GATv2-MultiOmics
This repository contains the implementation of "Multi-omics Data Integration for Bladder Cancer Subtype Classification Using GATv2 Graph Neural Networks." The project leverages advanced graph neural networks to enhance the classification of bladder cancer subtypes by integrating multi-omics data.
This study introduces GATv2, a novel approach utilizing Graph Attention Networks (GAT) for the integration of multi-omics data, aiming to improve bladder cancer subtype classification. By incorporating a multi-head attention mechanism, GATv2 optimizes information extraction and assigns unique attention coefficients based on pairwise metrics, enhancing the model's predictive accuracy and robustness.
- Authors: Oussama Haraja, Fedor Grigoryev, Savelii Komlev, Clara Martinez, Emil Sharafutdinov
- Supervisor: Ph.D. Farida Zehraoui
- Laboratory: IBISC Laboratory at Université d'Évry-Paris Saclay
- Date: 08/11/2023
- Model Architecture: GATv2 uses multi-head attention for integrating diverse omics data.
- Data Types Integrated: Protein expression, gene expression, miRNA expression, DNA methylation, and clinical data.
- Clone the repository:
git clone https://github.com/yourusername/GATv2-MultiOmics.git
Contributions to enhance the model or integrate additional datasets are welcome. Please submit issues or pull requests for consideration.
This project is licensed under the MIT License. See the LICENSE file for details.
Special thanks to Ph.D. Farida Zehraoui and the IBISC Laboratory team for their valuable contributions and guidance throughout this research.