Attention-Guided Multiview Deep Learning Framework Uncovers miRNA-Drug Associations for Therapeutic Discovery

Area of Science:

• Biomedical Informatics
• Computational Biology
• Genomics

Background:

• MicroRNAs (miRNAs) are crucial regulators of biological processes with therapeutic potential.
• Aberrant miRNA expression impacts drug response, complicating miRNA-based therapies.
• Current methods for identifying miRNA-drug associations (MDAs) are often slow and expensive.

Purpose of the Study:

• To develop an efficient and accurate computational framework for predicting miRNA-drug associations (MDAs).
• To overcome limitations of existing methods by integrating multiple data sources.
• To enhance the prediction of therapeutic targets by leveraging multi-view learning.

Main Methods:

• Developed an attention-guided multiview deep learning framework (DLMVF).
• Integrated miRNA and drug attribute data with interaction information.
• Employed view-level attention to adaptively weight feature importance for enhanced representation learning.

Main Results:

• DLMVF achieved high performance with an AUROC of 0.9611 and AUPRC of 0.9543 on a benchmark dataset.
• Demonstrated superior accuracy, robustness, and generalization compared to existing methods.
• Case studies successfully identified potential novel miRNA-drug associations for anticancer drugs.

Conclusions:

• The DLMVF framework offers a powerful and efficient approach for predicting miRNA-drug associations.
• This method facilitates the discovery of new therapeutic strategies by integrating multi-source data.
• The findings support the advancement of miRNA-based therapeutics through improved computational prediction.