circRNA-binding protein site prediction based on multi-view deep learning, subspace learning and multi-view...

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Circular RNAs (circRNAs) play a role in autoimmune diseases by binding to RNA-binding proteins (RBPs).
• Predicting RBP binding sites on circRNAs is crucial but challenging due to limited research and underutilized sequence characteristics.
• Existing methods primarily focus on linear RNAs and fail to fully exploit circRNA structure and composition information.

Purpose of the Study:

• To develop an efficient multi-view classification method for identifying circRNA-RBP interaction sites.
• To address the limitations of existing methods in exploiting diverse circRNA features for binding site prediction.
• To improve the accuracy of predicting specific RBP binding positions on circRNAs.

Main Methods:

• Proposed a multi-view deep learning method (DMSK) integrating sequence, structure, and composition features.
• Utilized pseudo-amino acid and pseudo-dipeptide sequences for feature extraction, RNAfold for secondary structure prediction, and sequence embedding for context-dependent features.
• Employed a hybrid CNN-LSTM network for deep feature extraction and subspace learning (VW-GCCA) for common feature extraction.
• Trained a multi-view TSK fuzzy system for classification and prediction of RBP binding sites.

Main Results:

• The DMSK method demonstrated improved prediction performance compared to existing methods.
• Successfully integrated multi-view features (sequence, structure, composition) for enhanced circRNA-RBP interaction site prediction.
• The developed multi-view classifier effectively predicted specific RBP binding sites on circRNAs.

Conclusions:

• DMSK offers a significant advancement in predicting circRNA-RBP binding sites by effectively utilizing multi-view data.
• The study highlights the importance of integrating diverse sequence and structural features for accurate circRNA-RBP interaction analysis.
• The proposed method provides a valuable tool for understanding the regulatory roles of circRNAs in diseases like autoimmune disorders.