DSCMF: prediction of LncRNA-disease associations based on dual sparse collaborative matrix factorization

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Long non-coding RNAs (lncRNAs) are increasingly linked to human diseases.
• Identifying lncRNA-disease associations (LDAs) is crucial for disease prevention and treatment.
• Current methods for predicting LDAs are time-consuming and labor-intensive.

Purpose of the Study:

• To develop an efficient computational method for predicting lncRNA-disease associations (LDAs).
• To enhance the accuracy and reliability of LDA prediction.
• To provide a valuable tool for researchers in the field of lncRNA and disease studies.

Main Methods:

• Proposed a Dual Sparse Collaborative Matrix Factorization (DSCMF) method.
• Incorporated L2,1-norm to enhance sparsity.
• Utilized Gaussian interaction profile kernel to improve network similarity between lncRNAs and diseases.

Main Results:

• Achieved an Area Under the Curve (AUC) value of 0.8523.
• Demonstrated the effectiveness of DSCMF through ten-fold cross-validation.
• Detailed analysis of experimental results for prostate cancer, breast cancer, ovarian cancer, and colorectal cancer.

Conclusions:

• The DSCMF method shows significant promise for advancing lncRNA-disease association research.
• The developed method offers a more efficient approach to predicting LDAs.
• The study provides a valuable resource for understanding the roles of lncRNAs in human diseases.