Dataset Information

Development of estrogen receptor beta binding prediction model using large sets of chemicals.

ABSTRACT: We developed an ER? binding prediction model to facilitate identification of chemicals specifically bind ER? or ER? together with our previously developed ER? binding model. Decision Forest was used to train ER? binding prediction model based on a large set of compounds obtained from EADB. Model performance was estimated through 1000 iterations of 5-fold cross validations. Prediction confidence was analyzed using predictions from the cross validations. Informative chemical features for ER? binding were identified through analysis of the frequency data of chemical descriptors used in the models in the 5-fold cross validations. 1000 permutations were conducted to assess the chance correlation. The average accuracy of 5-fold cross validations was 93.14% with a standard deviation of 0.64%. Prediction confidence analysis indicated that the higher the prediction confidence the more accurate the predictions. Permutation testing results revealed that the prediction model is unlikely generated by chance. Eighteen informative descriptors were identified to be important to ER? binding prediction. Application of the prediction model to the data from ToxCast project yielded very high sensitivity of 90-92%. Our results demonstrated ER? binding of chemicals could be accurately predicted using the developed model. Coupling with our previously developed ER? prediction model, this model could be expected to facilitate drug development through identification of chemicals that specifically bind ER? or ER?.

SUBMITTER: Sakkiah S

PROVIDER: S-EPMC5696238 | biostudies-literature | 2017 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Development of estrogen receptor beta binding prediction model using large sets of chemicals.

Sakkiah Sugunadevi S Selvaraj Chandrabose C Gong Ping P Zhang Chaoyang C Tong Weida W Hong Huixiao H

Oncotarget 20171010 54

We developed an ERβ binding prediction model to facilitate identification of chemicals specifically bind ERβ or ERα together with our previously developed ERα binding model. Decision Forest was used to train ERβ binding prediction model based on a large set of compounds obtained from EADB. Model performance was estimated through 1000 iterations of 5-fold cross validations. Prediction confidence was analyzed using predictions from the cross v ...[more]

PMID: 29190972

Similar Datasets

Project description:Many chemicals that disrupt endocrine function have been linked to a variety of adverse biological outcomes. However, screening for endocrine disruption using in vitro or in vivo approaches is costly and time-consuming. Computational methods, e.g., quantitative structure-activity relationship models, have become more reliable due to bigger training sets, increased computing power, and advanced machine learning algorithms, such as multilayered artificial neural networks. Machine learning models can be used to predict compounds for endocrine disrupting capabilities, such as binding to the estrogen receptor (ER), and allow for prioritization and further testing. In this work, an exhaustive comparison of multiple machine learning algorithms, chemical spaces, and evaluation metrics for ER binding was performed on public data sets curated using in-house cheminformatics software (Assay Central). Chemical features utilized in modeling consisted of binary fingerprints (ECFP6, FCFP6, ToxPrint, or MACCS keys) and continuous molecular descriptors from RDKit. Each feature set was subjected to classic machine learning algorithms (Bernoulli Naive Bayes, AdaBoost Decision Tree, Random Forest, Support Vector Machine) and Deep Neural Networks (DNN). Models were evaluated using a variety of metrics: recall, precision, F1-score, accuracy, area under the receiver operating characteristic curve, Cohen's Kappa, and Matthews correlation coefficient. For predicting compounds within the training set, DNN has an accuracy higher than that of other methods; however, in 5-fold cross validation and external test set predictions, DNN and most classic machine learning models perform similarly regardless of the data set or molecular descriptors used. We have also used the rank normalized scores as a performance-criteria for each machine learning method, and Random Forest performed best on the validation set when ranked by metric or by data sets. These results suggest classic machine learning algorithms may be sufficient to develop high quality predictive models of ER activity.

Project description:Estrogen receptor orthologues in molluscs may be targets for endocrine disruptors, although mechanistic evidence is lacking. Molluscs are reported to be highly susceptible to effects caused by very low concentrations of environmental estrogens which, if substantiated, would have a major impact on the risk assessment of many chemicals. The present paper describes the most thorough evaluation to-date of the susceptibility of Marisa cornuarietis ER and ERR gene transcription to modulation by vertebrate estrogens in vivo and in vitro. We investigated the effects of estradiol-17? and 4-tert-Octylphenol exposure on in vivo estrogen receptor (ER) and estrogen-related receptor (ERR) gene transcription in the reproductive and neural tissues of the gastropod snail M. cornuarietis over a 12-week period. There was no significant effect (p>0.05) of treatment on gene transcription levels between exposed and non-exposed snails. Absence of a direct interaction of estradiol-17? and 4-tert-Octylphenol with mollusc ER and ERR protein was also supported by in vitro studies in transfected HEK-293 cells. Additional in vitro studies with a selection of other potential ligands (including methyl-testosterone, 17?-ethinylestradiol, 4-hydroxytamoxifen, diethylstilbestrol, cyproterone acetate and ICI182780) showed no interaction when tested using this assay. In repeated in vitro tests, however, genistein (with mcER-like) and bisphenol-A (with mcERR) increased reporter gene expression at high concentrations only (>10(-6)M for Gen and >10(-5)M for BPA, respectively). Like vertebrate estrogen receptors, the mollusc ER protein bound to the consensus vertebrate estrogen-response element (ERE). Together, these data provide no substantial evidence that mcER-like and mcERR activation and transcript levels in tissues are modulated by the vertebrate estrogen estradiol-17? or 4-tert-Octylphenol in vivo, or that other ligands of vertebrate ERs and ERRs (with the possible exception of genistein and bisphenol A, respectively) would do otherwise.

Project description:Humans are exposed to thousands of man-made chemicals in the environment. Some chemicals mimic natural endocrine hormones and, thus, have the potential to be endocrine disruptors. Most of these chemicals have never been tested for their ability to interact with the estrogen receptor (ER). Risk assessors need tools to prioritize chemicals for evaluation in costly in vivo tests, for instance, within the U.S. EPA Endocrine Disruptor Screening Program.We describe a large-scale modeling project called CERAPP (Collaborative Estrogen Receptor Activity Prediction Project) and demonstrate the efficacy of using predictive computational models trained on high-throughput screening data to evaluate thousands of chemicals for ER-related activity and prioritize them for further testing.CERAPP combined multiple models developed in collaboration with 17 groups in the United States and Europe to predict ER activity of a common set of 32,464 chemical structures. Quantitative structure-activity relationship models and docking approaches were employed, mostly using a common training set of 1,677 chemical structures provided by the U.S. EPA, to build a total of 40 categorical and 8 continuous models for binding, agonist, and antagonist ER activity. All predictions were evaluated on a set of 7,522 chemicals curated from the literature. To overcome the limitations of single models, a consensus was built by weighting models on scores based on their evaluated accuracies.Individual model scores ranged from 0.69 to 0.85, showing high prediction reliabilities. Out of the 32,464 chemicals, the consensus model predicted 4,001 chemicals (12.3%) as high priority actives and 6,742 potential actives (20.8%) to be considered for further testing.This project demonstrated the possibility to screen large libraries of chemicals using a consensus of different in silico approaches. This concept will be applied in future projects related to other end points.Mansouri K, Abdelaziz A, Rybacka A, Roncaglioni A, Tropsha A, Varnek A, Zakharov A, Worth A, Richard AM, Grulke CM, Trisciuzzi D, Fourches D, Horvath D, Benfenati E, Muratov E, Wedebye EB, Grisoni F, Mangiatordi GF, Incisivo GM, Hong H, Ng HW, Tetko IV, Balabin I, Kancherla J, Shen J, Burton J, Nicklaus M, Cassotti M, Nikolov NG, Nicolotti O, Andersson PL, Zang Q, Politi R, Beger RD, Todeschini R, Huang R, Farag S, Rosenberg SA, Slavov S, Hu X, Judson RS. 2016.Collaborative Estrogen Receptor Activity Prediction Project. Environ Health Perspect 124:1023-1033;?http://dx.doi.org/10.1289/ehp.1510267.

Dataset Information

Development of estrogen receptor beta binding prediction model using large sets of chemicals.

Publications

Development of estrogen receptor beta binding prediction model using large sets of chemicals.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets