Global network random walk for predicting potential human lncRNA-disease associations

Area of Science:

• Genomics and Bioinformatics
• Computational Biology
• Molecular Medicine

Background:

• Long non-coding RNAs (lncRNAs) are increasingly implicated in disease pathogenesis, including various cancers.
• Experimental identification of lncRNA-disease associations is costly and time-intensive.
• There is a critical need for efficient computational methods to discover disease-related lncRNAs for biomarker development.

Purpose of the Study:

• To develop and validate a novel computational approach, GrwLDA, for predicting lncRNA-disease associations.
• To address limitations of existing methods, including the need for negative samples and applicability to novel lncRNAs and isolated diseases.

Main Methods:

• Developed a global network random walk model (GrwLDA) for predicting lncRNA-disease associations.
• The method is network-based and does not require negative samples.
• Evaluated performance using leave-one-out cross-validation (LOOCV).

Main Results:

• GrwLDA achieved high Area Under the Curve (AUC) values: 0.9449 (overall), 0.8562 (novel lncRNA), and 0.8374 (isolated disease).
• The model significantly outperformed existing computational methods.
• Case studies on colon, gastric, and kidney cancers identified top associations, with 13 of 15 validated by literature mining.

Conclusions:

• GrwLDA is a robust and universal computational tool for predicting lncRNA-disease associations.
• The method demonstrates significant potential for identifying novel lncRNA biomarkers for disease diagnosis and treatment.
• GrwLDA's ability to handle novel lncRNAs and isolated diseases enhances its utility in biomedical research.