Integrating Multiomics Data Using a Correlation Based Graph Attention Network for Subtype Classification in Lower Grade Glioma

This version provides the updated source codes, scripts, and data preprocessing pipelines associated with the BioGAT-LGG study titled “A Graph Attention Network Approach for Biomarker Identification and Subtype Classification in Lower-Grade Glioma.” The update includes refined data alignment procedures, feature selection using LASSO, correlation-based graph construction, and the final implementation of the GATv2 model with multi-omics integration. All experiments were conducted using Python (PyTorch Geometric) and R for differential expression and enrichment analysis. This version supersedes the previous upload by including the complete reproducible codebase and documentation for external validation and reproducibility purposes. License: CC-BY-4.0Citation: Please cite this Zenodo record when reusing any part of the code or data.