MLNGCF: circRNA-disease associations prediction with multilayer attention neural graph-based collaborative filtering

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Circular RNAs (circRNAs) are key regulators in biological processes, with abnormal expression linked to various diseases.
• Investigating circRNA-disease associations is vital for understanding disease mechanisms.
• Traditional wet-lab validation of these associations is costly and time-consuming, necessitating computational approaches.

Purpose of the Study:

• To develop an advanced computational method for predicting circRNA-disease associations.
• To address limitations in current methods regarding the consideration of latent features in circRNA-disease interactions.

Main Methods:

• A multilayer attention neural graph-based collaborative filtering (MLNGCF) model was proposed.
• Autoencoders were used to enhance initial features of circRNAs and diseases.
• A multilayer cooperative attention mechanism was applied to a central network for high-order feature extraction.
• Neural network-based collaborative filtering was employed for prediction and model parameter updates.

Main Results:

• The proposed MLNGCF method demonstrated superior performance compared to state-of-the-art approaches.
• Experimental results on benchmark datasets validated the effectiveness of MLNGCF.
• Case studies confirmed the biological relevance of the predicted circRNA-disease associations through literature support.

Conclusions:

• MLNGCF offers a powerful and accurate computational tool for predicting circRNA-disease associations.
• The method's ability to capture latent features enhances the prediction of these critical biological links.
• The open availability of source codes and datasets facilitates further research and application.