Docking unbound proteins with MIAX: a novel algorithm for protein-protein soft docking

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Protein-protein interactions are crucial for biological processes.
• Predicting protein complex structures from unbound states is challenging.
• Existing rigid docking methods struggle with dynamic protein surfaces.

Purpose of the Study:

• To develop a novel "soft" docking methodology for unbound protein molecules.
• To enable limited, realistic interpenetration of protein side chains.
• To create a fast screening engine for predicting protein complex structures.

Main Methods:

• A filtering process, akin to image processing, simplifies atomic details.
• Extraction of a characteristic flattened molecular shape.
• Definition of a "soft" atomic layer to facilitate smooth interpenetration.
• Integration into rigid body docking based on point complementarity.

Main Results:

• The method allows limited, non-unrealistic side chain interpenetration.
• Results show fair agreement with experimentally determined complex structures.
• The algorithm excels in predicting hetero-dimer complexes with complex surfaces.
• It functions effectively as a fast screening tool for known interacting proteins.

Conclusions:

• The soft docking methodology offers a simple and embeddable approach for protein complex prediction.
• It is particularly valuable for screening large numbers of proteins from genomic projects.
• The technique enhances the study of structural-function relationships in proteomics.
• It provides a computationally efficient way to predict interactions for proteins lacking experimental complex data.