MHRWR

MHRWR predicts associations between long non-coding RNAs (lncRNAs) and diseases by applying a multi-heterogeneous network-based random walk with restart to integrate lncRNA, disease, and gene relationships.

Key Features:

• Integration of Multiple Networks: Integrates lncRNA similarity, disease similarity, and gene similarity networks with known association networks (lncRNA–disease, lncRNA–gene, disease–gene) into a multi-layer heterogeneous network.
• Random Walk with Restart Algorithm: Applies the random walk with restart algorithm on the integrated network to capture both local and global association signals.
• Performance Evaluation: Evaluated using experimentally verified lncRNA–disease associations with leave-one-out cross-validation, achieving an AUC of 0.91344.
• Case Studies for Validation: Validated on case studies including colon cancer, colorectal cancer, and lung adenocarcinoma.

Scientific Applications:

• lncRNA–disease association prediction: Predicts potential lncRNA–disease associations to prioritize candidates for experimental validation.
• Biomarker and therapeutic target discovery: Aids identification of novel biomarkers and therapeutic targets by highlighting disease-associated lncRNAs.
• Cancer research: Applied to cancer studies, including colon cancer, colorectal cancer, and lung adenocarcinoma, to uncover disease-related lncRNAs.

Methodology:

Constructs an integrated multi-layer network combining lncRNA, disease, and gene similarity networks and known association networks (lncRNA–disease, lncRNA–gene, disease–gene), applies a random walk with restart on this network, and assesses performance via leave-one-out cross-validation using experimentally verified lncRNA–disease associations (AUC 0.91344).