Machine Learning for Evaluating the Cytotoxicity of Mixtures of Nano-TiO2 and Heavy Metals: QSAR Model Apply Random Forest Algorithm after Clustering Analysis
Abstract
:1. Introduction
2. Results and Discussion
2.1. Experimental Results
2.2. QSAR Model Calculation Results
2.3. Model Validation Results
2.4. Application domain analysis
2.5. Research Results of the Toxicity Mechanisms
3. Materials and Methods
3.1. Cell Experiments
3.2. Research on the QSAR Model
3.2.1. Selection and Calculation of Descriptors
3.2.2. Classification of Mixture Types
3.2.3. Data Set Division
3.2.4. Algorithm Application
3.2.5. Model Validation
3.2.6. Application Domain of the Model
4. Conclusions
Supplementary Materials
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Sharifi, S.; Behzadi, S.; Laurent, S.; Forrest, M.L.; Stroeve, P.; Mahmoudi, M. Toxicity of Nanomaterials. Chem. Soc. Rev. 2012, 41, 2323–2343. [Google Scholar] [CrossRef] [PubMed]
- Roy, J.; Roy, K. Assessment of Toxicity of Metal Oxide and Hydroxide Nanoparticles Using the QSAR Modeling Approach. Environ. Sci. Nano 2021, 8, 3395–3407. [Google Scholar] [CrossRef]
- Qi, Y.; ** a Self-Consistent AGB Wind Model–II. Non-Classical, Non-Equilibrium Polymer Nucleation in a Chemical Mixture. Mon. Not. R. Astron. Soc. 2019, 489, 4890–4911. [Google Scholar] [CrossRef]
- Przybyla, J.; Kile, M.; Smit, E. Description of Exposure Profiles for Seven Environmental Chemicals in a US Population Using Recursive Partition Mixture Modeling (RPMM). J. Expo. Sci. Environ. Epidemiol. 2019, 29, 61–70. [Google Scholar] [CrossRef]
- Udhayakala, P.; Rajendiran, T.V.; Gunasekaran, S. Quantum Chemical Investigations on Some Quinoxaline Derivatives as Effective Corrosion Inhibitors for Mild Steel. Pharm. Lett. 2012, 4, 1285–1298. [Google Scholar]
- Frisch, M.J.; Trucks, G.W.; Schlegel, H.B.; Scuseria, G.E.; Robb, M.A.; Cheeseman, J.R.; Scalmani, G.; Barone, V.; Petersson, G.A.; Nakatsuji, H.; et al. Gaussian 16, Rev. C.01; Gaussian Inc.: Wallingford, UK, 2016. [Google Scholar]
- Abendroth, J.A.; Blankenship, E.E.; Martin, A.R.; Roeth, F.W. Joint Action Analysis Utilizing Concentration Addition and Independent Action Models. Weed Technol. 2011, 25, 436–446. [Google Scholar] [CrossRef]
- Gafourian, T.; Safari, A.; Adibkia, K.; Parviz, F.; Nokhodchi, A. A Drug Release Study from Hydroxypropylmethylcellulose (HPMC) Matrices Using QSPR Modeling. J. Pharm. Sci. 2007, 96, 3334–3351. [Google Scholar] [CrossRef] [PubMed]
- Menze, B.H.; Kelm, B.M.; Masuch, R.; Himmelreich, U.; Bachert, P.; Petrich, W.; Hamprecht, F.A. A Comparison of Random Forest and Its Gini Importance with Standard Chemometric Methods for the Feature Selection and Classification of Spectral Data. BMC Bioinform. 2009, 10, 213. [Google Scholar] [CrossRef] [PubMed]
- Zakariah, M. Classification of Large Datasets Using Random Forest Algorithm in Various Applications: Survey. Int. J. Eng. Innov. Technol. (IJEIT) 2014, 4, 189–198. [Google Scholar]
- Schröder, W.; Holy, M.; Pesch, R.; Ilyin, I.; Harmens, H.; Gebhardt, H. Monitoring the Bioaccumulation of Metals and Nitrogen as Part of the Long-Term Integrated Environmental Monitoring in Baden-Württemberg. Umweltwiss. Schadst.-Forsch. 2010, 22, 721–735. [Google Scholar] [CrossRef]
- Altman, N.; Krzywinski, M. Ensemble Methods: Bagging and Random Forests. Nat. Methods 2017, 14, 933–934. [Google Scholar] [CrossRef]
- Alexander, D.L.J.; Tropsha, A.; Winkler, D.A. Beware of R2: Simple, Unambiguous Assessment of the Prediction Accuracy of QSAR and QSPR Models. J. Chem Inf. Model. 2015, 55, 1316–1322. [Google Scholar] [CrossRef]
- Mentaschi, L.; Besio, G.; Cassola, F.; Mazzino, A. Problems in RMSE-Based Wave Model Validations. Ocean Model. 2013, 72, 53–58. [Google Scholar] [CrossRef]
- Cedeño, W.; Agrafiotis, D.K. Using Particle Swarms for the Development of QSAR Models Based on K-Nearest Neighbor and Kernel Regression. J. Comput. Aided Mol. Des. 2003, 72, 53–58. [Google Scholar] [CrossRef]
Serial Number | CdCl2 (μmol/L) | ZnCl2 (μmol/L) | CuSO4 (μmol/L) | NiCl2 (μmol/L) | Pb(NO3)2 (μmol/L) | MnCl2 (μmol/L) | SbCl3 (μmol/L) | CoCl2 (μmol/L) |
---|---|---|---|---|---|---|---|---|
1 | 10 | 60 | 30 | 100 | 100 | 100 | 5 | 10 |
2 | 20 | 90 | 60 | 200 | 200 | 200 | 10 | 20 |
3 | 30 | 120 | 90 | 300 | 300 | 300 | 15 | 30 |
4 | 40 | 150 | 120 | 400 | 400 | 400 | 20 | 40 |
5 | 50 | 180 | 150 | 500 | 500 | 500 | 25 | 50 |
6 | 60 | 210 | 180 | 600 | 600 | 600 | 30 | 60 |
7 | 70 | 240 | 210 | 700 | 700 | 700 | 35 | 70 |
8 | 80 | 270 | 240 | 800 | 800 | 800 | 40 | 80 |
9 | 90 | 300 | 270 | 900 | 900 | 900 | 45 | 90 |
Descriptor | AdaBoost | RF | Model A | Model B |
---|---|---|---|---|
Highest orbital energy | ||||
Lowest orbital energy | 0.15 | 0.39 | ||
Ionization potentials | 0.10 | |||
Electron affinity | 0.14 | 0.07 | ||
Absolute electronegativity | 0.16 | 0.25 | ||
Absolute hardness | 0.25 | 0.20 | 0.22 | 0.22 |
Molecular energy | 0.11 | 0.14 | 0.31 | |
Adsorption energy | 0.49 | 0.40 | 0.16 | 0.21 |
Model Parameters | AdaBoost | RF | Model A | Model B | Model C | Model D |
---|---|---|---|---|---|---|
Training set samples | 54 | 54 | 27 | 27 | 27 | 27 |
Test set samples | 18 | 18 | 9 | 9 | 9 | 9 |
N estimators | 8 | 4 | 4 | 9 | 1 | 1 |
Random state | 93 | 35 | 95 | 83 | 19 | 79 |
R2 (train) | 0.86 | 0.95 | 0.97 | 0.97 | 0.88 | 0.90 |
R2 (test) | 0.78 | 0.85 | 0.85 | 0.95 | 0.31 | 0.35 |
RMSE (train) | 0.10 | 0.06 | 0.04 | 0.05 | 0.08 | 0.09 |
RMSE (test) | 0.12 | 0.10 | 0.08 | 0.06 | 0.20 | 0.16 |
0.69 | 0.70 | 0.73 | 0.81 | −0.06 | 0.64 | |
−0.20 | −0.44 | −0.45 | −0.47 | −0.86 | −0.79 | |
−0.25 | −0.45 | −0.49 | −0.50 | −1.01 | −1.03 | |
0.79 | 0.86 | 0.87 | 0.95 | 0.50 | 0.79 | |
0.78 | 0.85 | 0.85 | 0.95 | 0.31 | 0.35 | |
0.37 | 0.57 | 0.61 | 0.85 | −0.51 | 0.36 | |
CCC | 0.87 | 0.92 | 0.93 | 0.97 | 0.43 | 0.62 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Sang, L.; Wang, Y.; Zong, C.; Wang, P.; Zhang, H.; Guo, D.; Yuan, B.; Pan, Y. Machine Learning for Evaluating the Cytotoxicity of Mixtures of Nano-TiO2 and Heavy Metals: QSAR Model Apply Random Forest Algorithm after Clustering Analysis. Molecules 2022, 27, 6125. https://doi.org/10.3390/molecules27186125
Sang L, Wang Y, Zong C, Wang P, Zhang H, Guo D, Yuan B, Pan Y. Machine Learning for Evaluating the Cytotoxicity of Mixtures of Nano-TiO2 and Heavy Metals: QSAR Model Apply Random Forest Algorithm after Clustering Analysis. Molecules. 2022; 27(18):6125. https://doi.org/10.3390/molecules27186125
Chicago/Turabian StyleSang, Leqi, Yunlin Wang, Cheng Zong, Pengfei Wang, Huazhong Zhang, Dan Guo, Beilei Yuan, and Yong Pan. 2022. "Machine Learning for Evaluating the Cytotoxicity of Mixtures of Nano-TiO2 and Heavy Metals: QSAR Model Apply Random Forest Algorithm after Clustering Analysis" Molecules 27, no. 18: 6125. https://doi.org/10.3390/molecules27186125