食品中含有成千上萬種天然存在的化學物質，過去的毒性測試策略主要是透過暴露於高劑量化學物質之動物試驗來預測對人體健康的影響，但由於進行動物實驗昂貴且費時，所以仍有大量化學物質無動物毒性實驗成果。而美國國家研究委員會在2007年釋出了一份報告21st Century: A Vision and a Strategy，希望藉由計算毒理學的方式，透過結合計算模擬以及高通量體外毒性測試結果，用以了解未知化學物質在特定的狀況下是否具有毒性。
由於計算毒理學持續發展，目前國際上已建立各種毒理學資料庫，如美國環境保護署利用高通量篩選毒性測試資料庫(ToxCast)，以及彙整以往體內毒性測試結果資料庫(ToxRefDB)等。基於各國法規不一致，在大量新型化學物質產生與國際貿易發達的現狀下，各國的原物料、食品產品等都可能輸入至臺灣而導致食安事件的發生，然而目前臺灣尚未有針對國際食安相關法規之毒理學資料庫相關研究。因此本研究目的彙整目前臺灣法規中所涵蓋的食品相關化學物質清單，比對國際間目前運用於食品之化學物質差異，並分析臺灣法規未納入管理之化學物質於毒理學資料庫之試驗結果，以進一步評估因進口食品可能對臺灣食品安全造成威脅之化學物質。
本研究採用ToxCast及ToxRefDB之毒理學資料庫，並針對臺灣、美國及歐盟食品相關法規中化學物質進行彙整比較，利用R語言統計軟體分別將臺灣化學物質及美國與歐盟(但扣除臺灣法規已規範之物質)，以化學物質結構進行自組織映射網路(Self-Organizing Map, SOM)分群，並篩選具有細胞毒性試驗數量較多及平均活性試驗結果比例較高之網格中的化學物質，做為可能具有潛在危害之化學物質，進一步分析其在體外試驗結果中對哪些目標造成影響及濃度，並比對其是否有相關體內毒性試驗資訊以及所造成不良反應結果。
首先，以臺灣食品相關化學物質結構進行SOM分群，透過比對細胞毒性及平均活性試驗結果篩選出具有潛在危害之化學物質共22個，其全部為農藥類別之化學物質，主要用於各種蔬菜、水果和穀物之殺菌劑、殺蟲劑以及部分做為動物用藥使用，並且在體外毒性試驗結果顯示其在較低的濃度下即對Cytochrome P450產生活性影響，以及在體內毒性試驗結果顯示其對肝臟可能造成不良反應結果。
另外再以國際食品相關化學物質結構進行SOM分群，透過細胞毒性及平均活性試驗結果篩選出具有潛在危害之化學物質共有151個，其中有4個化學物質(4-(2-methylbutan-2-yl) phenol、o-Cresol、Benzene及Triphenyltin hydroxide)同時有進行體外及體內毒性測試，且對子宮、肝臟、皮膚及子房具有癌症相關試驗結果。Triphenyltin hydroxide被分類為農藥類別，僅可用於馬鈴薯、甜菜、山核桃；4-(2-methylbutan-2-yl)phenol及o-Cresol主要用於樹脂和聚合物塗料及模塑製品中的酚醛樹脂；Benzene作為食品添加物用於啤酒花萃取物及食品接觸物質之食品包裝黏著劑。
總和以上，以本研究方法所使用化學物質結構SOM分群，篩選出具有相似結構以及可能具有潛在危害之化學物質，分析其在體外及體內毒性試驗結果中對哪些目標造成不良影響，彙整上述化學物質於國際法規之規範及用途，可做為臺灣後續優先進行關注之化學物質，並於國際貿易上應特別加強稽查。

There are many toxicology databases in the world, such as in vitro toxicology database(ToxCast) and in vivo toxicology database(ToxRefDB). Due to international trade, food products from different countries may be imported into Taiwan and cause food safety incidents. Therefore, the purpose of this study will list food-related chemicals in Taiwan regulations, and compare the differences in chemicals used in foods internationally. Meanwhile, this study also aims to analyze toxicity test data and to predict the potential chemicals that may cause hazards to Taiwan's food safety. Food-relevant chemicals of Taiwan-only and the food-relevant chemicals of US and EU were clustered by Self-Organizing Map individually. Choose cytotoxicity and mean proportion bioactive of chemicals as the potential hazard chemicals. There are 22 pesticides in Taiwan regulations, which used in fungicides and insecticides of different vegetables, and veterinary medicine. They cause an impact on Cytochrome P450 at lower concentrations and show adverse effects on the liver. There are 151 potential hazard chemicals included in International regulations. 4 chemicals show cancer-related adverse effects on the uterus, liver, skin, and ovary. Triphenyltin hydroxide is a pesticide and can be applied for potatoes, beets, and pecans; 4-(2-methylbutan-2- yl) phenol and o-Cresol are used in polymer coatings and phenolic resins in molded products; Benzene is used as a food additive modified hop extract and food packaging adhesives. The potential hazard for these chemicals can be regarded as prior concern by the Taiwan government when international trade is common worldwide.
Key word: ToxCast, ToxRefDB, food-related chemicals, data mining

Al-Eryani L, Wahlang B, Falkner K, Guardiola J, Clair H, Prough R, et al. Identification of environmental chemicals associated with the development of toxicant-associated fatty liver disease in rodents. Toxicologic pathology 43:482-497. 2015.
Ali Hameed A, Karlik B, Salman MS, Eleyan G. Robust adaptive learning approach to self-organizing maps. Knowledge-Based Systems 171:25-36. 2019.
Andersen A. Final report on the safety assessment of sodium p -chloro- m -cresol, p -chloro- m -cresol, chlorothymol, mixed cresols, m -cresol, o -cresol, p -cresol, isopropyl cresols, thymol, o -cymen-5-ol, and carvacrol1. International Journal of Toxicology 25:29-127. 2006.
Andersen ME, Krewski D. Toxicity testing in the 21st century: Bringing the vision to life. Toxicol Sci 107:324-330. 2009.
Anzali S, Mederski W, Osswald M, Dorsch D. 1. Endothelin antagonists: Search for surrogates of methylendioxyphenyl by means of a kohonen neural network. Bioorg Med Chem Lett 8:11-16. 1998.
Arfaeinia L, Dobaradaran S, Nasrzadeh F, Shamsi S, Poureshgh Y, Arfaeinia H. Phthalate acid esters (paes) in highly acidic juice packaged in polyethylene terephthalate (pet) container: Occurrence, migration and estrogenic activity-associated risk assessment. Microchemical Journal 155:104719. 2020.
Arrigo P, Giuliano F, Scalia F, Rapallo A, Damiani G. Identification of a new motif on nucleic-acid sequence data using kohonen self-organizing map. Comput Appl Biosci 7:353-357. 1991.
Aylward LL, Hays SM. Consideration of dosimetry in evaluation of toxcast (tm) data. J Appl Toxicol 31:741-751. 2011.
Bauknecht H, Zell A, Bayer H, Levi P, Wagener M, Sadowski J, et al. Locating biologically active compounds in medium-sized heterogeneous datasets by topological autocorrelation vectors: Dopamine and benzodiazepine agonists. Journal of Chemical Information and Computer Sciences 36:1205-1213. 1996.
Benigni R. Predictive toxicology today: The transition from biological knowledge to practicable models. Expert Opin Drug Metab Toxicol 12:989-992. 2016.
Bernard P, Golbraikh A, Kireev D, Chretien JR, Rozhkova N. Comparison of chemical databases: Analysis of molecular diversity with self organising maps (som). Analusis 26:333-341. 1998.
Borrel A, Huang RL, Sakamuru S, Xia MH, Simeonov A, Mansouri K, et al. High-throughput screening to predict chemical-assay interference. Sci Rep 10:20. 2020.
Bottini AA, Hartung T. Food for thought... On the economics of animal testing. Altex 26:3-16. 2009.
Bromilow RH, Evans AA, Nicholls PH. Factors affecting degradation rates of five triazole fungicides in two soil types: 1. Laboratory incubations. Pesticide Science 55:1129-1134. 1999.
Bucher JR. The national toxicology program rodent bioassay. Annals of the New York Academy of Sciences 982:198-207. 2002.
Bushnell PJ, Kavlock RJ, Crofton KM, Weiss B, Rice DC. Behavioral toxicology in the 21st century: Challenges and opportunities for behavioral scientists summary of a symposium presented at the annual meeting of the neurobehavioral teratology society, june, 2009. Neurotoxicol Teratol 32:313-328. 2010.
Busquet F. The need for strategic development of safety sciences. ALTEX 34:3-21. 2017.
Carrington C, Devleesschauwer B, Gibb HJ, Bolger PM. Global burden of intellectual disability resulting from dietary exposure to lead, 2015. Environmental Research 172:420-429. 2019.
Cereto-Massague A, Ojeda MJ, Valls C, Mulero M, Garcia-Vallve S, Pujadas G. Molecular fingerprint similarity search in virtual screening. Methods 71:58-63. 2015.
Chhabra RS, Bucher JR, Wolfe M, Portier C. Toxicity characterization of environmental chemicals by the us national toxicology program: An overview. International journal of hygiene and environmental health 206:437-445. 2003.
Chi ZR, Wu J, Yan H. Handwritten numeral recognition using self-organizing maps and fuzzy rules. Pattern Recognit 28:59-66. 1995.
Chierici M, Giulini M, Bussola N, Jurman G, Furlanello C. Machine learning models for predicting endocrine disruption potential of environmental chemicals. J Environ Sci Health Pt C-Environ Carcinog Ecotoxicol Rev 36:237-251. 2018.
Crane M. Environmental risk evaluation report: 4-tert-pentylphenol (cas no. 80-46-6):Environment Agency. 2008.
Davies TS, Monro A. Marketed human pharmaceuticals reported to be tumorigenic in rodents. Journal of the American College of Toxicology 14:90-107. 1995.
Dix DJ, Houck KA, Martin MT, Richard AM, Setzer RW, Kavlock RJ. The toxcast program for prioritizing toxicity testing of environmental chemicals. Toxicol Sci 95:5-12. 2007.
Eddleston M, Singh S, Buckley N. Organophosphorus poisoning (acute). BMJ Clin Evid 2007:2102. 2007.
Ennever FK, Lave LB. Implications of the lack of accuracy of the lifetime rodent bioassay for predicting human carcinogenicity. Regulatory toxicology and pharmacology : RTP 38:52-57. 2003.
EPA. 2019a. The figure below provides a visual overview of the process toxcast follows to create an accessible database of high-throughput screening information on thousands of chemicals. Available: https://www.epa.gov/chemical-research/toxcast-owners-manual-guidance-exploring-data.
EPA. 2019b. Toxcast data generation: Toxcast pipeline (tcpl). Available: https://www.epa.gov/chemical-research/toxcast-data-generation-toxcast-pipeline-tcpl.
EPA. 2021. Toxcast data generation: Overview of toxcast assays. Available: https://www.epa.gov/chemical-research/toxcast-data-generation-overview-toxcast-assays.
Filer DL, Kothiya P, Setzer RW, Judson RS, Martin MT. Tcpl: The toxcast pipeline for high- throughput screening data. Bioinformatics 33:618-620. 2017.
Firestone M, Kavlock R, Zenick H, Kramer M. The u.S. Environmental protection agency strategic plan for evaluating the toxicity of chemicals. Journal of toxicology and environmental health Part B, Critical reviews 13:139-162. 2010.
Goldberg AM, Frazier JM. Alternatives to animals in toxicity testing. Scientific American 261:24-30. 1989.
Guha R. Chemical informatics functionality in r. J Stat Softw 18:16. 2007.
Henson S, Caswell J. Food safety regulation: An overview of contemporary issues. Food Policy 24:589-603. 1999.
Houck KA, Kavlock RJ. Understanding mechanisms of toxicity: Insights from drug discovery research. Toxicol Appl Pharmacol 227:163-178. 2008.
Hsu YY, Yang CC. Design of artificial neural networks for short-term load forecasting .1. Self-organizing feature maps for day type identification. Iee Proceedings-C Generation Transmission and Distribution 138:407-413. 1991.
Jagt K, Munn S, Torslov J, Bruijn J. Alternative approaches can reduce the use of test animals under reach. Report EUR 2. 2004.
Janesick AS, Dimastrogiovanni G, Vanek L, Boulos C, Chamorro-Garcia R, Tang WY, et al. On the utility of toxcast (tm) and toxpi as methods for identifying new obesogens. Environmental Health Perspectives 124:1214-1226. 2016.
JECFA. 1979. Evaluations of the joint fao/who expert committee on food additives (jecfa)-benzene. Available: https://apps.who.int/food-additives-contaminants-jecfa-database/chemical.aspx?chemID=170.
Judson R, Richard A, Dix DJ, Houck K, Martin M, Kavlock R, et al. The toxicity data landscape for environmental chemicals. Environ Health Perspect 117:685-695. 2009.
Judson R, Houck K, Martin M, Richard AM, Knudsen TB, Shah I, et al. Editor's highlight: Analysis of the effects of cell stress and cytotoxicity on in vitro assay activity across a diverse chemical and assay space. Toxicol Sci 152:323-339. 2016.
Judson RS, Houck KA, Kavlock RJ, Knudsen TB, Martin MT, Mortensen HM, et al. In vitro screening of environmental chemicals for targeted testing prioritization: The toxcast project. Environmental Health Perspectives 118:485-492. 2010.
Judson RS, Kavlock RJ, Setzer RW, Hubal EAC, Martin MT, Knudsen TB, et al. Estimating toxicity-related biological pathway altering doses for high-throughput chemical risk assessment. Chem Res Toxicol 24:451-462. 2011.
Judson RS, Martin MT, Patlewicz G, Wood CE. Retrospective mining of toxicology data to discover multispecies and chemical class effects: Anemia as a case study. Regulatory Toxicology and Pharmacology 86:74-92. 2017.
Kahle M, Buerge IJ, Hauser A, Müller MD, Poiger T. Azole fungicides: Occurrence and fate in wastewater and surface waters. Environ Sci Technol 42:7193-7200. 2008.
Kang J-H, Kondo F, Katayama Y. Human exposure to bisphenol a. Toxicology 226:79-89. 2006.
Karmaus AL, Filer DL, Martin MT, Houck KA. Evaluation of food-relevant chemicals in the toxcast high-throughput screening program. Food Chem Toxicol 92:188-196. 2016.
Kavlock R, Dix D. Computational toxicology as implemented by the u.S. Epa: Providing high throughput decision support tools for screening and assessing chemical exposure, hazard and risk. Journal of toxicology and environmental health Part B, Critical reviews 13:197-217. 2010.
Kim S, Chen J, Cheng T, Gindulyte A, He J, He S, et al. Pubchem 2019 update: Improved access to chemical data. Nucleic Acids Res 47:D1102-D1109. 2019.
Kleinstreuer NC, Judson RS, Reif DM, Sipes NS, Singh AV, Chandler KJ, et al. Environmental impact on vascular development predicted by high-throughput screening. Environ Health Perspect 119:1596-1603. 2011.
Knudsen TB, Martin MT, Kavlock RJ, Judson RS, Dix DJ, Singh AV. Profiling the activity of environmental chemicals in prenatal developmental toxicity studies using the u.S. Epa's toxrefdb. Reprod Toxicol 28:209-219. 2009.
Kohonen T. Self-organized formation of topologically correct feature maps. Biol Cybern 43:59-69. 1982.
Kohonen T, Makisara K. The self-organizing feature maps. Phys Scr 39:168-172. 1989.
Kokkonen M, Torkkola K. Using self-organizing maps and multilayered feedforward nets to obtain phonemic transcriptions of spoken utterances. Speech Commun 9:541-549. 1990.
Korolev D, Balakin KV, Nikolsky Y, Kirillov E, Ivanenkov YA, Savchuk NP, et al. Modeling of human cytochrome p450-mediated drug metabolism using unsupervised machine learning approach. J Med Chem 46:3631-3643. 2003.
Krewski D, Andersen ME, Mantus E, Zeise L. Toxicity testing in the 21st century: Implications for human health risk assessment. Risk Anal 29:474-479. 2009.
Krewski D, Westphal M, Al-Zoughool M, Croteau MC, Andersen ME. 2011. New directions in toxicity testing. In: Annual review of public health, vol 32, Vol. 32, (Fielding JE, Brownson RC, Green LW, eds). Palo Alto:Annual Reviews, 161-178.
Krewski D, Andersen ME, Tyshenko MG, Krishnan K, Hartung T, Boekelheide K, et al. Toxicity testing in the 21st century: Progress in the past decade and future perspectives. Archives of toxicology 94:1-58. 2020.
Leist M, Hartung T. Reprint: Inflammatory findings on species extrapolations: Humans are definitely no 70-kg mice. Altex 30:227-230. 2013.
Leist M, Hasiwa N, Rovida C, Daneshian M, Basketter D, Kimber I, et al. Consensus report on the future of animal-free systemic toxicity testing. ALTEX-Altern Anim Exp 31:341-356. 2014.
Li X, Gasteiger J, Zupan J. On the topology distortion in self-organizing feature maps. Biol Cybern 70:189-198. 1993.
Li Z-H, Zlabek V, Li P, Grabic R, Velisek J, Machova J, et al. Biochemical and physiological responses in liver and muscle of rainbow trout after long-term exposure to propiconazole. Ecotox Environ Safe 73:1391-1396. 2010.
Lien K-W, Hsieh DPH, Huang H-Y, Wu C-H, Ni S-P, Ling M-P. Food safety risk assessment for estimating dietary intake of sulfites in the taiwanese population. Toxicology Reports 3:544-551. 2016.
Liu J, Mansouri K, Judson RS, Martin MT, Hong HX, Chen MJ, et al. Predicting hepatotoxicity using toxcast in vitro bioactivity and chemical structure. Chem Res Toxicol 28:738-751. 2015.
MacCuish JD, MacCuish NE. Chemoinformatics applications of cluster analysis. Wiley Interdiscip Rev-Comput Mol Sci 4:34-48. 2014.
Magnuson B, Munro I, Abbot P, Baldwin N, Lopez-Garcia R, Ly K, et al. Review of the regulation and safety assessment of food substances in various countries and jurisdictions. Food Addit Contam Part A-Chem 30:1147-1220. 2013.
Martin MT, Judson RS, Reif DM, Kavlock RJ, Dix DJ. Profiling chemicals based on chronic toxicity results from the us epa toxref database. Environmental Health Perspectives 117:392-399. 2009.
Martin MT, Knudsen TB, Reif DM, Houck KA, Judson RS, Kavlock RJ, et al. Predictive model of rat reproductive toxicity from toxcast high throughput screening. Biol Reprod 85:327-339. 2011.
Moon KA, Oberoi S, Barchowsky A, Chen Y, Guallar E, Nachman KE, et al. A dose-response meta-analysis of chronic arsenic exposure and incident cardiovascular disease. International Journal of Epidemiology 46:1924-1939. 2017.
Muchmore SW, Edmunds JJ, Stewart KD, Hajduk PJ. Cheminformatic tools for medicinal chemists. J Med Chem 53:4830-4841. 2010.
NRC. Toxicity testing in the 21st century: A vision and a strategy. 2007.
Richard AM, Judson RS, Houck KA, Grulke CM, Volarath P, Thillainadarajah I, et al. Toxcast chemical landscape: Paving the road to 21st century toxicology. Chem Res Toxicol 29:1225-1251. 2016.
Ritter H, Kohonen T. Self-organizing semantic maps. Biol Cybern 61:241-254. 1989.
Russell WMS, Burch RL. The principles of humane experimental technique:Methuen. 1959.
Rusyn I, Daston GP. Computational toxicology: Realizing the promise of the toxicity testing in the 21st century. Environmental Health Perspectives 118:1047-1050. 2010.
Rykowska I, Szymański A, Wasiak W. Determination of bisphenol-a in drinking water using new spe sorbents with chemically bonded ketoimine groups. Polish Journal of Food and Nutrition Sciences 14:237-241. 2005.
Segler MHS, Kogej T, Tyrchan C, Waller MP. Generating focused molecule libraries for drug discovery with recurrent neural networks. ACS Central Sci 4:120-131. 2018.
Shah I, Houck K, Judson RS, Kavlock RJ, Martin MT, Reif DM, et al. Using nuclear receptor activity to stratify hepatocarcinogens. PLoS One 6:e14584. 2011.
Song N, Zhang J. An index for measuring functional diversity in plant communities based on neural network theory. Journal of Applied Mathematics 2013. 2013.
Suckow MA, Stevens KA, Wilson RP. The laboratory rabbit, guinea pig, hamster, and other rodents. Amsterdam; Boston:Elsevier Academic Press. 2012.
Sun J, Zhang A, Zhang J, Xie X, Liu W. Enantiomeric resolution and growth-retardant activity in rice seedlings of uniconazole. Journal of Agricultural and Food Chemistry 60:160-164. 2012.
Tannenbaum J, Bennett BT. Russell and burch's 3rs then and now: The need for clarity in definition and purpose. J Am Assoc Lab Anim Sci 54:120-132. 2015.
Taxvig C, Hass U, Axelstad M, Dalgaard M, Boberg J, Andeasen HR, et al. Endocrine-disrupting activities in vivo of the fungicides tebuconazole and epoxiconazole. Toxicol Sci 100:464-473. 2007.
Theunissen PT, Beken S, Cappon GD, Chen C, Hoberman AM, van der Laan JW, et al. Toward a comparative retrospective analysis of rat and rabbit developmental toxicity studies for pharmaceutical compounds. Reprod Toxicol 47:27-32. 2014.
Thomas RS, Paules RS, Simeonov A, Fitzpatrick SC, Crofton KM, Casey WM, et al. The us federal tox21 program: A strategic and operational plan for continued leadership. ALTEX 35:163-168. 2018.
Tomlin CD. The pesticide manual: A world compendium:British Crop Production Council. 2009.
Tox21. 2019. Tox21’s federal partners. Available: https://tox21.gov/overview/operational-model/.
Truong L, Ouedraogo G, Pham L, Clouzeau J, Loisel-Joubert S, Blanchet D, et al. Predicting in vivo effect levels for repeat-dose systemic toxicity using chemical, biological, kinetic and study covariates. Archives of Toxicology 92:587-600. 2018.
Vesanto J, Alhoniemi E. Clustering of the self-organizing map. IEEE transactions on neural networks 11:586-600. 2000.
Voicu A, Duteanu N, Voicu M, Vlad D, Dumitrascu V. The rcdk and cluster r packages applied to drug candidate selection. J Cheminformatics 12:8. 2020.
Wang Y, Zhu W, Qiu J, Wang X, Zhang P, Yan J, et al. Monitoring tryptophan metabolism after exposure to hexaconazole and the enantioselective metabolism of hexaconazole in rat hepatocytes in vitro. Journal of Hazardous Materials 295:9-16. 2015.
Watford S, Ly Pham L, Wignall J, Shin R, Martin MT, Friedman KP. Toxrefdb version 2.0: Improved utility for predictive and retrospective toxicology analyses. Reproductive Toxicology 89:145-158. 2019.
Wehrens R, Kruisselbrink J. Flexible self-organizing maps in kohonen 3.0. J Stat Softw 87:1-18. 2018.
Weininger D. Smiles, a chemical language and information system. 1. Introduction to methodology and encoding rules. Journal of Chemical Information and Computer Sciences 28:31-36. 1988.
World Health O, Food FAOPoEoPRi, the E, Residues WHOEGoP, International Programme on Chemical S. 1992. Pesticide residues in food : 1991, toxicology evaluations, joint meeting of the fao panel of experts on pesticide residues in food and the environment and the who expert group on pesticide residues, geneva, 16-25 september 1991. Part 2., toxicology. Geneva:World Health Organization.
Wu, Felicia, Rodricks JV. Forty years of food safety risk assessment: A history and analysis. Risk Analysis 40:2218-2230. 2020.
Wu CH. Artificial neural networks for molecular sequence analysis. Comput Chem 21:237-256. 1997.
Wu JVRDTFCF. 2020. Food constituents and contaminants. In: Environmental toxicants, 149-204.
Wu YY, Wang GY. Machine learning based toxicity prediction: From chemical structural description to transcriptome analysis. Int J Mol Sci 19:20. 2018.
Zang QD, Rotroff DM, Judson RS. Binary classification of a large collection of environmental chemicals from estrogen receptor assays by quantitative structure-activity relationship and machine learning methods. J Chem Inf Model 53:3244-3261. 2013.
許如君、吳昌昱. 農用藥劑分類及作用機制檢索第三版. 國立臺灣大學昆蟲學系:行政院農委會動植物防疫檢疫局. 2018年.
衛生福利部食品藥物管理署. 2016. 食品添加物手冊. Available: https://www.fda.gov.tw/tc/promotionalContent.aspx?pn=15&id=7.