Open-source QSAR models for pKa prediction using multiple machine learning approaches

Springer Science and Business Media LLC - Tập 11 - Trang 1-20 - 2019
Kamel Mansouri1, Neal F. Cariello1, Alexandru Korotcov2, Valery Tkachenko2, Chris M. Grulke3, Catherine S. Sprankle1, David Allen1, Warren M. Casey4, Nicole C. Kleinstreuer4, Antony J. Williams3
1Integrated Laboratory Systems, Inc., Research Triangle Park, USA
2Science Data Software LLC, Rockville, USA
3National Center for Computational Toxicology, U.S. Environmental Protection Agency, Research Triangle Park, USA
4National Institute of Environmental Health Sciences, Research Triangle Park, USA

Tóm tắt

The logarithmic acid dissociation constant pKa reflects the ionization of a chemical, which affects lipophilicity, solubility, protein binding, and ability to pass through the plasma membrane. Thus, pKa affects chemical absorption, distribution, metabolism, excretion, and toxicity properties. Multiple proprietary software packages exist for the prediction of pKa, but to the best of our knowledge no free and open-source programs exist for this purpose. Using a freely available data set and three machine learning approaches, we developed open-source models for pKa prediction. The experimental strongest acidic and strongest basic pKa values in water for 7912 chemicals were obtained from DataWarrior, a freely available software package. Chemical structures were curated and standardized for quantitative structure–activity relationship (QSAR) modeling using KNIME, and a subset comprising 79% of the initial set was used for modeling. To evaluate different approaches to modeling, several datasets were constructed based on different processing of chemical structures with acidic and/or basic pKas. Continuous molecular descriptors, binary fingerprints, and fragment counts were generated using PaDEL, and pKa prediction models were created using three machine learning methods, (1) support vector machines (SVM) combined with k-nearest neighbors (kNN), (2) extreme gradient boosting (XGB) and (3) deep neural networks (DNN). The three methods delivered comparable performances on the training and test sets with a root-mean-squared error (RMSE) around 1.5 and a coefficient of determination (R2) around 0.80. Two commercial pKa predictors from ACD/Labs and ChemAxon were used to benchmark the three best models developed in this work, and performance of our models compared favorably to the commercial products. This work provides multiple QSAR models to predict the strongest acidic and strongest basic pKas of chemicals, built using publicly available data, and provided as free and open-source software on GitHub.

Tài liệu tham khảo

Wikipedia (2019) Acid dissociation constant. https://en.wikipedia.org/w/index.php?title=Acid_dissociation_constant&oldid=897688731. Accessed 21 May 2019 US EPA-OCSPP (2015) Guidance for reporting on the environmental fate and transport of the stressors of concern in problem formulations. In: US EPA. https://www.epa.gov/pesticide-science-and-assessing-pesticide-risks/guidance-reporting-environmental-fate-and-transport. Accessed 21 May 2019 Klöpffer W, Rippen G, Frische R (1982) Physicochemical properties as useful tools for predicting the environmental fate of organic chemicals. Ecotoxicol Environ Saf 6:294–301. https://doi.org/10.1016/0147-6513(82)90019-7 Linde CD (1994) Physico-chemical properties and environmental fate of pesticides. In: Environmental hazards assessment program, state of California EPA. http://agris.fao.org/agris-search/search.do?recordID=US201300074742. Accessed 21 May 2019 National Research Council (2014) A framework to guide selection of chemical alternatives. The National Academies Press, Washington, D.C. https://doi.org/10.17226/18872 Cruciani G, Milletti F, Storchi L et al (2009) In silico pKa prediction and ADME profiling. Chem Biodivers 6:1812–1821. https://doi.org/10.1002/cbdv.200900153 Kerns EH, Di L (2004) Physicochemical profiling: overview of the screens. Drug Discov Today Technol 1:343–348. https://doi.org/10.1016/j.ddtec.2004.08.011 Wetmore BA, Wambaugh JF, Allen B et al (2015) Incorporating high-throughput exposure predictions with dosimetry-adjusted in vitro bioactivity to inform chemical toxicity testing. Toxicol Sci 148:121–136. https://doi.org/10.1093/toxsci/kfv171 Strope CL, Mansouri K, Clewell HJ et al (2018) High-throughput in silico prediction of ionization equilibria for pharmacokinetic modeling. Sci Total Environ 615:150–160. https://doi.org/10.1016/j.scitotenv.2017.09.033 Tetko IV, Bruneau P (2004) Application of ALOGPS to predict 1-octanol/water distribution coefficients, logP, and logD, of AstraZeneca in-house database. J Pharm Sci 93:3103–3110. https://doi.org/10.1002/jps.20217 Hansch C, Maloney PP, Fujita T, Muir RM (1962) Correlation of biological activity of phenoxyacetic acids with Hammett substituent constants and partition coefficients. Nature 194:178–180. https://doi.org/10.1038/194178b0 Fourches D, Muratov E, Tropsha A (2016) Trust, but verify II: a practical guide to chemogenomics data curation. J Chem Inf Model 56:1243–1252. https://doi.org/10.1021/acs.jcim.6b00129 Mansouri K, Grulke CM, Richard AM et al (2016) An automated curation procedure for addressing chemical errors and inconsistencies in public datasets used in QSAR modelling. SAR QSAR Environ Res 27:911–937. https://doi.org/10.1080/1062936X.2016.1253611 BioByte Corporation (2019) BioByte. http://www.biobyte.com/. Accessed 21 May 2019 Advanced Chemistry Development ACDLabs (2019) Chemistry software for analytical and chemical knowledge management. https://www.acdlabs.com/. Accessed 21 May 2019 Simulations Plus (2019) Simulations Plus: model-based drug development to make better data-driven decisions. https://www.simulations-plus.com/. Accessed 21 May 2019 ChemAxon Ltd. (2019) Chemicalize. https://chemaxon.com/products/chemicalize. Accessed 21 May 2019 Sushko I, Novotarskyi S, Körner R et al (2011) Online chemical modeling environment (OCHEM): web platform for data storage, model development and publishing of chemical information. J Comput Aided Mol Des 25:533–554. https://doi.org/10.1007/s10822-011-9440-2 Online Chemical Modeling Environment (OCHEM) (2019) Online chemical database with modeling environment. https://ochem.eu/home/show.do. Accessed 21 May 2019 QSAR DataBank (2019) Institute of Chemistry, University of Tartu, Tartu, Estonia. https://qsardb.org/. Accessed 21 May 2019 Chembench (2019) Carolina Exploratory Center for Cheminformatics Research, Chapel Hill, NC. https://chembench.mml.unc.edu/. Accessed 21 May 2019 Obama B (2013) Making open and machine readable the new default for government information. Office of the Executive, Washington, D.C Burwell SM, VanRoekel S, Mancini DJ (2013) Memorandum for the heads of executive departments and agencies—project open data. https://project-open-data.cio.gov/policy-memo/. Accessed 21 May 2019 Manallack DT (2007) The pK(a) distribution of drugs: application to drug discovery. Perspect Med Chem 1:25–38 Lee AC, Crippen GM (2009) Predicting pKa. J Chem Inf Model 49:2013–2033 Rupp M, Körner R, Tetko IV (2011) Predicting the pK a of small molecules. Comb Chem High Throughput Screen 14:307–327. https://doi.org/10.2174/138620711795508403 Mansouri K, Grulke CM, Judson RS, Williams AJ (2018) OPERA models for predicting physicochemical properties and environmental fate endpoints. J Cheminform. https://doi.org/10.1186/s13321-018-0263-1 Liao C, Nicklaus MC (2009) Comparison of nine programs predicting pKa values of pharmaceutical substances. J Chem Inf Model 49:2801–2812. https://doi.org/10.1021/ci900289x Sander T, Freyss J, von Korff M, Rufener C (2015) DataWarrior: an open-source program for chemistry aware data visualization and analysis. J Chem Inf Model 55:460–473. https://doi.org/10.1021/ci500588j Mansouri K, Abdelaziz A, Rybacka A et al (2016) CERAPP: collaborative estrogen receptor activity prediction project. Environ Health Perspect 124:1023–1033. https://doi.org/10.1289/ehp.1510267 Yap CW (2011) PaDEL-descriptor: an open source software to calculate molecular descriptors and fingerprints. J Comput Chem 32:1466–1474. https://doi.org/10.1002/jcc.21707 Sander T (2019) Openmolecules.org: free services all around molecules. http://www.openmolecules.org/. Accessed 21 May 2019 Yang C, Tarkhov A, Marusczyk J et al (2015) New publicly available chemical query language, CSRML, to support chemotype representations for application to data mining and modeling. J Chem Inf Model 55:510–528. https://doi.org/10.1021/ci500667v Berthold MR, Cebron N, Dill F et al (2008) KNIME: the Konstanz information miner. In: Preisach C, Burkhardt H, Schmidt-Thieme L, Decker R (eds) Data analysis, machine learning and applications: proceedings of the 31st annual conference of the Gesellschaft für Klassifikation e.V., Albert-Ludwigs-Universität Freiburg, March 7–9, 2007. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 319–326 McEachran AD, Mansouri K, Newton SR et al (2018) A comparison of three liquid chromatography (LC) retention time prediction models. Talanta 182:371–379. https://doi.org/10.1016/j.talanta.2018.01.022 Richard AM, Judson RS, Houck KA et al (2016) ToxCast chemical landscape: paving the road to 21st century toxicology. Chem Res Toxicol 29:1225–1251. https://doi.org/10.1021/acs.chemrestox.6b00135 Boser BE, Guyon IM, Vapnik VN (1992) A training algorithm for optimal margin classifiers. In: Proceedings of the fifth annual workshop on computational learning theory. ACM, New York, pp 144–152 Cortes C, Vapnik V (1995) Support-vector networks. Machine learning. McGraw Hill, New York, pp 273–297 Chang C-C, Lin C-J (2001) LIBSVM 3.1: a library for support vector machines. http://www.csie.ntu.edu.tw/~cjlin/libsvm. National Taiwan University, Department of Computer Science, Taipei 106, Taiwan Chang CC, Lin CJ (2011) LIBSVM: a library for support vector machines. ACM Trans Intell Syst Technol 2(3):1–27. https://doi.org/10.1145/1961189.1961199 Consonni V, Ballabio D, Todeschini R (2010) Evaluation of model predictive ability by external validation techniques. J Chemom 24:194–201. https://doi.org/10.1002/cem.1290 Todeschini R, Ballabio D, Grisoni F (2016) Beware of unreliable Q2! A comparative study of regression metrics for predictivity assessment of QSAR models. J Chem Inf Model 56:1905–1913. https://doi.org/10.1021/acs.jcim.6b00277 Ballabio D, Vasighi M, Consonni V, Kompany-Zareh M (2011) Genetic algorithms for architecture optimisation of counter-propagation artificial neural networks. Chemom Intell Lab Syst 105:56–64 Leardi R, Lupiáñez González A (1998) Genetic algorithms applied to feature selection in PLS regression: how and when to use them. Chemom Intell Lab Syst 41:195–207. https://doi.org/10.1016/S0169-7439(98)00051-3 Mansouri K (2019) OPERA—open structure–activity/property relationship app. National Institute of Environmental Health Science, Research Triangle Park, NC. https://github.com/NIEHS/OPERA Sahigara F, Mansouri K, Ballabio D et al (2012) Comparison of different approaches to define the applicability domain of QSAR models. Molecules 17:4791–4810. https://doi.org/10.3390/molecules17054791 MathWorks (2018) MATLAB 2018a. www.mathworks.com Chen T, Guestrin C (2016) XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining—KDD’16, pp 785–794. https://doi.org/10.1145/2939672.2939785 XGBoost (2019) XGBoost documentation. https://xgboost.readthedocs.io/en/latest/. Accessed 21 May 2019 Nishida K (2017) Introduction to extreme gradient boosting in exploratory. https://blog.exploratory.io/introduction-to-extreme-gradient-boosting-in-exploratory-7bbec554ac7. Accessed 21 May 2019 Sheridan RP, Wang WM, Liaw A et al (2016) Extreme gradient boosting as a method for quantitative structure–activity relationships. J Chem Inf Model 56:2353–2360. https://doi.org/10.1021/acs.jcim.6b00591 Kuhn M (2008) Building predictive models in R using the caret package. J Stat Softw 28:1–26. https://doi.org/10.18637/jss.v028.i05 Chen T, He T, Benesty M et al (2019) xgboost: extreme gradient boosting. https://CRAN.R-project.org/package=xgboost. Accessed 21 May 2019 Cariello N (2018) NIEHS/machine-learning-pipeline development. https://github.com/NIEHS/Machine-Learning-Pipeline. Accessed 21 May 2019 Angermueller C, Pärnamaa T, Parts L, Stegle O (2016) Deep learning for computational biology. Mol Syst Biol 12:878. https://doi.org/10.15252/msb.20156651 Jones W, Alasoo K, Fishman D, Parts L (2017) Computational biology: deep learning. Emerg Top Life Sci 1:257–274. https://doi.org/10.1042/ETLS20160025 Mamoshina P, Vieira A, Putin E, Zhavoronkov A (2016) Applications of deep learning in biomedicine. Mol Pharm 13:1445–1454. https://doi.org/10.1021/acs.molpharmaceut.5b00982 Goh GB, Hodas NO, Vishnu A (2017) Deep learning for computational chemistry. J Comput Chem 38:1291–1307. https://doi.org/10.1002/jcc.24764 Hughes TB, Swamidass SJ (2017) Deep learning to predict the formation of quinone species in drug metabolism. Chem Res Toxicol 30:642–656. https://doi.org/10.1021/acs.chemrestox.6b00385 Ma J, Sheridan RP, Liaw A et al (2015) Deep neural nets as a method for quantitative structure–activity relationships. J Chem Inf Model 55:263–274. https://doi.org/10.1021/ci500747n Chollet F Keras: the Python deep learning library. https://keras.io/. Accessed 21 May 2019 Google, Inc (2019) TensorFlow. https://www.tensorflow.org/. Accessed 21 May 2019 Sci-kit Learn Developers (2019) scikit-learn: machine learning in Python. https://scikit-learn.org/stable/. Accessed 21 May 2019 Voosen P, 2017, Pm 2:00 (2017) How AI detectives are cracking open the black box of deep learning. https://doi.org/10.1126/science.aan7059. Accessed 21 May 2019 Castelvecchi D (2016) Can we open the black box of AI? Nat News 538:20. https://doi.org/10.1038/538020a US EPA-NCCT (2019) EPA | TSCA: TSCA inventory, active non-confidential portion. https://comptox.epa.gov/dashboard/chemical_lists/tscaactivenonconf. Accessed 21 May 2019 US EPA-NCCT (2019) Chemistry Dashboard | Batch Search. https://comptox.epa.gov/dashboard/dsstoxdb/batch_search. Accessed 21 May 2019 Wambaugh JF, Hughes MF, Ring CL et al (2018) Evaluating in vitro–in vivo extrapolation of toxicokinetics. Toxicol Sci 163:152–169. https://doi.org/10.1093/toxsci/kfy020 Liu J, Patlewicz G, Williams AJ et al (2017) Predicting organ toxicity using in vitro bioactivity data and chemical structure. Chem Res Toxicol 30:2046–2059. https://doi.org/10.1021/acs.chemrestox.7b00084 US EPA-NCCT (2019) Chemistry dashboard predictions. https://comptox.epa.gov/dashboard/predictions/index. Accessed 23 Aug 2019