Optimized Weighting Factors for Protein Docking
Author Information
Author(s): Philipp Heuser, Dietmar Schomburg
Primary Institution: Cologne University Bioinformatics Center (CUBIC)
Hypothesis
Can optimized amino acid specific weighting factors improve the accuracy of unbound protein docking?
Conclusion
Using optimized amino acid specific parameters significantly increases the number of near-native structures in the top ranks of predictions.
Supporting Evidence
- The optimized parameters increased the percentage of near-native structures within the top 10% of predictions by 10-15%.
- For enzyme-inhibitor complexes, the percentage of near-native solutions found within the top 100 ranks increased from 32% to 68%.
- The method showed a significant improvement in identifying near-native structures with low RMS values.
Takeaway
The study found a way to make better predictions about how proteins interact by using special numbers for each amino acid that help the computer understand their importance.
Methodology
The study used a docking tool to calculate geometric correlations for unbound protein-protein complexes and optimized parameters for different complex classes.
Potential Biases
The optimization process may not account for all variables affecting docking accuracy, potentially leading to biased results.
Limitations
The method may not reach the ultimate goal of unambiguously identifying near-native geometries and relies on the availability of certain structural data.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website