Prediction of Ligand Binding Using an Approach Designed to Accommodate Diversity in Protein-Ligand Interactions
2011

Predicting Ligand Binding Using a New Method

Sample size: 39 publication Evidence: high

Author Information

Author(s): Lorraine Marsh

Primary Institution: Long Island University

Hypothesis

Can a new empirical potential method improve the prediction of protein-ligand binding interactions?

Conclusion

The internal consensus method outperforms traditional methods like Vina in predicting protein-ligand interactions.

Supporting Evidence

  • The internal consensus method achieved an ROC AUC of 0.996 for the DUD database.
  • It classified 87.9% of near-native complexes correctly.
  • Internal consensus outperformed Vina in distinguishing native ligands from decoys.

Takeaway

This study created a new way to predict how well small molecules stick to proteins, which can help in drug discovery.

Methodology

The method uses a neural network with 9 input factors to predict ligand binding affinities.

Potential Biases

Potential bias from the training data if it does not represent the diversity of ligand interactions.

Limitations

The method may overfit if not properly trained, and its performance can vary with different protein types.

Statistical Information

P-Value

p<2.0×10−7

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0023215

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication