GPACDA – circRNA-Disease Association Prediction with Generating Polynomials

Circular RNA, a molecule with partially understood functions, has been implicated in various diseases. Therefore, there is a vast effort to predict associations between circular RNAs and diseases. In our recent study, we introduced circGPA, an algorithm that enables the annotation of circular RNAs with gene ontology terms through interactions with miRNAs and mRNAs. Recognizing the analytical similarity in predicting disease associations, we developed GPACDA, an extension of circGPA tailored for disease associations. The benefits of our methods include explainability, as the outputs are based on known interactions and associations, as well as the rigorous calculation of the p-value, which the circGPA algorithm can compute. We compared our method with two other tools, NCPCDA and DWNCPCDA, using a subset of the CDASOR dataset and showed that GPACDA overcomes its competitors in terms of true association ranks. Our method’s code and predictions are publicly accessible.