Rescore protein-protein docked ensembles with an interface contact statistics

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Protein complex structures are crucial for understanding biological functions.
• Accurate prediction and validation of protein complex interfaces remain challenging.
• Statistical measures can aid in assessing the quality of proposed protein complex models.

Purpose of the Study:

• To introduce and evaluate a novel statistical contact score for protein complex interfaces.
• To assess the score's ability to differentiate native protein complex structures from models.
• To explore the score's utility in improving protein docking and dimer determination.

Main Methods:

• Development of a parameter-free statistical measure for residue-residue contacts at protein interfaces.
• Calculation of a contact score based on this measure.
• Comparison of contact scores for native and model protein complex structures.
• Application of the contact score to redocking experiments and analysis of unbound protein docking.
• Exploration of the score's use in determining biological dimers from crystal structures.

Main Results:

• The developed contact score generally favors native protein complex structures over models.
• The score alone did not reliably identify the native structure as the best.
• Adjusting docking scores with the contact score improved the ranking of native-like structures.
• Tests on unbound protein docking showed the contact score favors native-like conformations.
• The contact score demonstrated potential for improving biological dimer determination.

Conclusions:

• The novel contact score is a valuable tool for assessing the quality of protein complex structures.
• The score enhances the reliability of protein docking predictions.
• This statistical measure shows promise for improving the identification of biological dimers.