DRGCNCDA: Predicting circRNA-disease interactions based on knowledge graph and disentangled relational graph...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Circular RNAs (circRNAs) are increasingly recognized for their roles in human disease diagnosis and prognosis.
• Existing computational methods for predicting circRNA-disease associations often overlook interactions with other biological entities and latent factors.

Purpose of the Study:

• To propose a novel computational model, DRGCNCDA, to address limitations in current circRNA-disease association prediction.
• To improve the accuracy and comprehensiveness of predicting circRNA-disease associations by incorporating multi-relational data.

Main Methods:

• Constructed multi-relational graphs integrating circRNA, disease, miRNA, and lncRNA data.
• Employed a disentangled relational graph convolutional network (DRGCNCDA) to derive feature vectors for circRNAs and diseases.
• Utilized a knowledge graph model to predict circRNA-disease association affinity scores based on learned embeddings.

Main Results:

• The DRGCNCDA model demonstrated superior performance compared to existing computational methods.
• Experimental validation using 5-fold cross-validation confirmed the model's effectiveness.
• A case study highlighted the model's capability in predicting circRNA-disease associations and identifying potential candidates for biological experiments.

Conclusions:

• The DRGCNCDA model offers a significant advancement in predicting circRNA-disease associations.
• The integration of multi-relational data and advanced network techniques enhances prediction accuracy.
• DRGCNCDA provides a valuable tool for biological research, aiding in the identification of potential disease-related circRNAs.