Modeling of protein complexes in CAPRI Round 37 using template-based approach combined with model selection

Area of Science:

• Computational biology
• Structural bioinformatics
• Protein structure prediction

Background:

• Protein-protein interactions (PPIs) are crucial for cellular functions.
• Accurate prediction of protein complex structures is essential for understanding biological processes.
• Computational methods are vital for predicting these complex structures.

Purpose of the Study:

• To assess the PPI3D web server for template identification in comparative modeling of protein complexes.
• To evaluate the VoroMQA method for scoring and ranking structural models of protein-protein interactions.
• To test the utility of these tools in the Critical Assessment of PRediction of Interactions (CAPRI) and Critical Assessment of protein Structure Prediction (CASP12) challenges.

Main Methods:

• Utilized PPI3D and HHpred for sequence searches to identify multimeric templates.
• Employed clustering and visual analysis in PPI3D for detecting alternative interaction interfaces.
• Applied VoroMQA, combining full structure and interface scores, for model accuracy estimation and ranking.

Main Results:

• Identified suitable templates for 7 out of 11 CAPRI targets using PPI3D.
• Successfully constructed acceptable quality models for 6 targets.
• VoroMQA effectively ranked structural models, selecting higher quality predictions from CAPRI Scoring sets.
• Manual intervention was sometimes required to select the correct interaction template among alternatives.

Conclusions:

• PPI3D and VoroMQA demonstrate utility in the structural modeling of protein-protein interactions.
• The tools show promise for aiding in the prediction of protein complex structures.
• Further improvements to PPI3D and VoroMQA are suggested based on the evaluation results.