MicroRNA target gene prediction model based on input-feature dependency and sample data expansion technique.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Predicting microRNA (miRNA) target genes is crucial for elucidating miRNA biological functions.
• Current prediction methods face challenges due to data limitations and class imbalance.

Purpose of the Study:

• To develop an advanced miRNA target gene prediction model.
• To enhance the accuracy and reliability of miRNA-gene interaction predictions.

Main Methods:

• Utilized input-feature dependency modeling.
• Employed Gaussian mixture models (GMM) for marginal density estimation.
• Applied regular vine (R-vine) copula to capture variable dependencies.
• Implemented Bayes' rule for posterior probability calculation.
• Employed hybrid distribution mega-trend diffusion for data augmentation.

Main Results:

• The developed model demonstrated high predictive performance, even with limited training data (30%).
• Experimental validation confirmed a predicted interaction (miR-8485 targeting JAK2) using dual-luciferase, cellular, and animal models.
• The model effectively addresses data insufficiency and class imbalance issues.

Conclusions:

• The novel miRNA target gene prediction model offers a valuable tool for biological research.
• Findings contribute to a deeper understanding of miRNA functions and their roles in disease mechanisms.
• This approach facilitates more accurate identification of miRNA-gene interactions.