ChemBCPP: A freely available web server for calculating commonly used physicochemical properties

• Published in: Chemometrics and intelligent laboratory systems Vol. 171; pp. 65 - 73
• Main Authors: Dong, Jie, Wang, Ning-Ning, Liu, Ke-Yi, Zhu, Min-Feng, Yun, Yong-Huan, Zeng, Wen-Bin, Chen, Alex F., Cao, Dong-Sheng
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.12.2017
• Subjects: Cheminformatics; Machine learning; Physiochemical property prediction; QSPR; Web server; Physiochemical property prediction; QSPR; Cheminformatics; Web server; Machine learning
• ISSN: 0169-7439; 1873-3239
• DOI: 10.1016/j.chemolab.2017.10.006

The behavior of a chemical in human or environment mostly depends on its several key physicochemical properties, such as aqueous solubility, octanol-water partition coefficient (logP), boiling point (BP), density, flash point (FP), viscosity, surface tension (ST), vapor pressure (VP) and melting point (MP). Commonly, these properties are important for the environmental sciences and drug discovery, such as the absorption, distribution, metabolism, excretion, and toxicity (ADMET) for medicinal compounds and the common risk assessment for problematic chemicals. At present, the quantitative structure-property relationship (QSPR) model was widely applied to save time and money investment in the early stage of chemical research. Although some satisfactory models were already obtained, most of them are not available for the public researchers and thus cannot be directly applied to practical research projects. Herein, in this study, we developed a user-friendly web server named ChemBCPP that can be used to predict aforementioned 8 important physicochemical properties and calculate several other commonly used properties just by uploading a molecular structure or file. In addition, for a new chemical entity, users can not only get its predicted value but also obtain a leverage value (h value) which can be used to evaluate the reliability of predictive result. We believe that ChemBCPP could be widely applied in environmental science, chemical synthesis and drug ADMET fields with the demand for high quality of chemical properties. ChemBCPP could be freely available via http://chembcpp.scbdd.com. [Display omitted] •Obtain 32 well-optimized QSPR models for 8 key properties.•Building models by RF, SVM, Cubist, Boosting methods.•Consensus models were proposed to stabilize the model performance.•Leverage values were calculated to evaluate the reliability of the prediction.•A user-friendly webserver was developed for rapid estimation of 11 key properties.