# MetaMDA

**Repository Path**: Tomhappy/MetaMDA

## Basic Information

- **Project Name**: MetaMDA
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-10-09
- **Last Updated**: 2026-10-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# MetaMDA 
## Brief description:
The MetaMDA is a novel computational framework for predicting microbe–drug associations (MDAs) by performing random walks on a microbe–metabolite–drug heterogeneous network. 
## Run steps:

 1. To generate training data and test data.
 2. Run run_demo.py (located in the code folder) to train the model and obtain the predicted scores for microbe-drug associations.

# Data description:
## Input data:

 - networkf: A microbe–metabolite–drug (MMD) heterogeneous network file with the following column order: source, target, edge_type, weight, edge_id.
 - simf: A similarity file for microbe–microbe or drug–drug pairs (similarity threshold: 0.65), with columns: source, target, edge_type, weight, and edge_id.
 - nodetypef: Node type information for all entities in the MMD heterogeneous network.
 - nodef: A complete list of all nodes in the MMD heterogeneous network.
 - simnodef: A subset of nodes corresponding to microbes and drugs in the MMD network.
 - pairf: A list of positive and negative microbe–drug association pairs.

## Output data:

 - embeddingf: Vector representations (embeddings) of all nodes in the network.
 - mdaf: Prediction results of microbe–drug associations for all candidate pairs.

# Requirements:

 - python=3.8
 - networkx=2.8.8
 - numpy=1.23.3
 - pandas=1.4.4
 - keras=2.13.1
 - scikit-learn=1.1.3
 - scipy=1.9.1
 - tqdm=4.64.1
 - parmap=1.6.0
 - xgboost=1.7.4

# Contact 
Any questions, problems, or bugs are welcome and should be reported to [Bingqiang Liu]( bingqiang@sdu.edu.cn) or [Qi Wang](wangqi1994_sdu@163.com).