Since the complexity and structural diversity of man-made compounds are considered, quantitative structure-activity relationships (QSARs)-based fast screening approaches are urgently needed for the assessment of the p...Since the complexity and structural diversity of man-made compounds are considered, quantitative structure-activity relationships (QSARs)-based fast screening approaches are urgently needed for the assessment of the potential risk of endocrine disrupting chemicals (EDCs). The artificial neural net-works (ANN) are capable of recognizing highly nonlinear relationships, so it will have a bright applica-tion prospect in building high-quality QSAR models. As a popular supervised training algorithm in ANN, back-propagation (BP) converges slowly and immerses in vibration frequently. In this paper, a research strategy that BP neural network was improved by conjugate gradient (CG) algorithm with a variable selection method based on genetic algorithm was applied to investigate the QSAR of EDCs. This re-sulted in a robust and highly predictive ANN model with R2 of 0.845 for the training set, q2pred of 0.81 and root-mean-square error (RMSE) of 0.688 for the test set. The result shows that our method can provide a feasible and practical tool for the rapid screening of the estrogen activity of organic compounds.展开更多
The molecular electronegativity-distance vector (MEDV) is employed to describe the chemical structure of bisphenol A analogs and their correlated estrogen activities. The result shows that the constructed models have ...The molecular electronegativity-distance vector (MEDV) is employed to describe the chemical structure of bisphenol A analogs and their correlated estrogen activities. The result shows that the constructed models have good predictability and indicates substructures that may influence estrogen activities of chemicals.展开更多
Computer-based quantitative structure-activity relationship (QSAR) model has been becoming a pow- erful tool in understanding the structural requirements for chemicals to bind the estrogen receptor (ER), designing dru...Computer-based quantitative structure-activity relationship (QSAR) model has been becoming a pow- erful tool in understanding the structural requirements for chemicals to bind the estrogen receptor (ER), designing drugs for human estrogen replacement therapy, and identifying potential estrogenic endo- crine disruptors. In this study, a simple yet powerful neural network technique, generalized regression neural network (GRNN) was used to develop a QSAR model based on 131 structurally diverse estro- gens (training set). Only nine descriptors calculated solely from the molecular structures of com- pounds selected by objective and subjective feature selections were used as inputs of the GRNN model. The predictive power of the built model was found to be comparable to that of the more traditional techniques but requiring significantly easy implementation and a shorter computation-time. The ob- tained result indicates that the proposed GRNN model is robust and satisfactory, and can provide a feasible and practical tool for the rapid screening of the estrogenic activity of organic compounds.展开更多
基金the National Natural Science Foundation of China (Grant No. 20507008)the National Natural Science Foundation Key Project of China (Grant No. 20737001)+1 种基金the Natural Science Foundation of Jiangsu Province,China (Grant No. BK200418)the National Basic Research Program of China (973 Program) (Grant No. 2003CB415002)
文摘Since the complexity and structural diversity of man-made compounds are considered, quantitative structure-activity relationships (QSARs)-based fast screening approaches are urgently needed for the assessment of the potential risk of endocrine disrupting chemicals (EDCs). The artificial neural net-works (ANN) are capable of recognizing highly nonlinear relationships, so it will have a bright applica-tion prospect in building high-quality QSAR models. As a popular supervised training algorithm in ANN, back-propagation (BP) converges slowly and immerses in vibration frequently. In this paper, a research strategy that BP neural network was improved by conjugate gradient (CG) algorithm with a variable selection method based on genetic algorithm was applied to investigate the QSAR of EDCs. This re-sulted in a robust and highly predictive ANN model with R2 of 0.845 for the training set, q2pred of 0.81 and root-mean-square error (RMSE) of 0.688 for the test set. The result shows that our method can provide a feasible and practical tool for the rapid screening of the estrogen activity of organic compounds.
基金This work was supported by the National Natural Science Foundation of China （Grant No. 20477018） the ＂973＂ Program （Grant No. 2003CB415001）the ＂863＂ Program （Grant No. 2001 AA640601-4）.
文摘The molecular electronegativity-distance vector (MEDV) is employed to describe the chemical structure of bisphenol A analogs and their correlated estrogen activities. The result shows that the constructed models have good predictability and indicates substructures that may influence estrogen activities of chemicals.
基金Supported by the National Natural Science Foundation of China (Grant Nos. 20507008 and 20737001)the Natural Science Foundation of Jiangsu Province, China (Grant No. BK200418)the National Basic Research Program of China (973 Program)(Grant No. 2003CB415002)
文摘Computer-based quantitative structure-activity relationship (QSAR) model has been becoming a pow- erful tool in understanding the structural requirements for chemicals to bind the estrogen receptor (ER), designing drugs for human estrogen replacement therapy, and identifying potential estrogenic endo- crine disruptors. In this study, a simple yet powerful neural network technique, generalized regression neural network (GRNN) was used to develop a QSAR model based on 131 structurally diverse estro- gens (training set). Only nine descriptors calculated solely from the molecular structures of com- pounds selected by objective and subjective feature selections were used as inputs of the GRNN model. The predictive power of the built model was found to be comparable to that of the more traditional techniques but requiring significantly easy implementation and a shorter computation-time. The ob- tained result indicates that the proposed GRNN model is robust and satisfactory, and can provide a feasible and practical tool for the rapid screening of the estrogenic activity of organic compounds.
