ProteinMPNN MCP

Protein sequence design using ProteinMPNN via Docker

An MCP (Model Context Protocol) server for protein design with 7 core tools:

• Design protein sequences from PDB structures
• Score protein sequences for likelihood
• Constrained design with fixed/redesigned positions
• CA-only design using backbone coordinates
• Submit large-scale batch design jobs with async job tracking
• Monitor and retrieve design results
• List available example structures

Quick Start with Docker

Approach 1: Pull Pre-built Image from GitHub

The fastest way to get started. A pre-built Docker image is automatically published to GitHub Container Registry on every release.

# Pull the latest image
docker pull ghcr.io/macromnex/proteinmpnn_mcp:latest

# Register with Claude Code (runs as current user to avoid permission issues)
claude mcp add proteinmpnn -- docker run -i --rm --user `id -u`:`id -g` --gpus all --ipc=host -v `pwd`:`pwd` ghcr.io/macromnex/proteinmpnn_mcp:latest

Note: Run from your project directory. `pwd` expands to the current working directory.

Requirements:

• Docker with GPU support (nvidia-docker or Docker with NVIDIA runtime)
• Claude Code installed

That's it! The ProteinMPNN MCP server is now available in Claude Code.

---

Approach 2: Build Docker Image Locally

Build the image yourself and install it into Claude Code. Useful for customization or offline environments.

# Clone the repository
git clone https://github.com/MacromNex/proteinmpnn_mcp.git
cd proteinmpnn_mcp

# Build the Docker image
docker build -t proteinmpnn_mcp:latest .

# Register with Claude Code (runs as current user to avoid permission issues)
claude mcp add proteinmpnn -- docker run -i --rm --user `id -u`:`id -g` --gpus all --ipc=host -v `pwd`:`pwd` proteinmpnn_mcp:latest

Note: Run from your project directory. `pwd` expands to the current working directory.

Requirements:

• Docker with GPU support
• Claude Code installed
• Git (to clone the repository)

About the Docker Flags:

• -i — Interactive mode for Claude Code
• --rm — Automatically remove container after exit
• --user `id -u`:`id -g` — Runs the container as your current user, so output files are owned by you (not root)
• --gpus all — Grants access to all available GPUs
• --ipc=host — Uses host IPC namespace for PyTorch shared memory
• -v — Mounts your project directory so the container can access your data

---

Verify Installation

After adding the MCP server, you can verify it's working:

# List registered MCP servers
claude mcp list

# You should see 'proteinmpnn' in the output

In Claude Code, you can now use all 7 ProteinMPNN tools:

• simple_design
• sequence_scoring
• constrained_design
• ca_only_design
• submit_batch_design
• get_job_status
• get_job_result

---

Next Steps

• Detailed documentation: See detail.md for comprehensive guides on:
  • Available MCP tools and parameters
  • Local Python environment setup (alternative to Docker)
  • Example workflows and use cases
  • Configuration file options
  • Troubleshooting

---

Usage Examples

Once registered, you can use the ProteinMPNN tools directly in Claude Code. Here are some common workflows:

Example 1: Basic Protein Sequence Design

I have a protein structure at /path/to/3HTN.pdb with chains A and B. Can you use simple_design to generate 5 diverse sequences with temperature 0.1 and save results to /path/to/results/?

Example 2: Constrained Design with Fixed Positions

I want to redesign /path/to/protein.pdb while keeping key residues fixed at positions 1, 2, 3 in chain A and positions 10, 11 in chain B. Can you use constrained_design and generate 3 sequences?

Example 3: Batch Design for Multiple Structures

I have many PDB files in /path/to/structures/ directory. Can you submit a batch design job using submit_batch_design for all .pdb files, generate 5 sequences each, and save results to /path/to/results/? Monitor the job until it finishes.

---

Troubleshooting

Docker not found?

docker --version  # Install Docker if missing

GPU not accessible?

• Ensure NVIDIA Docker runtime is installed
• Check with: docker run --gpus all ubuntu nvidia-smi

Claude Code not found?

# Install Claude Code
npm install -g @anthropic-ai/claude-code

---

License

MIT — Based on ProteinMPNN by Dauparas et al.