Binding Ligand Prediction for Proteins Using Partial Matching of Local Surface Patches
Author Information
Author(s): Sael Lee, Kihara Daisuke
Primary Institution: Purdue University
Hypothesis
Can local surface features of ligand binding pockets improve the prediction of protein functions?
Conclusion
The proposed method improves prediction results over previous global pocket shape-based methods.
Supporting Evidence
- The method showed a better performance over the previous Pocket-Surfer method.
- AUC values increased significantly with the new method compared to random predictions.
- Local surface patch comparisons can identify functional similarities between non-homologous proteins.
Takeaway
This study shows a new way to predict how proteins bind to molecules by looking closely at small parts of their surfaces.
Methodology
The study used a weighted bipartite matching algorithm to compare local surface patches of protein binding pockets.
Potential Biases
Potential bias in ligand predictions due to similarities in local surface patches across different proteins.
Limitations
The method may not account for all factors influencing protein-ligand interactions.
Participant Demographics
Proteins selected from different homologous families in the CATH database.
Statistical Information
P-Value
0.30
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website