iCircDA-LTR: identification of circRNA-disease associations based on Learning to Rank

Hang Wei, Yong Xu, Bin Liu*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

23 Citations (Scopus)

Abstract

Motivation: Due to the inherent stability and close relationship with the progression of diseases, circRNAs are serving as important biomarkers and drug targets. Efficient predictors for identifying circRNA-disease associations are highly required. The existing predictors consider circRNA-disease association prediction as a classification task or a recommendation problem, failing to capture the ranking information among the associations and detect the diseases associated with new circRNAs. However, more and more circRNAs are discovered. Identification of the diseases associated with these new circRNAs remains a challenging task. Results: In this study, we proposed a new predictor called iCricDA-LTR for circRNA-disease association prediction. Different from any existing predictor, iCricDA-LTR employed a ranking framework to model the global ranking associations among the query circRNAs and the diseases. The Learning to Rank (LTR) algorithm was employed to rank the associations based on various predictors and features in a supervised manner. The experimental results on two independent test datasets showed that iCircDA-LTR outperformed the other competingmethods, especially for predicting the diseases associated with new circRNAs. As a result, iCircDA-LTR ismore suitable for the real-world applications.

Original languageEnglish
Pages (from-to)3302-3310
Number of pages9
JournalBioinformatics
Volume37
Issue number19
DOIs
Publication statusPublished - 1 Oct 2021

Fingerprint

Dive into the research topics of 'iCircDA-LTR: identification of circRNA-disease associations based on Learning to Rank'. Together they form a unique fingerprint.

Cite this