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

Area of Science:

• Biomolecular Informatics
• Genomics
• Computational Biology

Background:

• Circular RNAs (circRNAs) are crucial biomarkers and drug targets due to their stability and disease relevance.
• Accurate prediction of circRNA-disease associations is essential but challenging, especially for newly discovered circRNAs.
• Existing methods often fail to capture ranking information and predict associations for novel circRNAs.

Purpose of the Study:

• To develop an efficient predictor for circRNA-disease association identification.
• To address the limitations of existing classification or recommendation-based approaches.
• To improve the detection of diseases linked to newly discovered circRNAs.

Main Methods:

• Proposed iCricDA-LTR, a novel predictor utilizing a ranking framework for circRNA-disease associations.
• Employed the Learning to Rank (LTR) algorithm for supervised ranking of associations based on diverse features.
• Modeled global ranking associations between query circRNAs and diseases.

Main Results:

• iCricDA-LTR demonstrated superior performance compared to existing methods on two independent test datasets.
• The predictor showed particular effectiveness in identifying diseases associated with novel circRNAs.
• Experimental results indicate iCricDA-LTR's suitability for real-world applications.

Conclusions:

• iCricDA-LTR offers an advanced approach to circRNA-disease association prediction.
• The ranking framework effectively captures complex association patterns.
• The tool provides a valuable resource for researchers in the field.