VoroMQA: Assessment of protein structure quality using interatomic contact areas

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein structure prediction

Background:

• Computational structural models are vital for biological research when experimental structures are unavailable.
• The accuracy of these models hinges on their quality, making quality estimation a critical problem.
• Existing methods often rely on statistical potentials derived from distances and angles of atomic interactions.

Purpose of the Study:

• To introduce VoroMQA, a novel method for assessing the quality of protein structural models.
• To leverage interatomic contact areas, derived from Voronoi tessellation, as a more comprehensive measure of interactions.
• To provide atomic, residue, and global quality scores within a consistent range.

Main Methods:

• VoroMQA employs Voronoi tessellation to calculate interatomic contact areas.
• It integrates both explicit protein-protein and implicit protein-solvent interactions using these contact areas.
• The method generates quality scores ranging from 0 to 1 at multiple levels (atomic, residue, global).

Main Results:

• VoroMQA demonstrated strong performance in tests using CASP data.
• The method excelled in recognizing native structures and in structural model selection tasks.
• Results indicate the effectiveness of interatomic contact areas for protein structure quality assessment.

Conclusions:

• Interatomic contact areas derived from Voronoi tessellation are effective for protein model quality assessment.
• VoroMQA offers a robust and versatile approach to estimating protein structure quality.
• The VoroMQA software is publicly available as a standalone application and web server.