Reinforcement Learning-enhanced Dual-view GAT-based Multi-task Learning for Non-coding RNA-Disease Association...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Non-coding RNAs (ncRNAs), including long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), are crucial gene expression regulators implicated in disease.
• Identifying associations between ncRNAs and diseases is vital for understanding disease mechanisms.
• Existing computational methods often treat lncRNA-disease and miRNA-disease association predictions independently, missing crucial cross-task biological signals.

Purpose of the Study:

• To develop an integrated computational framework for predicting lncRNA-disease associations (LDA), miRNA-disease associations (MDA), and lncRNA-miRNA interactions (LMI).
• To address the limitations of existing models by jointly learning these interconnected tasks and adaptively tuning hyperparameters.

Main Methods:

• Proposed RL-DMGLMD (Reinforcement Learning-enhanced Dual-view Multi-task Graph learning for LncRNA-MiRNA-Disease association prediction).
• Employed a Soft Actor-Critic (SAC) controller for adaptive hyperparameter tuning.
• Utilized a unified multi-task framework with shared encoders and task-specific decoders for knowledge transfer.
• Implemented a dual-view multi-head Graph Attention Network (GAT) to learn from heterogeneous interaction and attribute graphs.

Main Results:

• RL-DMGLMD achieved high AUROC values: 0.9900/0.9872/0.9867 on Dataset 1 and 0.9946/0.9903/0.9954 on Dataset 2 for LDA, MDA, and LMI, respectively.
• The proposed method significantly outperformed state-of-the-art baseline approaches.
• Demonstrated the effectiveness of joint learning and adaptive hyperparameter optimization.

Conclusions:

• RL-DMGLMD provides a powerful and integrated approach for predicting ncRNA-disease and miRNA-disease associations.
• The method's ability to capture cross-task signals enhances understanding of disease pathogenesis.
• RL-DMGLMD serves as a practical tool for biomarker discovery and prioritizing therapeutic targets.