Machine Learning Methods and Docking For Predicting Human Pregnane X Receptor Activation

Akash Khandelwal; Matthew D Krasowski; Erica J Reschly; Michael W Sinz; Peter W Swaan; Sean Ekins

doi:10.1021/tx800102e

Back

Machine Learning Methods and Docking For Predicting Human Pregnane X Receptor Activation

Journal article

Open access

Peer reviewed

Machine Learning Methods and Docking For Predicting Human Pregnane X Receptor Activation

Akash Khandelwal, Matthew D Krasowski, Erica J Reschly, Michael W Sinz, Peter W Swaan and Sean Ekins

Chemical research in toxicology, Vol.21(7), pp.1457-1467

07/2008

DOI: 10.1021/tx800102e

PMCID: PMC2574557

PMID: 18547065

Files and links (1)

url

https://doi.org/10.1021/tx800102eView

Published (Version of record) Open Access

Abstract

The pregnane X receptor (PXR) regulates the expression of genes involved in xenobiotic metabolism and transport. In vitro methods to screen for PXR agonists are used widely. In the current study, computational models for human PXR activators and PXR non-activators were developed using recursive partitioning (RP), random forest (RF) and support vector machine (SVM) algorithms with VolSurf descriptors. Following 10 fold randomization, the models correctly predicted 82.6 % to 98.9 % of activators and 62.0 % to 88.6 % of non-activators. The models were validated using separate test sets. The overall (n = 15) test set prediction accuracy for PXR activators with RP, RF and SVM PXR models is 80 to 93.3 %, representing an improvement over models previously reported. All models were tested with a second test set (n =145) and prediction accuracy ranged from 63−67 % overall. These test set molecules were found to cover the same area in a principal component analysis plot as the training set, suggesting the predictions were within the applicability domain. The FlexX docking method combined with logistic regression performed poorly in classifying this PXR test set compared with RP, RF and SVM, but may be useful for qualitative interpretion of interactions within the LBD. From this analysis, VolSurf descriptors and machine learning methods had good classification accuracy and made reliable predictions within the model applicability domain. These methods could be used for high throughput virtual screening to assess for PXR activation, prior to in vitro testing to predict potential drug-drug interactions.

Details

Title: Subtitle: Machine Learning Methods and Docking For Predicting Human Pregnane X Receptor Activation
Creators: Akash Khandelwal - Department of Pharmaceutical Sciences, University of Maryland, 20 Penn Street, Baltimore, MD 21201, USA
Matthew D Krasowski - Department of Pathology, University of Pittsburgh, Pittsburgh, PA 15261, USA
Erica J Reschly - Department of Pathology, University of Pittsburgh, Pittsburgh, PA 15261, USA
Michael W Sinz - Bristol-Myers Squibb Company, Research Parkway, Wallingford, CT 06492, USA
Peter W Swaan - Department of Pharmaceutical Sciences, University of Maryland, 20 Penn Street, Baltimore, MD 21201, USA
Sean Ekins - Department of Pharmaceutical Sciences, University of Maryland, 20 Penn Street, Baltimore, MD 21201, USA
Resource Type: Journal article
Publication Details: Chemical research in toxicology, Vol.21(7), pp.1457-1467
DOI: 10.1021/tx800102e
PMID: 18547065
PMCID: PMC2574557
NLM abbreviation: Chem Res Toxicol
ISSN: 0893-228X
eISSN: 1520-5010
Language: English
Date published: 07/2008
Academic Unit: Pathology
Record Identifier: 9984047740502771

Metrics

39 Record Views

63 Times Cited - Web of Science