentials in shared configurations. Use your MCP client’s secure storage or environment management to inject the API key at runtime.

If you run analyses with sensitive variant data, ensure you follow your institution’s data handling policies and the AlphaGenome API terms of service.

Performance

Initial setup may incur a longer startup time as the bridge initializes. Subsequent variant analyses typically complete within seconds, depending on the number of modalities queried.

Available tools

predict_variant_effect

Full regulatory impact prediction across 11 modalities via a single API call wrapper.

assess_pathogenicity

Clinical pathogenicity scoring with evidence breakdown (expression, splicing, TF binding) for a given variant.

predict_tissue_specific

Compare variant effects across multiple tissues to reveal tissue-differential expression.

batch_tissue_comparison

Batch analysis of multiple variants across tissues for comparative insights.

compare_variants

Side-by-side comparison of two or more variants to rank relative impact.

compare_alleles

Allele-level comparisons at a single position (e.g., T>C vs T>G).

compare_protective_risk

Contrast protective vs risk alleles to highlight differential effects.

compare_variants_same_gene

Rank and compare multiple variants within the same gene.

predict_splice_impact

Splicing-focused analysis to assess splice site changes.

predict_expression_impact

Gene expression changes associated with the variant.

predict_tf_binding_impact

Transcription factor binding changes due to the variant.

predict_chromatin_impact

Chromatin accessibility or histone-modification related effects.

batch_modality_screen

Screen multiple variants across selected modalities for rapid prioritization.

batch_score_variants

Rank variants by aggregated regulatory impact scores.

analyze_gwas_locus

Fine-mapping and causal variant identification within a locus.

batch_pathogenicity_filter

Filter large variant sets by pathogenicity thresholds.

annotate_regulatory_context

Provide comprehensive regulatory context for a variant.

predict_allele_specific_effects

Allele-specific regulatory effects to assess differential impact.

generate_variant_report

Generate a clinical report summarizing key variant findings.

explain_variant_impact

Human-readable explanation of the variant’s impact.