Biolinguistic graph fusion model for circRNA-miRNA association prediction.

Area of Science:

• Bioinformatics
• Molecular Biology
• Computational Biology

Background:

• Circular RNAs (circRNAs) and microRNAs (miRNAs) are key regulators in human diseases.
• Experimental validation of circRNA-miRNA associations (CMAs) is challenging due to cost and labor.
• Existing computational methods for CMA prediction are limited by reliance on single data types.

Purpose of the Study:

• To develop an advanced computational model for predicting circRNA-miRNA associations (CMAs).
• To overcome limitations of existing methods by integrating multiple data features.
• To provide a reliable tool for understanding disease mechanisms involving circRNAs and miRNAs.

Main Methods:

• Proposed BGF-CMAP model integrating gradient boosting decision tree, natural language processing, and graph embedding.
• Feature extraction using Word2vec for sequence attributes and graph embedding (LINE, GraphFactor) for interaction behaviors.
• Validation using extensive experimental analysis and comparison with existing computational models.

Main Results:

• BGF-CMAP achieved high prediction accuracy (82.90%) and AUC (0.9075) for circRNA-miRNA associations.
• The model demonstrated superior performance compared to other existing methods.
• Experimental validation confirmed 23 of the top 30 predicted miRNA-associated circRNAs.

Conclusions:

• BGF-CMAP offers a robust and accurate computational approach for predicting circRNA-miRNA associations.
• The model provides a valuable tool for advancing research in molecular biology and disease pathogenesis.
• BGF-CMAP can serve as a scientific basis for future studies on circRNA-miRNA interactions.