Variational gated autoencoder-based feature extraction model for inferring disease-miRNA associations based on...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• MicroRNAs (miRNAs) are crucial regulators in biological processes and disease development.
• Accurate inference of disease-miRNA associations is vital for understanding complex human diseases.

Purpose of the Study:

• To develop an advanced computational model for predicting disease-miRNA associations.
• To extract complex contextual features from heterogeneous biological networks.

Main Methods:

• A variational gated autoencoder was employed for feature extraction.
• Integrated miRNA and disease similarity networks were constructed.
• A graph autoencoder with variational gate mechanisms extracted multilevel representations.
• A gate-based predictor utilized contrastive cross-entropy loss for association inference.

Main Results:

• The proposed model demonstrated significant performance in predicting disease-miRNA associations.
• The variational gate mechanism and contrastive cross-entropy loss proved effective.
• The model successfully extracted complex contextual features from heterogeneous networks.

Conclusions:

• The developed model offers a powerful tool for inferring disease-miRNA associations.
• This approach enhances our understanding of miRNA roles in disease.
• The findings highlight the potential of advanced deep learning techniques in bioinformatics.