Evo2

A Python-based MCP server exposing Evo 2 genomic sequence scoring, embedding, generation, and SNP scoring.

python

1

GitHub Stars

python

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "not-a-feature-evo2-mcp": {
      "command": "python",
      "args": [
        "-m",
        "evo2_mcp.main"
      ]
    }
  }
}

You run Evo 2 as an MCP server to access genomic sequence analysis tools from any MCP-compatible client. This server exposes capabilities for scoring, embedding, generating DNA sequences, and assessing the effects of variants, using multiple Evo 2 model checkpoints. It is designed for researchers and developers who want to integrate Evo 2's capabilities into larger MEC workflows or pipelines.

How to use

You interact with the Evo2 MCP server through an MCP client. The server exposes a set of tools that let you score DNA sequences, extract embeddings, generate new sequences, score SNPs, and list embedding layers or available model checkpoints. Use an MCP client to invoke each tool by name and pass the required inputs, then interpret the results returned by the server.

How to install

Prerequisites: Python 3.12 is required to run Evo 2 MCP server components.

Install Evo2 dependencies and the MCP server package with these steps.

conda install -c nvidia cuda-nvcc cuda-cudart-dev
conda install -c conda-forge transformer-engine-torch=2.3.0
pip install flash-attn==2.8.0.post2 --no-build-isolation
pip install evo2

pip install evo2-mcp

Configuration and starting the MCP server

Configure your MCP server by adding a short configuration block that defines how to start the Evo2 MCP server. The following example shows how to register the Evo2 MCP server as a local stdio process.

{
  "mcpServers": {
    "evo2-mcp": {
      "command": "python",
      "args": ["-m", "evo2_mcp.main"]
    }
  }
}

Available tools

score_sequence

Compute log probabilities for DNA sequences to assess how likely a sequence is under the Evo 2 model.

embed_sequence

Extract learned representations from intermediate layers to capture sequence features.

generate_sequence

Generate novel DNA sequences with controlled sampling settings for diversity and quality.

score_snp

Predict variant effects by scoring SNP mutations and prioritizing variants of interest.

get_embedding_layers

List available embedding layers that can be queried for representations.

list_available_checkpoints

Show supported Evo 2 model checkpoints (e.g., 7B, 40B, 1B).