Association prediction of lncRNAs and diseases using multiview graph convolution neural network

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Long noncoding RNAs (lncRNAs) are crucial regulators of physiological processes and form complex disease-associated networks.
• Predicting lncRNA-disease associations is challenging due to intricate network structures and data sparsity.

Purpose of the Study:

• To develop an advanced computational method for predicting long noncoding RNA (lncRNA)-disease associations.
• To address the limitations of existing methods in handling network complexity and isolated biological entities.

Main Methods:

• Proposed MVIGCN, a graph convolutional network (GCN) framework integrating multimodal data.
• Constructed a heterogeneous network incorporating disease semantics, lncRNA similarity, and miRNA-lncRNA-disease interactions.
• Employed deep learning with attention mechanisms to model topological features and multiscale relationships.

Main Results:

• MVIGCN demonstrated enhanced prediction accuracy by prioritizing critical nodes and edges within the heterogeneous network.
• Cross-validation confirmed improved reliability compared to single-view prediction methods.
• Successfully identified potential disease-related lncRNA biomarkers.

Conclusions:

• MVIGCN offers a powerful, scalable computational strategy for decoding lncRNA functions in disease biology.
• The method advances network-based approaches for identifying therapeutic targets and understanding disease mechanisms.
• Highlights the potential of multimodal data integration for improving lncRNA-disease association prediction.