DeepWalk-aware graph attention networks with CNN for circRNA-drug sensitivity association identification

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Circular RNAs (circRNAs) are noncoding RNA molecules crucial in human health and disease.
• Predicting circRNA-drug sensitivity associations experimentally is inefficient, costly, and time-consuming.
• There is a pressing need for advanced computational methods to predict circRNA-drug sensitivity links.

Purpose of the Study:

• To develop an efficient and accurate deep learning-based computational method for identifying circRNA-drug sensitivity associations.
• To introduce DGATCCDA, a novel approach for predicting potential circRNA-drug interactions.

Main Methods:

• Constructed multimodal networks using circRNA and drug feature information.
• Employed DeepWalk-aware graph attention networks to extract features from multimodal networks.
• Fused extracted features using layer attention and utilized an inner product approach for association matrix construction.

Main Results:

• DGATCCDA achieved an average Area Under the Receiver Operating Characteristic Curve (AUC) of 91.18% in 5-fold cross-validation.
• The method outperformed five existing state-of-the-art computational approaches.
• A case study demonstrated DGATCCDA's effectiveness in uncovering latent circRNA-drug sensitivity associations.

Conclusions:

• DGATCCDA is a highly effective computational tool for predicting circRNA-drug sensitivity associations.
• The method offers a significant improvement in efficiency and accuracy over existing techniques.
• DGATCCDA holds promise for accelerating the discovery of novel therapeutic strategies involving circRNAs.