The Recognition And Classification Model Of Epoxide Hydrolases

Epoxide hydrolase (EH) as a family presents in all living organisms and represents important roles in proceeding metabolism and transformation of mutagenic xenobiotics [66,48,118]. It is evolved in transforming epoxides into compounds with decreased chemical reactivity, increased water solubility [141], and altered biological activity by the addition of water. Recent findings suggest that the EHs are novel drug targets for regulation of blood pressure, inflammation, cancer progression, and the onset of several other diseases[47,79,174,170]. However, study of the enzyme activity, substrates selectivity and specificity are still limited by the available sequences, structures and experiments. In this study, a fully analysis of EH family has been conducted and provided a solid foundation for the rational design of EH inhibitors.1. 101 EH structures and potential binding pockets were predicted by homology modeling. Secondary and three-dimensional structures indicated that the variable loop length of EH proteins as well as the physical and chemical properties of the catalytic center play a crucial role in the substrates specificity and selectivity.2. A two-layer model has been developed to distinguish epoxide hydrolases (EHs) from the other enzymes and to classify its subfamilies using its primary protein sequences. Considering the low sequence similarity in EH subfamily, statistical learning method support vector machine (SVM) has been employed in this study. The accuracies of training and testing model of distinguishing EHs from other proteins were 95.3% and 97.2% respectively. The 5-fold cross-validation accuracy of EH classification models ranged from 89.2% to 98.6%. These models also achieved a relatively higher performance with an accuracy higher than 94.9% by evaluated with the independent datasets.3. With the increasing information of EH sequences, structures, chemical kinetics and drug pharmacokinetics, this dynamic model could be used to predict the substrats specificity and selectivity of each EH subfamilies and the rational design of EH inhibitors .