Protein-Protein Interactions Prediction Based on Graph Energy and Protein Sequence Information

Area of Science:

• Bioinformatics
• Computational Biology
• Biochemistry

Background:

• Protein-protein interactions (PPIs) are crucial for understanding cellular functions.
• Experimental methods for PPI identification are often time-consuming and labor-intensive.
• Developing efficient computational approaches for PPI prediction is essential.

Purpose of the Study:

• To propose a novel computational method, PPI-GE, for predicting protein-protein interactions.
• To introduce innovative feature extraction techniques based on graph energy.
• To enhance the accuracy and reliability of PPI prediction.

Main Methods:

• Developed physicochemical graph energy and contact graph energy for feature extraction.
• Incorporated dipeptide composition for amino acid order information.
• Utilized multi-information fusion, Principal Component Analysis (PCA) for noise reduction, and a Weighted Sparse Representation-based Classification (WSRC) classifier.

Main Results:

• Achieved high prediction accuracies: 99.49% for human, 97.15% for *H. pylori*, and 99.56% for yeast using five-fold cross-validation.
• Demonstrated superior performance compared to existing methods across five independent datasets and two PPI networks.

Conclusions:

• The proposed PPI-GE method offers a reliable and accurate computational approach for identifying protein-protein interactions.
• PPI-GE's novel feature extraction and classification strategies significantly advance the field of bioinformatics.
• This method holds practical significance for accelerating biological research by efficiently predicting PPIs.