Potential circRNA-disease association prediction using DeepWalk and network consistency projection.

Area of Science:

• Biochemistry
• Genomics
• Computational Biology

Background:

• Circular RNAs (circRNAs) are increasingly recognized for their roles in disease mechanisms and as potential diagnostic biomarkers.
• Current experimental methods for identifying circRNA-disease interactions are resource-intensive, necessitating advanced computational approaches.
• Existing computational methods do not fully leverage known circRNA-disease relationships.

Purpose of the Study:

• To develop a novel computational method, DWNCPCDA, for predicting circRNA-disease associations.
• To improve the accuracy and efficiency of identifying circRNA-disease interactions.
• To provide a tool for prioritizing potential circRNAs linked to complex human diseases.

Main Methods:

• The DWNCPCDA method integrates the deep learning technique DeepWalk with network consistency projection.
• DeepWalk is used to learn node features from known circRNA-disease associations, generating circRNA-circRNA and disease-disease similarity networks.
• Network consistency projection is applied to these similarity networks to predict unobserved circRNA-disease interactions.

Main Results:

• DWNCPCDA achieved high accuracy in predicting circRNA-disease interactions, with an AUC of 0.9647 (leave-one-out cross-validation) and an average AUC of 0.9599 (five-fold cross-validation).
• Case studies demonstrated the method's ability to prioritize latent circRNAs associated with complex human diseases.
• The proposed method shows significant potential for enhancing existing similarity-based prediction techniques.

Conclusions:

• DWNCPCDA offers a promising and accurate solution for predicting circRNA-disease associations.
• The method effectively exploits known relationships to uncover novel interactions.
• This approach can aid in the discovery of new diagnostic biomarkers and enhance understanding of disease pathogenesis.