Neighborhood-Regularized Matrix Factorization for lncRNA-Disease Association Identification.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Long non-coding RNAs (lncRNAs) play critical roles in biological processes and human diseases.
• Identifying lncRNA-disease associations is vital for disease biomarker discovery and therapeutic development.

Purpose of the Study:

• To develop an effective computational model for inferring disease-related lncRNAs.
• To enhance the accuracy and reliability of lncRNA-disease association predictions.

Main Methods:

• Proposed a recommendation-system-based model named NRMFLDA.
• Utilized matrix factorization with disease neighborhood regularization.
• Employed leave-one-out and five-fold cross-validation for performance assessment.

Main Results:

• NRMFLDA achieved high performance with AUC scores of 0.9143 and 0.8993.
• Outperformed four previously established models in predicting lncRNA-disease associations.
• Demonstrated robustness and effectiveness in identifying disease-related lncRNAs.

Conclusions:

• NRMFLDA offers an innovative approach to uncover lncRNA-disease associations.
• The model can significantly contribute to identifying novel disease biomarkers.
• Advancements in diagnostic and therapeutic strategies are expected through this research.