Benchmarking Cross-Docking Strategies in Kinase Drug Discovery
Author Information
Author(s): David A. Schaller, Clara D. Christ, John D. Chodera, Andrea Volkamer
Primary Institution: Charité − Universitätsmedizin Berlin
Hypothesis
How well do different docking and pose selection strategies recapitulate experimentally observed binding modes in a realistic cross-docking scenario?
Conclusion
The study found that docking methods biased by the cocrystallized ligand are more successful in recovering binding poses than standard physics-based docking alone.
Supporting Evidence
- The docking methods biased by the cocrystallized ligand achieved a success rate of 70.4%.
- Docking into multiple structures significantly increases the chance of generating a low RMSD docking pose.
- Posit was the most successful docking method, achieving a mean bootstrapped success rate of 33.0%.
Takeaway
This study looks at how well different methods can predict where drugs will bind to proteins, finding that some methods work better than others.
Methodology
The study used a cross-docking benchmark with 589 protein structures and 423 ligands, employing three docking algorithms from the OpenEye Toolkits.
Potential Biases
The performance of docking methods may be influenced by the choice of protein structure and the similarity of cocrystallized ligands.
Limitations
The benchmark is limited to a single protein class (kinases) and may not generalize to other protein families.
Participant Demographics
The study focused on protein kinases, specifically human and mouse kinases.
Statistical Information
P-Value
0.0001
Confidence Interval
95%
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website