NGCICM: A Novel Deep Learning-Based Method for Predicting circRNA-miRNA Interactions.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Circular RNAs (circRNAs) and microRNAs (miRNAs) are key in human disease development and serve as diagnostic biomarkers.
• Their roles and interactions in various diseases are not fully understood, necessitating advanced computational methods.
• Understanding circRNA-miRNA interactions is vital for disease diagnosis and therapeutic strategies.

Purpose of the Study:

• To develop a novel computational algorithm for predicting interactions between circRNAs and miRNAs.
• To address the gap in knowledge regarding the vast number of unknown circRNA-miRNA associations.
• To provide a robust tool for identifying potential disease-related molecular interactions.

Main Methods:

• A deep learning algorithm, NGCICM, was developed, integrating Node2vec, Graph Attention Network (GAT), Conditional Random Field (CRF), and Inductive Matrix Completion (IMC).
• A GAT-based encoder with attention mechanisms and a CRF layer was used for deep feature learning.
• An IMC-based decoder was employed to calculate interaction scores.

Main Results:

• The NGCICM method achieved high performance across cross-validation experiments.
• Area Under the Curve (AUC) values ranged from 0.9697 to 0.9980.
• Area Under the Precision-Recall Curve (AUPR) values ranged from 0.9671 to 0.9981, demonstrating significant predictive accuracy.

Conclusions:

• The NGCICM algorithm is highly effective in predicting circRNA-miRNA interactions.
• The findings highlight the potential of deep learning approaches in uncovering complex molecular relationships in diseases.
• This method offers a valuable tool for advancing research in circRNA and miRNA-mediated disease mechanisms.