Assessment of predictivity of volatile organic compounds carcinogenicity and mutagenicity by freeware in silico models.

Autor: Guerra LR; Post-Graduation Program on Science and Biotechnology, Fluminense Federal University, Niteroi, Brazil., de Souza AMT; Faculty of Pharmacy, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil., Côrtes JA; Post-Graduation Program on Science and Biotechnology, Fluminense Federal University, Niteroi, Brazil., Lione VOF; Faculty of Pharmacy, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil., Castro HC; Cell and Molecular Biology Department, Institute of Biology, Fluminense Federal University, Niteroi, Brazil., Alves GG; Cell and Molecular Biology Department, Institute of Biology, Fluminense Federal University, Niteroi, Brazil. Electronic address: gutopepe@yahoo.com.br.
Jazyk: angličtina
Zdroj: Regulatory toxicology and pharmacology : RTP [Regul Toxicol Pharmacol] 2017 Dec; Vol. 91, pp. 1-8. Date of Electronic Publication: 2017 Sep 29.
DOI: 10.1016/j.yrtph.2017.09.030
Abstrakt: The application of in silico methods is increasing on toxicological risk prediction for human and environmental health. This work aimed to evaluate the performance of three in silico freeware models (OSIRIS v.2.0, LAZAR, and Toxtree) on the prediction of carcinogenicity and mutagenicity of thirty-eight volatile organic compounds (VOC) related to chemical risk assessment for occupational exposure. Theoretical data were compared with assessments available in international databases. Confusion matrices and ROC curves were used to evaluate the sensitivity, specificity, and accuracy of each model. All three models (OSIRIS, LAZAR and Toxtree) were able to identify VOC with a potential carcinogenicity or mutagenicity risk for humans, however presenting differences concerning the specificity, sensitivity, and accuracy. The best predictive performances were found for OSIRIS and LAZAR for carcinogenicity and OSIRIS for mutagenicity, as these softwares presented a combination of negative predictive power and lower risk of false positives (high specificity) for those endpoints. The heterogeneity of results found with different softwares reinforce the importance of using a combination of in silico models to occupational toxicological risk assessment.
(Copyright © 2017 Elsevier Inc. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: