Multiview Joint Learning-Based Method for Identifying Small-Molecule-Associated MiRNAs by Integrating...

Area of Science:

• Computational biology
• Pharmacogenomics
• Bioinformatics

Background:

• Drug discovery and development face challenges integrating pharmacological, genomic, and network data.
• Identifying small molecule-miRNA associations is vital for drug repositioning and understanding disease treatment mechanisms.
• Accurately predicting small molecule-miRNA links from multiomics data remains a challenge.

Purpose of the Study:

• To develop a novel framework, SMAJL, for improved prediction of small molecule-miRNA associations.
• To leverage joint learning by integrating diverse data types for more accurate predictions.
• To enhance the understanding of how small molecules regulate miRNAs in disease treatment.

Main Methods:

• Utilized enhancing matrix completion to derive small molecule-miRNA network knowledge.
• Extracted small molecule fingerprints and miRNA sequences into feature vectors.
• Developed a joint learning model using Restricted Boltzmann Machine (RBM) incorporating structure, sequence, and network data.

Main Results:

• The SMAJL model demonstrated superior performance in 5-fold cross-validation compared to four state-of-the-art methods.
• Achieved significantly higher Area Under the Curve (AUC) and Area Under the Precision-Recall Curve (AUPRC) values.
• Showcased strong robustness and predictive power through case studies.

Conclusions:

• SMAJL offers a powerful and effective approach for predicting small molecule-miRNA associations.
• The joint learning framework successfully integrates heterogeneous data for enhanced prediction accuracy.
• This method holds significant potential for advancing drug discovery and repositioning efforts.