🌌 Aurora-MCP

Model Context Protocol (MCP) server providing access to datasets of natural and synthetic small molecules, with a focus on identifying potential mitochondrial Complex I inhibitors that may occur in plant species.

🔍 Overview

Aurora-MCP is a Model Context Protocol (MCP) server and data integration layer that connects natural-product, biodiversity, and mitochondrial-inhibitor datasets. It enables LLMs and users to query relationships between plant species, small molecules, and mitochondrial Complex I inhibition—bridging COCONUT, Laji.fi, GBIF, and AI-derived PubMed data through structured joins and metadata schemas.

Aurora-MCP is a lightweight MCP server + Hugging Face Space designed to bridge two complementary knowledge sources:

🌿 Aurora — natural-product and plant biodiversity data, mapping compounds to genera and species found in Nordic ecosystems.
🧬 Aurora-Mito-ETL — curated PubMed-derived corpus of small-molecule inhibitors of mitochondrial Complex I (NADH dehydrogenase).

Together they form a conversational dataset where ChatGPT (or any MCP-compatible LLM) can reason over structured biological data, ask questions, and perform targeted searches on small compounds, plants, and mechanistic links between them.

🧠 Concept

Goal: allow scientific dialogue with an LLM grounded in domain data, for example:

“Show me plant-derived compounds that inhibit mitochondrial Complex I.”
“Find PubMed evidence for arctigenin as a Complex I inhibitor.”
“List Nordic plants whose metabolites overlap with known ETC inhibitors.”

Aurora-MCP turns your static text/TSV data into an interactive semantic backend, exposing programmatic tools for searching, linking, and reasoning.

🧩 Key Features

Capability

Description

🗂️

File introspection

List and read data files under

./data

(merged from Aurora + Aurora-Mito-ETL).

🔎

Regex & keyword search

Query files using regex (e.g. *inhibit.

complex I

,

NADH oxidoreductase

).

🧬

PubMed integration

Auto-generate PubMed URLs from PMIDs for fast evidence lookup.

🧾

MeSH & compound tag extraction

Parse MeSH IDs, chemical codes, or compound names from TSV files.

🌿

Plant–compound linkage

Bridge plant genera/species (Aurora) with small-molecule inhibitors (Aurora-Mito-ETL).

🧩

Hugging Face Space

Streamlit demo UI to browse data and test queries visually.

🤖

MCP tools for ChatGPT

Connect directly to ChatGPT via MCP for grounded conversational access.

🧱 Architecture

aurora-mcp/ ├─ data/ # Unified domain data (from Aurora + Aurora-Mito-ETL) │ ├─ plants/ # e.g. plant_compounds.tsv, gbif_species.tsv │ ├─ inhibitors/ # e.g. pubmed_gpt.txt, mito_complex_I.tsv │ └─ reference/ │ ├─ mcp_server/ # Core MCP server implementation │ ├─ server.py # stdio entrypoint (for ChatGPT MCP) │ ├─ http.py # optional FastAPI/uvicorn HTTP endpoint │ └─ tools/ │ ├─ files.py # list_files, read_text │ ├─ search.py # regex grep │ └─ bio.py # pubmed_url, mesh_tags, unique_values │ ├─ app/ # Streamlit demo for Hugging Face Space │ └─ app.py │ ├─ configs/mcp.json # Example MCP manifest ├─ requirements.txt ├─ Dockerfile ├─ huggingface.yaml └─ README.md

🚀 Quick Start (local)

git clone https://github.com/ndaniel/aurora-mcp.git cd aurora-mcp # Optional: add Aurora + Aurora-Mito-ETL data under ./data/ mkdir -p data && cp -r ../aurora/data/* data/ mkdir -p data/inhibitors && cp -r ../aurora-mito-etl/data/* data/inhibitors/ python -m venv .venv source .venv/bin/activate pip install -r requirements.txt # Run the MCP server (stdio mode) python -m mcp_server.server

➕ Add to ChatGPT (MCP)

In ChatGPT, open Settings → Model Context Protocol → Add server
Command: python -m mcp_server.server
Working directory: this repository root (aurora-mcp)
Start chatting about compounds, plants, or PubMed evidence — all grounded in your data!

💻 Hugging Face Space

Option 1 — Python Space (recommended):

Push this repo to a new Space.
Spaces runs app/app.py (Streamlit UI).
Use the console or HTTP /mcp endpoint for MCP access.

Option 2 — Docker Space:

Build via the included Dockerfile.
Exposes:
- Streamlit UI on :7860
- MCP HTTP gateway on :8000/mcp

# Local test docker build -t aurora-mcp . docker run -p 7860:7860 -p 8000:8000 aurora-mcp

🌍 Example MCP Tools

Tool	Purpose
`list_files(root, pattern)`	List domain files under `data/` .
`read_text(path)`	Read contents of a text file.
`grep(pattern, paths, flags)`	Regex search within selected files.
`pubmed_url(pmid)`	Build PubMed link for given PMID.
`mesh_tags(paths)`	Extract MeSH or chemical IDs.
`unique_values(path, column)`	List unique values from TSV column.

You can easily extend these to include:

Compound → plant cross-references
Frequency summaries
Co-occurrence matrices
Filtered outputs for downstream ML models

🧩 Example Conversations (MCP-ChatGPT)

User:

List natural compounds found in Arctium lappa that have PubMed evidence of Complex I inhibition.

Aurora-MCP:

Arctigenin — PubMed ID 22095235
https://pubmed.ncbi.nlm.nih.gov/22095235/
Described as an AMPK activator via inhibition of mitochondrial Complex I.
Related lignans: arctiin, matairesinol.

🧪 Development

Extend tools in mcp_server/tools/.
Keep MCP responses lightweight (avoid full-file dumps).
Test new tools locally:
pytest -v

⚖️ License

GNU General Public License v3.0 (GPL-3.0)
See LICENSE for details.

👤 Author

Daniel N.
University of Helsinki — Department of Computer Science / Precision Medicine
Bioinformatics • Machine Learning • Mitochondrial Metabolism • Natural-Product Discovery

This server cannot be installed

-

security - not tested

A

license - permissive license

-

quality - not tested

How are these scores calculated?

hybrid server

The server is able to function both locally and remotely, depending on the configuration or use case.

Enables querying relationships between plant species, small molecules, and mitochondrial Complex I inhibitors by bridging natural-product, biodiversity, and PubMed datasets. Allows LLMs to perform structured searches and reasoning over biological data to identify potential plant-derived mitochondrial inhibitors.

Aurora-MCP