You are here:
CERAPP: Collaborative Estrogen Receptor Activity Prediction Project
Mansouri, K., A. Abdelaziz, A. Rybacka, A. Roncaglioni, A. Tropsha, A. Varnek, A. Zakharov, A. Worth, A. Richard, Chris Grulke, D. Trisciuzzi, D. Fourches, D. Horvath, E. Benfenati, E. Muratov, E. Wedebye, F. Grisoni, G. Mangiatordi, G. Incisivo, H. Hong, H. Ng, I. Tetko, I. Balabin, J. Kancherla, J. Shen, J. Burton, M. Nicklaus, M. Cassotti, N. Nikolov, O. Nicolotti, P. Andersson, Q. Zang, R. Politi, R. Beger, R. Todeschini, R. Huang, S. Farag, S. Rosenberg, S. Slavov, X. Hu, AND R. Judson. CERAPP: Collaborative Estrogen Receptor Activity Prediction Project. ENVIRONMENTAL HEALTH PERSPECTIVES. National Institute of Environmental Health Sciences (NIEHS), Research Triangle Park, NC, 124(7):1023-1033, (2016).
Here, we describe a large-scale modeling project called CERAPP (Collaborative Estrogen Receptor Activity Prediction Project) demonstrating the efficacy of using predictive computational models on high-throughput screening data to screen thousands of chemicals against the ER.
Humans potentially 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. Many of these chemicals never have been tested for their ability to interact with the estrogen receptor (ER). Risk assessors need tools to prioritize chemicals for assessment in costly in vivo tests, for instance, within the EPA Endocrine Disruptor Screening Program. Here, we describe a large-scale modeling project called CERAPP (Collaborative Estrogen Receptor Activity Prediction Project) demonstrating the efficacy of using predictive computational models on high-throughput screening data to screen thousands of chemicals against the ER. CERAPP combined multiple models developed in collaboration among 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 1677 compounds provided by EPA, to build a total of 40 categorical and 8 continuous models for binding, agonist, and antagonist ER activity. All predictions were tested using an evaluation set of 7522 chemicals collected from the literature. To overcome the limitations of single models, a consensus was built weighting models using a scoring function (0 to 1) based on their 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 4001 (12.3%) as ER actives and 6742 (20.8%) as potential actives to be considered as high priority for further testing.