Deep belief network-Based Matrix Factorization Model for MicroRNA-Disease Associations Prediction

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• MicroRNAs (miRNAs) are crucial regulators of gene expression implicated in human diseases.
• miRNAs show potential as biomarkers for disease diagnosis and prognosis.
• Computational methods offer economical and time-efficient alternatives to experimental approaches for predicting miRNA-disease associations.

Purpose of the Study:

• To propose a novel computational model, deep belief network (DBN)-based matrix factorization (DBN-MF), for predicting miRNA-disease associations.
• To leverage unsupervised learning for feature extraction of miRNAs and diseases.
• To develop an accurate and efficient computational tool for miRNA-disease association prediction.

Main Methods:

• Utilized raw interaction features from a miRNA-disease adjacent matrix.
• Employed two deep belief networks (DBNs) for unsupervised feature learning of miRNAs and diseases.
• Trained a classifier using DBNs and a cosine score function, incorporating matrix factorization for feature representation.

Main Results:

• The proposed DBN-MF model demonstrated superior performance compared to state-of-the-art methods in miRNA-disease association prediction.
• Achieved high accuracy based on 10-fold cross-validation.
• Case studies further validated the model's effectiveness.

Conclusions:

• The DBN-MF model provides a powerful computational approach for predicting miRNA-disease associations.
• This method can aid in identifying potential miRNA biomarkers for various diseases.
• The model's efficiency and accuracy contribute to advancing miRNA-related disease research.