Answers are grounded in this paper and cite their sources.
Recent studies have revealed that certain circRNAs are closely associated with cancer progression, making them potential tumor biomarkers or therapeutic targets. However, traditional CDA validation experiments are limited in scale, costly, and time-consuming. It is necessary to establish an efficient and rational computational method to predict associations of circRNAs and complex diseases. In this work, we develop a novel model called MMNNCDA, which is a matrix completion method based on minimizing the nuclear norm to identify potential CDAs. Firstly, we construct the circRNA similarity network relying on known CDAs and circRNA functional information. Additionally, we construct the disease similarity network, relying on CDAs and disease semantic information. Subsequently, the circRNA–disease heterogeneous network is generated through the amalgamation of disease similarity network, circRNA similarity network, and circRNA–disease association (CDA) network. Finally, the minimizing matrix nuclear norm is utilized to acquire the corresponding CDA scores. MMNNCDA achieved AUC values of 0.9757, 0.9831, 0.9773, and 0.9655 under five-fold cross validation on four benchmark datasets, respectively. Furthermore, we performed a case study on the CircR2Disease dataset, where 17 of the top 20 predicted candidate associations were confirmed by relevant literature. Another case study on gastric cancer further validates the persuasiveness of our method in predicting unknown associations. The experimental results demonstrated that MMNNCDA surpasses other competing prediction models and can effectively infer new associations of circRNA with disease.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: