Prediction of miRNA-Disease Association Using Deep Collaborative Filtering.

Area of Science:

• Biomedical Informatics
• Computational Biology
• Genomics

Background:

• MicroRNAs (miRNAs) play a crucial role in regulating gene expression and are implicated in various human diseases.
• Experimental identification of miRNA-disease associations is costly, time-consuming, and prone to failure.
• Computational approaches are increasingly used to predict potential miRNA-disease associations.

Purpose of the Study:

• To propose a novel computational method, DCFMDA, for predicting potential miRNA-disease associations.
• To enhance prediction accuracy by integrating deep collaborative filtering with neural network matrix factorization (NNMF) and multilayer perceptron (MLP).
• To provide an efficient alternative to experimental methods for identifying miRNA-disease links.

Main Methods:

• Developed a deep collaborative filtering framework (DCFMDA).
• Integrated NNMF to capture miRNA-disease interaction features from known associations.
• Utilized MLP with miRNA and disease similarity data to extract feature vectors.
• Merged NNMF and MLP outputs to generate a prediction matrix.

Main Results:

• Achieved an Area Under the Curve (AUC) of 0.9466 using 10-fold cross-validation.
• Demonstrated superior performance compared to existing computational methods.
• Case studies successfully predicted candidate miRNAs for breast neoplasms, colon neoplasms, kidney neoplasms, leukemia, and lymphoma.

Conclusions:

• DCFMDA is an effective computational tool for predicting miRNA-disease associations.
• The method offers a promising approach for accelerating the discovery of disease-related miRNAs.
• Findings support the utility of DCFMDA in identifying potential therapeutic targets and diagnostic biomarkers.