SVM classification and CoMSIA modeling of UGT1A6 interacting molecules
Leo Ghemtio1, Anne Soikkeli, Marjo Yliperttula
1Centre for Drug Research, ‡Division of Pharmaceutical Technology, §Division of Biopharmaceutics and Pharmacokinetics, and ∥Division of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Helsinki , 00100 Helsinki, Finland.
Journal of Chemical Information and Modeling
|March 5, 2014
Summary
This study developed predictive models for human UDP-glucuronosyltransferase 1A6 (UGT1A6) inhibitors, crucial for drug metabolism. The models accurately predict compound interactions, aiding in the design of safer pharmaceuticals.
Area of Science:
- Pharmacology
- Biochemistry
- Computational Chemistry
Background:
- The human UDP-glucuronosyltransferase 1A6 (UGT1A6) enzyme is vital for metabolizing xenobiotics, including medications.
- Understanding UGT1A6 inhibition is critical for predicting drug-drug interactions and optimizing drug efficacy.
Purpose of the Study:
- To experimentally assess inhibitory properties of various compounds against UGT1A6.
- To build and validate computational models for predicting UGT1A6-compound interactions.
Main Methods:
- Tested 46 compounds for UGT1A6 inhibitory activity during 1-naphthol glucuronidation.
- Employed support vector machines (SVM) for classification modeling with training and external test sets.
- Utilized Comparative Molecular Similarity Index Analysis (CoMSIA) with homology modeling for detailed interaction analysis.
Main Results:
- SVM models achieved high accuracy (80-81%) and correlation (0.61-0.63) in predicting compound interactions.
- The best CoMSIA model demonstrated strong predictive power (q²=0.62, r²=0.91 training; r²(pred)=0.82 test).
- CoMSIA analysis identified hydrogen bond donors and electrostatic interactions as key factors in UGT1A6 inhibition.
Conclusions:
- Developed robust computational models for predicting UGT1A6 inhibitors.
- Highlighted the significance of specific molecular features (H-bond donors, electrostatic fields) for UGT1A6 interaction.
- Findings support structure-based drug design for modulating UGT1A6 activity.
