Study of human dopamine sulfotransferases based on gene expression programming.

A quantitative model is developed to predict the Km of 47 human dopamine sulfotransferases by gene expression programming. Each kind of compound is represented by several calculated structural descriptors of moment of inertia A, average electrophilic reactivity index for a C atom, relative number of triple bonds, RNCG relative negative charge, HA-dependent HDSA-1, and HBCA H-bonding charged surface area. Eight fitness functions of the gene expression programming method are used to find the best nonlinear model. The best quantitative model with squared standard error and square of correlation coefficient are 0.096 and 0.91 for training data set, and 0.102 and 0.88 for test set, respectively. It is shown that the gene expression programming-predicted results with fitness function are in good agreement with experimental ones.

Location of Si vacancies and [Ti(OSi)4] and [Ti(OSi)3OH] sites in the MFI framework: a large cluster and full ab initio study.

Titanium silicalite-1 (TS-1) is an important catalyst for selective oxidation reactions. However, the nature and structure of the active sites and the mechanistic details of the catalytic reactions over TS-1 have not been well-understood, leaving a continuous debate on the genesis of active sites on the TS-1 surface in the literature. In this work, the location of Si vacancies and [Ti(OSi)(4)] and [Ti(OSi)(3)OH] sites in the MFI (Framework Type Code of ZSM-5 (Zeolite Socony Mobile-Five)) framework has been studied using a full ab initio method with 40T clusters with a Si:Ti molar ratio of 39:1. It was shown that the former four energetically favorable sites for Si vacancies are T6, T12, T4, and T8 and for Ti centers of [Ti(OSi)(4)] are T10, T4, T8 and T11, being partially the same sites. Whether by replacing Si vacancies or substituting the fully coordinated Si sites, the most preferential site for Ti is T10, which indicates that the insertion mechanism does not affect the favorable sites of Ti in the MFI lattice. For the defective [Ti(OSi)(3)OH] sites, it was found that the Si vacancy at T6 with a Ti at its neighboring T9 site (T6-def-T9-Ti pair) is the most energetically favorable one, followed by a T6-def-T5-Ti pair with a small energy gap. These findings are significant to elucidate the nature of the active sites and the mechanism of reactions catalyzed by TS-1 and to design the TS-1 catalyst.

Predicting the activity of drugs for a group of imidazopyridine anticoccidial compounds.

Gene expression programming (GEP) is a novel machine learning technique. The GEP is used to build nonlinear quantitative structure-activity relationship model for the prediction of the IC(50) for the imidazopyridine anticoccidial compounds. This model is based on descriptors which are calculated from the molecular structure. Four descriptors are selected from the descriptors' pool by heuristic method (HM) to build multivariable linear model. The GEP method produced a nonlinear quantitative model with a correlation coefficient and a mean error of 0.96 and 0.24 for the training set, 0.91 and 0.52 for the test set, respectively. It is shown that the GEP predicted results are in good agreement with experimental ones.

Quantitative structure activity relationship model for predicting the depletion percentage of skin allergic chemical substances of glutathione.

A quantitative model was developed to predict the depletion percentage of glutathione (DPG) compounds by gene expression programming (GEP). Each kind of compound was represented by several calculated structural descriptors involving constitutional, topological, geometrical, electrostatic and quantum-chemical features of compounds. The GEP method produced a nonlinear and five-descriptor quantitative model with a mean error and a correlation coefficient of 10.52 and 0.94 for the training set, 22.80 and 0.85 for the test set, respectively. It is shown that the GEP predicted results are in good agreement with experimental ones, better than those of the heuristic method.