A Novel Probability Model for LncRNA⁻Disease Association Prediction Based on the Naïve Bayesian Classifier

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Long non-coding RNAs (lncRNAs) are increasingly recognized for their roles in disease.
• Experimentally validated lncRNA-disease associations are limited, necessitating computational approaches.
• Integrating diverse biological data is a key strategy for discovering novel associations.

Purpose of the Study:

• To develop a novel computational method for predicting potential lncRNA-disease associations.
• To construct comprehensive biological networks integrating multiple data types.
• To validate the predictive performance of the proposed method.

Main Methods:

• Construction of global tripartite and quadruple networks integrating miRNA-disease, miRNA-lncRNA, lncRNA-disease, gene-lncRNA, gene-disease, and gene-miRNA interactions.
• Development of a naïve Bayesian classifier (NBCLDA) for predicting lncRNA-disease associations.
• Performance evaluation using leave-one-out cross-validation and comparison with existing methods.

Main Results:

• The NBCLDA method demonstrated reliable performance with effective area under the ROC curve (AUCs).
• The approach does not solely depend on known lncRNA-disease associations for predictions.
• Case studies on colorectal cancer, prostate cancer, and glioma showed promising results.

Conclusions:

• NBCLDA is a robust computational tool for predicting lncRNA-disease associations.
• The method effectively leverages integrated biological networks for discovery.
• NBCLDA shows potential as a valuable asset for future biomedical research.