Predicting interface residues for protein polymers based on module division

Area of Science:

• Computational biology
• Structural biology
• Biochemistry

Background:

• AlphaFold has advanced protein structure prediction, but accuracy for multimeric complexes requires improvement.
• Accurate prediction of binding sites in protein oligomers is crucial for understanding complex function, affinity, and specificity.

Purpose of the Study:

• To develop a non-black-box method for predicting protein-protein interaction interfaces.
• To enhance the prediction of binding sites in protein oligomers, particularly trimers and tetramers.

Main Methods:

• Investigated protein interaction sites in the Protein Data Bank (PDB) using a modular division method.
• Ranked modules using a combination of solvent-accessible surface area and internal contact area (SSAIA).

Main Results:

• Interface modules were found in surface modules with lower SSAIA values.
• 92.25% of monomer interface residues were located in the three modules with the smallest SSAIA values.
• High prediction accuracy was achieved for trimers (98.63%) and tetramers (98.92%), with an average of 3 out of 4 residues in top modules being interface residues.

Conclusions:

• The SSAIA-based method effectively predicts binding sites in multimeric protein interactions.
• This approach offers a new perspective for improving protein complex structure prediction and elucidating biological mechanisms.