Related Experiment Video
Updated: Jul 8, 2026
05:08
Application of I TASSER, trRosetta, UCSF Chimera, HADDOCK server, and HEX loria for De Novo and In Silico Design of Proteins
Published on: July 8, 2025
A combination of rescoring and refinement significantly improves protein docking performance
1Bioinformatics Program, Boston University, Boston, Massachusetts 02215, USA.
Proteins
|January 25, 2008
Summary
Combining ZDOCK/ZRANK and RosettaDock protein docking programs improves protein complex structure prediction. This integrated approach refines initial predictions, enhancing accuracy and increasing the number of top-ranked correct models.
Area of Science:
- Computational Biology
- Structural Biology
- Bioinformatics
Background:
- Protein docking is crucial for understanding protein-protein interactions and characterizing protein complexes.
- Initial docking predictions often require refinement to achieve accurate structural models.
- Existing methods like ZDOCK provide global predictions, but refinement is necessary for precise contacts and conformations.
Purpose of the Study:
- To enhance protein docking accuracy by combining the ZDOCK/ZRANK and RosettaDock algorithms.
- To evaluate the efficacy of ZRANK-based refinement using RosettaDock for protein-protein docking.
- To optimize the ZRANK energy function for improved performance in protein docking.
Main Methods:
- Applied the ZRANK algorithm to rerank ZDOCK global docking predictions.
- Utilized RosettaDock for local side-chain and rigid-body refinement of docked models.
- Compared ZRANK scoring against RosettaDock scoring and evaluated different RosettaDock perturbation sizes.
- Validated the combined protocol on a protein-protein docking benchmark.
- Optimized the ZRANK energy function using refined models.
Main Results:
- The optimal protocol involved larger RosettaDock perturbation sizes combined with ZRANK scoring.
- Refinement improved top-ranked structures in 20 out of 27 benchmark cases.
- The number of cases with top 20 predictions increased from 23 to 27.
- Optimizing the ZRANK energy function significantly improved performance.
- Using the optimized function and protocol, top-ranked (number one) hits increased from 12 to 19 and 7 to 15 for different ZDOCK versions.
Conclusions:
- The integration of independently developed docking protocols (ZDOCK/ZRANK and RosettaDock) effectively improves protein docking results.
- Employing diverse search and scoring functions enhances the accuracy of protein complex structure prediction.
- The optimized ZRANK energy function combined with the refinement protocol offers substantial improvements in identifying near-native protein docking models.
