Prediction of ionic liquid toxicity by interpretable machine learning

doi:10.1016/j.cjche.2025.04.018

中国化学工程学报 ›› 2025, Vol. 84 ›› Issue (8): 201-210.DOI: 10.1016/j.cjche.2025.04.018

• Full Length Article • 上一篇下一篇

Prediction of ionic liquid toxicity by interpretable machine learning

Haijun Feng¹, Li Jiajia², Zhou Jian²

1. School of Computer and Software, Shenzhen Institute of Information Technology, Shenzhen 518172, China;
2. School of Chemistry and Chemical Engineering, Guangdong Provincial Key Laboratory for Green Chemical Product Technology, South China University of Technology, Guangzhou 510640, China

收稿日期:2025-01-15 修回日期:2025-03-08 接受日期:2025-04-23 出版日期:2025-08-28 发布日期:2025-06-07
通讯作者: Haijun Feng,E-mail:fenghj@sziit.edu.cn;Zhou Jian,E-mail:jianzhou@scut.edu.cn
基金资助:
This study was funded by Research Platforms and Projects for Higher Education Institutions of Department of Education of Guangdong Province in 2024 (2024KTSCX256), and 2023 Guangdong Province Higher Vocational Education Teaching Quality and Teaching Reform Project (2023JG080).

Prediction of ionic liquid toxicity by interpretable machine learning

Haijun Feng¹, Li Jiajia², Zhou Jian²

1. School of Computer and Software, Shenzhen Institute of Information Technology, Shenzhen 518172, China;
2. School of Chemistry and Chemical Engineering, Guangdong Provincial Key Laboratory for Green Chemical Product Technology, South China University of Technology, Guangzhou 510640, China

Received:2025-01-15 Revised:2025-03-08 Accepted:2025-04-23 Online:2025-08-28 Published:2025-06-07
Contact: Haijun Feng,E-mail:fenghj@sziit.edu.cn;Zhou Jian,E-mail:jianzhou@scut.edu.cn
Supported by:
This study was funded by Research Platforms and Projects for Higher Education Institutions of Department of Education of Guangdong Province in 2024 (2024KTSCX256), and 2023 Guangdong Province Higher Vocational Education Teaching Quality and Teaching Reform Project (2023JG080).

摘要/Abstract

摘要： The potential toxicity of ionic liquids (ILs) affects their applications; how to control the toxicity is one of the key issues in their applications. To understand its toxicity structure relationship and promote its greener application, six different machine learning algorithms, including Bagging, Adaptive Boosting (AdaBoost), Gradient Boosting (GBoost), Stacking, Voting and Categorical Boosting (CatBoost), are established to model the toxicity of ILs on four distinct datasets including Leukemia rat cell line IPC-81 (IPC-81), Acetylcholinesterase (AChE), Escherichia coli (E.coli) and Vibrio fischeri. Molecular descriptors obtained from the simplified molecular input line entry system (SMILES) are used to characterize ILs. All models are assessed by the mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE) and correlation coefficient (R²). Additionally, an interpretation model based on SHapley Additive exPlanations (SHAP) is built to determine the positive and negative effects of each molecular feature on toxicity. With additional parameters and complexity, the Catboost model outperforms the other models, making it a more reliable model for ILs' toxicity prediction. The results of the model's interpretation indicate that the most significant positive features, SMR_VSA5, PEOE_VSA8, Kappa2, PEOE_VSA6, SMR_VSA5, PEOE_VSA6 and EState_VSA1, can increase the toxicity of ILs as their levels rise, while the most significant negative features, VSA_EState7, EState_VSA8, PEOE_VSA9 and FpDensityMorgan1, can decrease the toxicity as their levels rise. Also, an IL's toxicity will grow as its average molecular weight and number of pyridine rings increase, whereas its toxicity will decrease as its hydrogen bond acceptors increase. This finding offers a theoretical foundation for rapid screening and synthesis of environmentally-benign ILs.

关键词: Ionic liquids, Toxicity, Machine learning, Model, Prediction, Interpretation

Abstract: The potential toxicity of ionic liquids (ILs) affects their applications; how to control the toxicity is one of the key issues in their applications. To understand its toxicity structure relationship and promote its greener application, six different machine learning algorithms, including Bagging, Adaptive Boosting (AdaBoost), Gradient Boosting (GBoost), Stacking, Voting and Categorical Boosting (CatBoost), are established to model the toxicity of ILs on four distinct datasets including Leukemia rat cell line IPC-81 (IPC-81), Acetylcholinesterase (AChE), Escherichia coli (E.coli) and Vibrio fischeri. Molecular descriptors obtained from the simplified molecular input line entry system (SMILES) are used to characterize ILs. All models are assessed by the mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE) and correlation coefficient (R²). Additionally, an interpretation model based on SHapley Additive exPlanations (SHAP) is built to determine the positive and negative effects of each molecular feature on toxicity. With additional parameters and complexity, the Catboost model outperforms the other models, making it a more reliable model for ILs' toxicity prediction. The results of the model's interpretation indicate that the most significant positive features, SMR_VSA5, PEOE_VSA8, Kappa2, PEOE_VSA6, SMR_VSA5, PEOE_VSA6 and EState_VSA1, can increase the toxicity of ILs as their levels rise, while the most significant negative features, VSA_EState7, EState_VSA8, PEOE_VSA9 and FpDensityMorgan1, can decrease the toxicity as their levels rise. Also, an IL's toxicity will grow as its average molecular weight and number of pyridine rings increase, whereas its toxicity will decrease as its hydrogen bond acceptors increase. This finding offers a theoretical foundation for rapid screening and synthesis of environmentally-benign ILs.

Key words: Ionic liquids, Toxicity, Machine learning, Model, Prediction, Interpretation

Haijun Feng, Li Jiajia, Zhou Jian. Prediction of ionic liquid toxicity by interpretable machine learning[J]. 中国化学工程学报, 2025, 84(8): 201-210.

Haijun Feng, Li Jiajia, Zhou Jian. Prediction of ionic liquid toxicity by interpretable machine learning[J]. Chinese Journal of Chemical Engineering, 2025, 84(8): 201-210.

参考文献

[1] M.Y. Zhu, Y. Yang, Poly(ionic liquid)s: an emerging platform for green chemistry, Green Chem. 26 (9) (2024) 5022-5102.
[2] L. Zhang, J. Liu, S. Lou, H. Jiang, S. Wang, S. Li, T. Tang, A. Zhang, Synergistic Flame-retardant Epoxy Resin Using Quaternary Phosphate Naphthalene Sulfonate Ionic Liquid and Dimethyl Methylphosphonate, Chem. J. Chin. U., 45 (04) (2024) 17-26. (in Chinese).
[3] F. Li, T. Zhang, L. Lv, W.X. Tang, Y. Wang, S.W. Tang, Effects of ionic liquids on the vapor-liquid equilibriumof 1, 3, 5-trioxane-water system at 101.3 kPa, Chin. J. Chem. Eng. 73 (2024) 42-50.
[4] P.M. Probert, A.C. Leitch, M.P. Dunn, S.K. Meyer, J.M. Palmer, T.M. Abdelghany, A.F. Lakey, M.P. Cooke, H. Talbot, C. Wills, W. McFarlane, L.I. Blake, A.K. Rosenmai, A. Oskarsson, R. Figueiredo, C. Wilson, G.E. Kass, D.E. Jones, P.G. Blain, M.C. Wright, Identification of a xenobiotic as a potential environmental trigger in primary biliary cholangitis, J. Hepatol. 69 (5) (2018) 1123-1135.
[5] A. Jakubowska, E. Grabinska-Sota, Study of the toxicity of five quaternary ammonium ionic liquids to aquatic organisms, Desalin. Water Treat. 117 (2018) 202-210.
[6] G. Jeremias, F. Jesus, S.P.M. Ventura, F.J.M. Goncalves, J. Asselman, J.L. Pereira, New insights on the effects of ionic liquid structural changes at the gene expression level: Molecular mechanisms of toxicity in Daphnia Magna, J. Hazard. Mater. 409 (2021) 124517.
[7] C. Zhang, L.S. Zhu, J.H. Wang, J. Wang, T.T. Zhou, Y.Q. Xu, C. Cheng, The acute toxic effects of imidazolium-based ionic liquids with different alkyl-chain lengths and anions on zebrafish (Danio rerio), Ecotoxicol. Environ. Saf. 140 (2017) 235-240.
[8] D.M. Makarov, Y.A. Fadeeva, E.A. Safonova, L.E. Shmukler, Predictive modeling of antibacterial activity of ionic liquids by machine learning methods, Comput. Biol. Chem. 101 (2022) 107775.
[9] M. Villanueva, P. Vallet, T. Teijeira, A. Santiago-Alonso, A. Amigo, E. Tojo, L.M. Varela, J.J. Parajo, J. Salgado, Effect of alkyl chain length on the thermal properties and toxicity of n-alkyl-ammonium nitrate ionic liquids (n = 2, 3, 4, 5, 6, 8) for energy applications, J. Therm. Anal. Calorim. (2024), https://doi.org/10.1007/s10973-024-13333-y.
[10] H.J. Feng, L.L. Qin, B.X. Zhang, J. Zhou, Prediction and interpretability of melting points of ionic liquids using graph neural networks, ACS Omega 9 (14) (2024) 16016-16025.
[11] V. Venkatraman, S. Evjen, H.K. Knuutila, A. Fiksdahl, B.K. Alsberg, Predicting ionic liquid melting points using machine learning, J. Mol. Liq. 264 (2018) 318-326.
[12] H.J. Feng, W. Qin, G.W. Hu, H.J. Wang, Intelligent prediction of nitrous oxide capture in designable ionic liquids, Appl. Sci. 13 (12) (2023) 6900.
[13] H.J. Feng, P.G. Zhang, W. Qin, W.M. Wang, H.J. Wang, Estimation of solubility of acid gases in ionic liquids using different machine learning methods, J. Mol. Liq. 349 (2022) 118413.
[14] Y. Jian, Y.Y. Wang, A. Barati Farimani, Predicting CO₂ absorption in ionic liquids with molecular descriptors and explainable graph neural networks, ACS Sustainable Chem. Eng. 10 (50) (2022) 16681-16691.
[15] D.M. Makarov, Y.A. Fadeeva, V.A. Golubev, A.M. Kolker, Designing deep eutectic solvents for efficient CO₂ capture: a data-driven screening approach, Sep. Purif. Technol. 325 (2023) 124614.
[16] F.Z. Meng, S.Y. Han, L. Lin, J.L. Li, K.L. Chen, J.G. Jiang, Process optimization and mechanism study of ionic liquid-based mixed amine biphasic solvents for CO₂ capture in biogas upgrading procedure, Front. Environ. Sci. Eng. 18 (8) (2024) 95.
[17] S.J. Zeng, X.Q. Sun, Y.G. Bai, L. Bai, S. Zheng, X.P. Zhang, S.J. Zhang, Research progress of CO₂ capture and separation by functionalized ionic liquids and materials, Acta Chim. Sin. 81 (6) (2023) 627.
[18] Y.X. Fu, X.Y. Liu, J.Z. Gao, Y. Lei, Y.Q. Chen, X.P. Zhang, Machine learning models for the density and heat capacity of ionic liquid-water binary mixtures, Chin. J. Chem. Eng. 73 (2024) 244-255.
[19] Y. Lei, Y. Shu, X.G. Liu, X.Y. Liu, X.Q. Wu, Y.Q. Chen, Predictive modeling on the surface tension and viscosity of ionic liquid-organic solvent mixtures via machine learning, J. Taiwan Inst. Chem. Eng. 151 (2023) 105140.
[20] Y.Q. Chen, S.L. Ma, Y. Lei, X.D. Liang, X.Y. Liu, G.M. Kontogeorgis, R. Gani, Ionic liquid binary mixtures: Machine learning-assisted modeling, solvent tailoring, process design, and optimization, AlChE. J. 70 (5) (2024) e18392.
[21] N.H. Abdullah, D. Zaini, B. Lal, Prediction of ionic liquids toxicity using machine learning models for application to gas hydrate, Process. Saf. Prog. 43 (2024) S199-S212.
[22] D.C. Fan, K. Xue, R.Q. Zhang, W.G. Zhu, H.R. Zhang, J.G. Qi, Z.Y. Zhu, Y.L. Wang, P.Z. Cui, Application of interpretable machine learning models to improve the prediction performance of ionic liquids toxicity, Sci. Total Environ. 908 (2024) 168168.
[23] Y.S. Zhao, J.H. Zhao, Y. Huang, Q. Zhou, X.P. Zhang, S.J. Zhang, Toxicity of ionic liquids: database and prediction via quantitative structure-activity relationship method, J. Hazard. Mater. 278 (2014) 320-329.
[24] A. Sosnowska, M. Grzonkowska, T. Puzyn, Global versus local QSAR models for predicting ionic liquids toxicity against IPC-81 leukemia rat cell line: The predictive ability, J. Mol. Liq. 231 (2017) 333-340.
[25] L.D. Cao, P. Zhu, Y.S. Zhao, J.H. Zhao, Using machine learning and quantum chemistry descriptors to predict the toxicity of ionic liquids, J. Hazard. Mater. 352 (2018) 17-26.
[26] S.Y. Ma, M. Lv, X.Y. Zhang, H.L. Zhai, W.J. Lv, Computational study of the effects of cations and anions to the cytotoxicity of diverse ionic liquids by supervised machine learning, Chemom. Intell. Lab. Syst. 144 (2015) 138-147.
[27] J.C. Yan, G.H. Liu, H.L. Chen, S. Hu, X.H. Wang, B. Yan, X.L. Yan, ILTox: a curated toxicity database for machine learning and design of environmentally friendly ionic liquids, Environ. Sci. Technol. Lett. 10 (11) (2023) 983-988.
[28] J.C. Yan, X.L. Yan, S. Hu, H. Zhu, B. Yan, Comprehensive interrogation on acetylcholinesterase inhibition by ionic liquids using machine learning and molecular modeling, Environ. Sci. Technol. 55 (21) (2021) 14720-14731.
[29] I.U. Ekanayake, D.P.P. Meddage, U. Rathnayake, A novel approach to explain the black-box nature of machine learning in compressive strength predictions of concrete using Shapley additive explanations (SHAP), Case Stud. Constr. Mater. 16 (2022) e01059.
[30] L. Breiman, Bagging predictors, Mach. Learn. 24 (2) (1996) 123-140.
[31] Y. Freund, R.E. Schapire, A decision-theoretic generalization of on-line learning and an application to boosting, J. Comput. Syst. Sci. 55 (1) (1997) 119-139.
[32] J.H. Friedman, Greedy function approximation: a gradient boosting machine, Ann. Statist. 29 (5) (2001) 1189-1232.
[33] D.H. Wolpert, Stacked generalization, Neural Netw. 5 (2) (1992) 241-259.
[34] A.V. Dorogush, V. Ershov, A. Gulin, CatBoost: gradient boosting with categorical features support, (2018): 1810.11363.
[35] J.T. Hancock, T.M. Khoshgoftaar, CatBoost for big data: an interdisciplinary review, J. Big Data 7 (1) (2020) 94.
[36] P. Geurts, D. Ernst, L. Wehenkel, Extremely randomized trees, Mach. Learn. 63 (1) (2006) 3-42.
[37] L. Breiman, Random forests, Mach. Learn. 45 (1) (2001) 5-32.
[38] P. Labute, A widely applicable set of descriptors, J. Mol. Graph. Model. 18 (4-5) (2000) 464-477.
[39] B. Ramsundar, P. Eastman, P. Walters, V. Pande, Deep Learning for the Life Sciences: Applying Deep Learning to Genomics, Microscopy, Drug Discovery, and More, O'Reilly Media Inc., 2019.
[40] Molecular descriptors, https://datagrok.ai/help/datagrok/solutions/domains/chem/descriptors.
[41] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V. "Scikit-learn: Machine learning in Python", the Journal of machine Learning research, 12, 2825-2830(2011).
[42] L. Buitinck, G. Louppe, M. Blondel, F. Pedregosa, A. Mueller, O. Grisel, V. Niculae, P. Prettenhofer, A. Gramfort, J. Grobler, R. Layton, J. Vanderplas, A. Joly, B. Holt, G. Varoquaux, API design for machine learning software: experiences from the scikit-learn project, (2013): 1309.0238.
[43] Fan, R., Chang, K., Hsieh, C., Wang, X., Lin, C. "LIBLINEAR: A library for large linear classification", Mach. Learn. Res. 9 (2008), 1871-1874.
[44] H. Feng, B. Zhang, J. Zhou, Predicting and interpreting the toxicity of ionic liquids using graph neural network", CIESC Journal 76 (2025) 93-106. (in Chinese).
[45] G. E. Hinton, Connectionist learning procedures, Artif. Intell. 40 (1-3) (1989) 185-234.
[46] T.Q. Chen, C. Guestrin, XGBoost: A Scalable Tree Boosting System, In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco California USA, ACM, 2016.
[47] A. Krogh, What are artificial neural networks? Nat. Biotechnol. 26 (2008) 195-197.

Prediction of ionic liquid toxicity by interpretable machine learning

Prediction of ionic liquid toxicity by interpretable machine learning

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	Yujing Zhao, Qilei Liu, Jian Du, Qingwei Meng, Liang Sun, Lei Zhang. High-throughput and intelligent design of potential GRK2 inhibitor candidates using deep learning and mathematical programming methods[J]. 中国化学工程学报, 2025, 84(8): 11-22.
[2]	Ali Tarik Karagoz, Omar Alqusair, Chao Liu, Jie Li. Advances in conceptual process design: From conventional strategies to AI-assisted methods[J]. 中国化学工程学报, 2025, 84(8): 60-76.
[3]	Dian Zhang, Bo Ouyang, Zheng-Hong Luo. Reaction process optimization based on interpretable machine learning and metaheuristic optimization algorithms[J]. 中国化学工程学报, 2025, 84(8): 77-85.
[4]	Zhi Ma, Peng Cui, Xu Wang, Lanyu Li, Haoxiang Xu, Adrian Fisher, Daojian Cheng. The integration of artificial intelligence and high-throughput experiments: An innovative driving force in catalyst design[J]. 中国化学工程学报, 2025, 84(8): 117-132.
[5]	Zheyuan Pang, Siying Liu, Yiting Lin, Xiangchen Fang, Honglai Liu, Chong Peng, Cheng Lian. Large language model-based multi-objective modeling framework for vacuum gas oil hydrotreating[J]. 中国化学工程学报, 2025, 84(8): 133-145.
[6]	Jianfeng Jiao, Xi Gao, Jie Li. Pure component property estimation framework using explainable machine learning methods[J]. 中国化学工程学报, 2025, 84(8): 158-178.
[7]	Raúl Díaz, Hongliang Xin. Knowledge graphs in heterogeneous catalysis: Recent advances and future opportunities[J]. 中国化学工程学报, 2025, 84(8): 179-189.
[8]	Xiaodong Hong, Yudong Shen, Zuwei Liao, Yongrong Yang. Machine-learning-assisted high-throughput computational screening of the n-hexane cracking initiator[J]. 中国化学工程学报, 2025, 84(8): 190-200.
[9]	Feifei Chen, Zhenyuan Xiao, Zhongfan Luo, Peng Jiang, Jingjing Chen, Yuanhui Ji, Jiahua Zhu, Xiaohua Lu, Liwen Mu. Prediction of mass transfer performance in gas-liquid stirred bioreactor using machine learning[J]. 中国化学工程学报, 2025, 84(8): 211-226.
[10]	Yuan Tian, Honghua Zhang, Yueyang Qiao, Han Yang, Yanrong Liu, Xiaoyan Ji. Intelligent prediction of ionic liquids and deep eutectic solvents by machine learning[J]. 中国化学工程学报, 2025, 84(8): 227-243.
[11]	Yuan Wang, Mengyue Chen, Jingwei Tian, Weidong Zhang, Dahuan Liu. A data-driven predictive model for solubility: A case study of the NaCl-Na₂SO₄-H₂O system[J]. 中国化学工程学报, 2025, 84(8): 254-265.
[12]	Yongqi Pan, Yazi Yu, Lijie Wang, Guogang Hu, Yujun Wang, Guangsheng Luo. Intelligent chemical synthesis based on microchemical engineering technology[J]. 中国化学工程学报, 2025, 84(8): 274-288.
[13]	Yulong Zhang, Hongchao Yue, Yanru Zhang, Yun Li, Hongfei Guo, Dong Xu, Jilin Cao. Separation of chlorineebromine-based solid solution in bittern tail solution by phase equilibrium[J]. 中国化学工程学报, 2025, 83(7): 62-71.
[14]	Mika Sillanpää, Mohammad Reza kimiaei, Soheil Balsini Gavanaroudi, Nezamaddin Mengelizadeh, Najmeh Ahmadi, Davoud Balarak. Synthesis of magnetically separable and recyclable MnFe₂O₄@SiO₂@NH₂ nanocomposite coupled-acylated MWCNTS with enhanced photocatalytic performance under visible-light irradiation[J]. 中国化学工程学报, 2025, 83(7): 229-243.
[15]	Yuping Luo, Wenyang Wang, Yuyan Zhang, Muxin Chen, Peng Shao. A transformer-based model for predicting and analyzing light olefin yields in methanol-to-olefins process[J]. 中国化学工程学报, 2025, 83(7): 266-276.