DSSGNN-PPI: A Protein-Protein Interactions prediction model based on Double Structure and Sequence graph neural...

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Experimentally validating protein-protein interactions (PPIs) is resource-intensive.
• Accurate prediction of PPIs is crucial for understanding cellular mechanisms.

Purpose of the Study:

• To develop an efficient and accurate method for predicting protein-protein interactions (PPIs).
• To leverage graph neural networks (GNNs) for integrating protein structure and sequence information.

Main Methods:

• A novel Double Structure and Sequence GNN for PPIs (DSSGNN-PPI) model was designed.
• Constructed a distance graph from amino acid residues and used a graph attention network to learn structural embeddings.
• Obtained sequence-based embeddings and fused both structural and sequence features.
• Employed a gated graph attention network to learn complex topological patterns from fused features.

Main Results:

• DSSGNN-PPI demonstrated remarkable effectiveness in enhancing the prediction of multi-type protein-protein interactions.
• The model successfully integrated heterogeneous multi-source information for improved PPI understanding.
• Multilevel representation learning and information fusion strategies proved effective.

Conclusions:

• DSSGNN-PPI offers a new paradigm for addressing complex problems in structural biology, particularly PPI prediction.
• The developed framework provides an effective solution for enhancing the prediction accuracy of protein interaction networks.
• The source code is publicly available on GitHub for further research and application.