Collaborative deep learning improves disease-related circRNA prediction based on multi-source functional information.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Circular RNAs (circRNAs) are increasingly recognized for their roles in biological processes and disease.
• Existing methods for predicting circRNA-disease associations often fail to fully exploit multi-view functional data.
• Efficiently integrating diverse circRNA data remains a challenge in the field.

Purpose of the Study:

• To develop a computational model for predicting potential circRNA-disease associations.
• To effectively utilize multi-view functional annotations of circRNAs for improved prediction accuracy.
• To enhance the understanding of circRNA roles in disease diagnosis and treatment.

Main Methods:

• Extraction of circRNA multi-view functional annotations and construction of circRNA association networks.
• Development of a collaborative deep learning framework for multi-view information integration.
• Application of a graph autoencoder model for predicting circRNA-disease associations.

Main Results:

• The proposed CLCDA model demonstrates superior performance in predicting candidate disease-related circRNAs compared to existing methods.
• Case studies on common diseases identified unknown circRNAs associated with them, showcasing the method's practicability.
• The model effectively integrates circRNA multi-source information and internal relationships.

Conclusions:

• CLCDA offers an efficient approach for predicting disease-related circRNAs by leveraging multi-view functional data.
• The findings support the utility of CLCDA in aiding the diagnosis and treatment of human diseases.
• This work highlights the potential of collaborative learning frameworks in circRNA research.