DepMap Gene Correlation Analysis Tool - MCP Server#

TOU validation and deployment status (2026-08-16)#

Loopback discovery and correlation retrieval passed against a deterministic safe artifact fixture with finite bounded output. Four simultaneous fixture calls passed with event-loop offloading and serialized provider state. Production DepMap 24Q2 data, public publication, cross-user isolation, scale, and scientific-value validation remain incomplete. Authenticated private Platform import and owner testing passed on 2026-08-16; public publication and independent-caller authorization/isolation remain untested.

  • Operation: compute_depmap24q2_gene_correlations

  • Start: python -m tooluniverse.remote.depmap_24q2.depmap_24q2_mcp_tool

  • Endpoint: http://127.0.0.1:7002/mcp

  • Provider configuration: set DEPMAP_DATA_PATH to the reviewed provider artifact root; it initializes once per process.

  • TOU check: tu doctor --forward http://127.0.0.1:7002/mcp --json

  • Private relay: tu serve --share --forward http://127.0.0.1:7002/mcp --name validation-depmap-24q2 --workers 1

Non-loopback binding requires TOOLUNIVERSE_API_TOKEN; otherwise keep the server on loopback. The relay requires TOOLUNIVERSE_SERVICE_KEY. The current result is an artifact-contract pass, not production-dataset validation.

New-user check: python scripts/remote_validation/setup_skill_preflight.py --implementation depmap-24q2. Add --check-provider-env before launch, --live after launch, and --check-connect-prereqs before sharing. Live preflight checks exact MCP discovery only; it does not run or validate a model. The pinned relay SDK is not on PyPI and currently requires authorized GitHub repository access plus a configured SSH key, so a working local MCP server does not by itself prove that a new operator can share it.

The authenticated 2026-08-16 Platform matrix found all 30 private owner relays online and all 41 operations discoverable. This implementation was imported as unpublished owner draft(s), configured with a 120-second timeout, and invoked through /expert-sessions/{id}/test. Across the set, 38 unique operations passed return-schema and semantic validation; the three USPTO operations returned exact provider HTTP 403 and remain credential-blocked. Public publication, independent-caller authorization/isolation, broad saturation, and persistent supervision were not tested.

See the complete setup and verification guide.

A MCP tool from Prism ToolSpace for analyzing gene-gene correlations from the DepMap (Dependency Map) CRISPR knockout screening dataset. This tool processes systematic CRISPR-Cas9 knockout data from over 1,320 cancer cell lines from DepMap 24Q2 to identify genetic dependencies and co-essential gene pairs.

Authorize once, then share with one short command#

After installing this provider’s dependencies and the pinned Connect relay SDK, and exporting its required resources, run once per machine (and again after key rotation):

tu remote login
# Or import an existing protected 0600 file without sourcing it:
tu remote login --env-file /path/to/tooluniverse-service.env

Then each private share is:

tu remote share depmap-24q2

By default, tu remote login requests a short-lived device code, opens the TU Platform approval page, and polls until the signed-in user approves. No key copy/paste is required. On a headless machine, add --no-browser and open the printed link elsewhere. The CLI exchanges approval for a computer-only key, verifies /remote-servers/preflight, stores it in a local 0600 config file, and never displays it.

The share command selects the reviewed environment, checks it, starts or reuses the exact loopback MCP tool set, runs the TU Platform preflight, and keeps the private relay in the foreground until Ctrl-C. Override defaults only when needed:

tu remote share depmap-24q2 --name my-depmap-24q2-remote --workers 1

Use tu remote run depmap-24q2 for local-only operation. Sharing does not publish a tool or prove scientific accuracy.

In an interactive terminal, sharing automatically starts the same browser flow when the key is missing, expired, or revoked. A malformed or revoked explicit TOOLUNIVERSE_SERVICE_KEY fails fast instead of being silently replaced; unset or correct it, then run tu remote login. Non-interactive jobs also fail fast. Use tu remote logout to remove only the locally stored key.

Prerequisites#

1. Install Required Dependencies#

Install the required Python packages for the DepMap correlation analysis:

# Create a virtual environment for DepMap setup
uv venv depmap --python 3.10
source depmap/bin/activate
uv pip install -r requirements.txt

Data Setup#

1. Download DepMap 24Q2 Dataset#

Download the preprocessed DepMap correlation data from the Prism ToolSpace or prepare your own correlation matrices:

# Install CLI if not already
uvx --from huggingface_hub hf

# Download only the depmap_24q2 folder
uvx --from huggingface_hub hf download mims-harvard/ToolSpace \
  --repo-type dataset \
  --include "depmap_24q2/*" \
  --local-dir ./path/to/your/depmap/

Required Files:

  • Gene correlation matrix - Pairwise correlations between genes

  • P-value matrix - Statistical significance of correlations

  • Gene index - Mapping of gene symbols to matrix indices

  • Adjusted p-values (optional) - FDR-corrected p-values

Data Sources:

  • DepMap Portal: https://depmap.org/portal/download/

  • DepMap 24Q2 Release: Contains CRISPR knockout data for 1,320+ cell lines

  • CERES Algorithm: Standardized gene effect scores for dependency analysis

2. Directory Structure Setup#

Create the following directory structure for your DepMap data:

/path/to/your/depmap/
├── depmap_24q2/              # DepMap data directory
│   ├── corr_matrix.npy       # Gene correlation matrix (dense format)
│   ├── p_val_matrix.npy      # P-value matrix (dense format)
│   ├── p_adj_matrix.npy      # Adjusted p-values (optional)
│   ├── gene_idx_array.npy    # Gene symbol index array
│   └── gene_names.txt        # Gene symbols (alternative format)
│
│   # Alternative sparse format for large datasets:
│   └── gene_correlations.h5  # HDF5 sparse matrices

3. Set Environment Variable#

Set the DEPMAP_DATA_PATH environment variable to point to your DepMap installation:

# Add to your ~/.bashrc or ~/.zshrc
export DEPMAP_DATA_PATH="/path/to/your/depmap"

Input and Output Specifications#

Input Format#

The tool accepts gene symbol pairs for correlation analysis:

  • Gene Symbols: Standard HUGO gene nomenclature (e.g., “BRAF”, “TP53”, “MAPK1”)

  • Case Insensitive: Tool automatically standardizes gene symbols

  • Validation: Checks gene availability in the correlation matrix

Output Format#

The tool returns a structured JSON response with comprehensive correlation analysis:

{
  "correlation_data": {
    "correlation": 0.756,
    "p_value": 1.23e-15,
    "adjusted_p_value": 4.56e-12
  },
  "interpretation": {
    "strength": "strong",
    "significance": "significant (FDR corrected)",
    "direction": "similar",
    "biological_relationship": "co-dependent relationship (shared essential functions)",
    "summary": "DepMap analysis reveals a strong, similar correlation (r=0.756) in knockout effects between BRAF and MAPK1, suggesting co-dependent relationship (shared essential functions). This finding is significant (FDR corrected)."
  },
  ...
}

Output Fields:

  • correlation_data (dict): Statistical measures

    • correlation (float): Pearson correlation coefficient (-1.0 to 1.0)

    • p_value (float): Statistical significance of correlation

    • adjusted_p_value (float, optional): FDR-corrected p-value

  • interpretation (dict): Biological and statistical context

    • strength (str): Correlation strength classification

    • significance (str): Statistical significance interpretation

    • direction (str): Relationship type (similar vs opposing effects)

    • biological_relationship (str): Biological interpretation

    • summary (str): Comprehensive analysis summary

  • context_info (list): Analysis metadata and messages

  • error (str, optional): Error description if analysis failed

Running the MCP Server#

1. Start the Server#

# Activate the virtual environment
source depmap/bin/activate

# Set environment variable (if not in bashrc)
export DEPMAP_DATA_PATH="/path/to/your/depmap"

# Run the MCP server from the ToolUniverse repository root
python -m tooluniverse.remote.depmap_24q2.depmap_24q2_mcp_tool

2. Server Configuration#

The server runs with the following default settings:

  • Host: 127.0.0.1 (loopback; use the outbound relay for Connect)

  • Port: 7002 (configured to avoid conflicts)

  • Transport: streamable-http

  • Mode: Stateless HTTP

A deliberate non-loopback bind requires TOOLUNIVERSE_API_TOKEN; do not expose the unauthenticated provider directly.

Common Issues#

  1. Data Directory Not Found

    FileNotFoundError: DepMap data directory not found at /path/to/data

    • Ensure DEPMAP_DATA_PATH is set correctly

    • Verify the depmap_24q2/ subdirectory exists

    • Check that correlation matrices are properly downloaded

  2. Gene Symbol Not Found

    KeyError: Gene 'INVALID' not available in the DepMap correlation matrix

    • Verify gene symbol spelling (use standard HUGO nomenclature)

    • Check if gene is present in the DepMap 24Q2 dataset

    • Try alternative gene symbols or aliases

  3. Missing Correlation Data

    FileNotFoundError: No correlation data found in directory

    • Ensure correlation matrices are in the correct format (.npy or .h5)

    • Verify gene index files are present (gene_idx_array.npy or gene_names.txt)

    • Check file permissions and accessibility

References#