Establishing a GRU-GCN coordination-based prediction model for miRNA-disease associations.

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• MicroRNAs (miRNAs) are key regulators of gene expression implicated in disease.
• Wet-lab experiments for miRNA-disease associations are laborious and complex.
• Machine learning (ML) and deep learning (DL) offer powerful tools for analyzing biological data.

Purpose of the Study:

• To develop a DL model for predicting miRNA-disease associations.
• To enhance prediction accuracy using novel labeling strategies.
• To improve the efficiency of identifying miRNA-disease relationships.

Main Methods:

• Utilized experimental miRNA-disease association data from HMDD.
• Developed a DL model integrating Gated Recurrent Units (GRU) and Graph Convolutional Network (GCN).
• Implemented weight-based and majority-based labeling strategies for data classification into 'upregulated', 'downregulated', and 'nonspecific' categories.

Main Results:

• The GRU-GCN model achieved an Area Under the Curve (AUC) score of 0.8.
• Demonstrated robust efficacy in predicting potential miRNA-disease relationships.
• Successfully classified miRNA-disease associations using refined labeling approaches.

Conclusions:

• Innovative label-preprocessing methods addressed ambiguity in miRNA-disease association results.
• The developed DL model refines and predicts miRNA-disease associations.
• This approach complements traditional methods, enhancing understanding of miRNA-related disease mechanisms.