Data Quality in Predictive Toxicology: Reproducibility of Rodent Carcinogenicity Experiments Eva Gottmann,1,2 Stefan Kramer,3 Bernhard Pfahringer,4 and Christoph Helma1,2,3 1Institute for Cancer Research, and 2Institute for Environmental Hygiene, University Vienna, Vienna, Austria 3Institute for Computer Science, Machine Learning Lab, University Freiburg, Freiburg, Germany
4Austrian Research Institute for Artificial Intelligence, Vienna, Austria Abstract We compared 121 replicate rodent carcinogenicity assays from the two parts (National Cancer Institute/National Toxicology Program and literature) of the Carcinogenic Potency Database (CPDB) to estimate the reliability of these experiments. We estimated a concordance of 57% between the overall rodent carcinogenicity classifications from both sources. This value did not improve substantially when additional biologic information (species, sex, strain, target organs) was considered. These results indicate that rodent carcinogenicity assays are much less reproducible than previously expected, an effect that should be considered in the development of structure-activity relationship models and the risk assessment process. Key words: carcinogenicity, machine learning, predictive toxicology, quality assurance, structure-activity relationships. Environ Health Perspect 109:509-514 (2001) . [Online 9 May 2001] http://ehpnet1.niehs.nih.gov/docs/2001/109p509-514gottmann/ abstract.html Address correspondence to E. Gottmann, Federal Ministry of Education, Science and Culture, Abt. VII/B/5 - PROVISO, Minoritenplatz 5, A-1014 Vienna Austria. Telephone: +43-1-53120-5850. Fax: +43-1-53120-5805. E-mail: eva.gottmann@bmbwk.gv.at We thank M. Kundi for helpful discussions and suggestions for improvements. This project was funded by the Austrian Federal Ministry of Science and Transport (GZ 70.017/2-Pr/4/87) . Partial support was also provided by the "Jubiläumsfond der Österreichischen Nationalbank" under grant 6930. The Austrian Federal Ministry of Science and Transport provides general financial support for the Austrian Research Institute for Artificial Intelligence. Received 18 August 2000 ; accepted 6 December 2000. The full version of this article is available for free in HTML or PDF formats. |