Prediction of Potential Associations Between MicroRNA and Disease Based on Bayesian Probabilistic Matrix...

Area of Science:

• Computational biology
• Genomics
• Biomedical informatics

Background:

• Understanding microRNA (miRNA) and disease associations is crucial for human disease research.
• Biological experiments for identifying these associations are costly and complex.
• Computational methods offer an effective complement to experimental approaches.

Purpose of the Study:

• To develop a computational model, MDBPMF, for predicting potential associations between miRNAs and diseases.
• To leverage a fully Bayesian probabilistic matrix factorization approach.
• To utilize the HMDDv2.0 database of known miRNA-disease associations.

Main Methods:

• Developed a model named microRNA and disease based on Bayesian probabilistic matrix factorization (MDBPMF).
• Employed a fully Bayesian treatment of probabilistic matrix factorization.
• Utilized Markov chain Monte Carlo methods for efficient model training on the HMDDv2.0 database.

Main Results:

• MDBPMF achieved reliable predictions for miRNA-disease associations.
• The model demonstrated an average area under the receiver operating characteristic curve of 0.8755 for eight complex diseases.
• Performance surpassed state-of-the-art methods in fivefold cross-validation.
• A case study on lung cancer validated the method's utility.

Conclusions:

• The MDBPMF model provides an efficient and reliable computational approach for predicting miRNA-disease associations.
• This method can significantly aid in understanding disease development and progression.
• The findings highlight the potential of Bayesian probabilistic matrix factorization in bioinformatics.