Toxicological Research

Korean Society of Toxicology & Korea Environmental Mutagen Society (한국독성학회)
권호 표지
DOI QR코드

DOI QR Code

Efficiency of pharmaceutical toxicity prediction in computational toxicology

  • Yoshihiro Uesawa (Department of Medical Molecular Informatics, Meiji Pharmaceutical University)
  • Received : 2023.07.30
  • Accepted : 2023.10.11
  • Published : 2024.01.15
⟨ Previous Next ⟩

Abstract

The adverse effects and toxicity of chemical substances pose substantial challenges in drug discovery and environmental science. Their management, most especially in the early development stage, is crucial in preventing costly failures in clinical trials. Predictive methodologies, such as computational toxicology, offer an effective means of managing risks, particularly for new compounds with insufficient post-marketing surveillance and those lacking information on adverse effects. Computational approaches have become increasingly important in environmental science, in which the sheer number and diversity of chemicals present similar challenges to toxicity control. Traditional animal-based evaluation methods are resource intensive, time consuming, and ethically problematic, making them unsuitable for use in assessing the vast compound range. It is an urgent task for the academic community to minimize the risks associated with drug discovery and environmental exposure. This study focuses on systems used to predict toxicity from chemical structure information and outlines the prediction accuracy and systems developed in Japan.

Keywords

Acknowledgement

The Ministry of Economy, Trade and Industry's AI-SHIPS Project in Japan was accomplished as a result of collaborative research by all project participants, beginning with Professor Kimito Funatsu (Nara Institute of Science and Technology), who is the team leader. This writing was supported by the Technology Innovation Program (20023658, Development of AI-based toxicity prediction/evaluation technology to support toxicity verification for chemical safety management) funded By the Ministry of Trade, Industry & Energy (MOTIE, Korea).

References

  1. Uesawa Y (2018) Adverse effect predictions based on computational toxicology techniques and large-scale databases. Yakugaku Zasshi 138:185-190. https://doi.org/10.1248/yakushi.17-00174-4 
  2. Reisfeld B, Mayeno AN (2012) What is computational toxicology? Methods in molecular biology (CliftonNJ). Springer, pp 3-7. https://doi.org/10.1007/978-1-62703-050-2 
  3. Ansari M, Moraiet M, Ahmad S (2014) Insecticides: impact on the environment and human health. In: Malik A, Grohmann E, Akhtar R (eds) Environmental deterioration and human health. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-7890-0_6 
  4. Dix DJ, Houck KA, Martin MT, Richard AM, Setzer RW, Kavlock RJ (2007) The ToxCast program for prioritizing toxicity testing of environmental chemicals. Toxicol Sci 95:5-12. https://doi.org/10.1093/toxsci/kf103 
  5. Amari S, Wu S (1999) Improving support vector machine classifiers by modifying kernel functions. Neural Netw 12:783-789. https://doi.org/10.1016/S0893-6080(99)00032-5 
  6. Breiman L (2001) Random forests. Mach Learn 45:5-32. https://doi.org/10.1023/A:1010933404324 
  7. LeCun Y, Bengio Y, Hinton GE (2015) Deep learning. Nature 521:436-444. https://doi.org/10.1038/nature14539 
  8. Tox21 Data Challenge 2014. https://tripod.nih.gov/tox21/challenge/index.jsp. Accessed 15 Nov 2023 
  9. Tox21. https://ntp.niehs.nih.gov/whatwestudy/tox21. Accessed 15 Nov 2023 
  10. The PubChem Project. NCBI. https://pubchem.ncbi.nlm.nih.gov/. Accessed 15 Nov 2023 
  11. 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 
  12. Hansch C, Fujita T (1964) ρ-σ-π Analysis. A method for the correlation of biological activity and chemical structure. J Am Chem Soc 86:1616-1626. https://doi.org/10.1021/ja01062a035 
  13. Fujita T, Iwasa J, Hansch C (1964) A new substituent constant, π, derived from partition coefficients. J Am Chem Soc 86:5175-5180. https://doi.org/10.1021/ja01077a028 
  14. Deeb O, Goodarzi M (2012) In silico quantitative structure toxicity relationship of chemical compounds: some case studies. Curr Drug Saf 7:289-297. https://doi.org/10.2174/157488612804096533 
  15. Final results for blind set prediction. http://www.cadaster.eu/sites/cadaster.eu/files/final_results.html. Accessed 15 Nov 2023 
  16. Tox21. https://www.epa.gov/chemical-research/toxicology-testing-21st-century-tox21. Accessed 15 Nov 2023 
  17. Ettlin RA (2012) Toxicologic pathology in the 21st century. Toxicol Pathol 41(5):689-708. https://doi.org/10.1177/0192623312466192 
  18. Jeong J, Kim D, Choi J (2022) Application of ToxCast/Tox21 data for toxicity mechanism-based evaluation and prioritization of environmental chemicals: perspective and limitations. Toxicol In Vitro 84:105451. https://doi.org/10.1016/j.tiv.2022.105451 
  19. Ankley GT, Bennett RS, Erickson RJ, Hof DJ, Hornung MW, Johnson RD, Mount DR, Nichols JW, Russom CL, Schmieder PK, Serrrano JA, Tietge JE, Villeneuve DL (2010) Adverse outcome pathways: a conceptual framework to support ecotoxicology research and risk assessment. Environ Toxicol Chem 29:730-741. https://doi.org/10.1002/etc.34 
  20. Attene-Ramos MS, Miller N, Huang R, Michael S, Itkin M, Kavlock RJ, Austin CP, Shinn P, Simeonov A, Tice RR, Xia M (2013) The Tox21 robotic platform for the assessment of environmental chemicals-from vision to reality. Drug Discov Today 18:716-723. https://doi.org/10.1016/j.drudis.2013.05.015 
  21. NIH Tox21 Data Challenge 2014. https://tripod.nih.gov/tox21/challenge/. Accessed 15 Nov 2023 
  22. Uesawa Y (2015) Current status of the NIH sponsored competition on toxicity prediction using chemical structure. Farumashia 51:952-956 
  23. Kurosaki K, Wu R, Uesawa Y (2020) A toxicity prediction tool for potential agonist/antagonist activities in molecular initiating events based on chemical structures. Int J Mol Sci 21:7853. https://doi.org/10.3390/ijms21217853 
  24. Huang R, Xia M, Nguyen D-T, Zhao T, Sakamuru S, Zhao J, Shahane SA, Rossoshek A, Simeonov A (2016) Tox21Challenge to build predictive models of nuclear receptor and stress response pathways as mediated by exposure to environmental chemicals and drugs. Front Environ Sci 14:1-9. https://doi.org/10.3389/fenvs.2015.00085 
  25. Uesawa Y (2016) Rigorous selection of random forest models for identifying compounds that activate toxicity-related pathways. Front Environ Sci 4:1-6. https://doi.org/10.3389/fenvs.2016.00009 
  26. Asako Y, Uesawa Y (2017) High-performance prediction of human estrogen receptor agonists based on chemical structures. Molecules 22:1-10. https://doi.org/10.3390/molecules22040675 
  27. Mayr A, Klambauer G, Unterthiner T, Hochreiter S (2016) Deep-Tox: toxicity prediction using deep learning. Front Environ Sci 3:80. https://doi.org/10.3389/fenvs.2015.00080 
  28. Uesawa Y (2018) Quantitative structure-activity relationship analysis using deep learning based on a novel molecular image input technique. Bioorg Med Chem Lett 28:3400-3403. https://doi.org/10.1016/j.bmcl.2018.08.032 
  29. Matsuzaka Y, Uesawa Y (2019) Optimization of a deep-learning method based on the classification of images generated by parameterized deep snap a novel molecular-image-input technique for quantitative structure-activity relationship (QSAR) analysis. Front Bioeng Biotechnol 7:65. https://doi.org/10.3389/fbioe.2019.00065 
  30. Matsuzaka Y, Uesawa Y (2019) Prediction model with high-performance constitutive androstane receptor (CAR) using Deep-Snap-deep learning approach from the Tox21 10K compound library. Int J Mol Sci 20:4855. https://doi.org/10.3390/ijms20194855 
  31. Matsuzaka Y, Uesawa Y (2022) A deep learning-based quantitative structure-activity relationship system construct prediction model of agonist and antagonist with high performance. Int J Mol Sci 23:2141. https://doi.org/10.3390/ijms23042141 
  32. Matsuzaka Y, Uesawa Y (2023) Ensemble learning, deep learning-based and molecular descriptor-based quantitative structure-activity relationships. Molecules 28:2410. https://doi.org/10.3390/molecules28052410 
  33. Kurosaki K, Uesawa Y (2022) Development of in silico prediction models for drug-induced liver malignant tumors based on the activity of molecular initiating events: biologically interpretable features. J Toxicol Sci 47:89-98. https://doi.org/10.2131/jts.47.89 
  34. AI-SHIPS HP. http://www-dsc.naist.jp/ai-ships/. Accessed 15 Nov 2023 
  35. Matsumoto M, Kobayashi K, Takahashi M, Hirata-Koizumi M, Ono A, Hirose A (2015) Summary information of human health hazard assessment of existing chemical substances (I). Kokuritsu Iyakuhin Shokuhin Eisei Kenkyusho Hokoku (133):42-47 
  36. Akahori Y, Yamashita K, Ishida K, Saito F, Nakai M (2020) Transcriptomics-driven evaluation on liver toxicity using adverse outcome pathways (AOP). Yakugaku Zasshi 140:491-498. https://doi.org/10.1248/yakushi.19-00190-3 
  37. The Hazard Evaluation Support System Integrated Platform (HESS). https://integbio.jp/dbcatalog/record/nbdc01319. Accessed 15 Nov 2023 
  38. Lin Z, Chou WC (2022) Machine learning and artificial intelligence in toxicological sciences. Toxicol Sci 189:7-19. https://doi.org/10.1093/toxsci/kfac075