Comment on 'Assessment of chemical screening outcomes based on different partitioning property estimation methods'

摘要

In their article, Zhang et al. (2010) investigate several software prediction methods for equilibrium partition coefficients between octanol and water (K_(oW)), air and water (K_(aw)), and octanol and air (K_(Oa)) in the context of regulatory categorization methods for persistence, bioaccumulation, toxicity, and long range transport potential of various chemicals. The authors use a database of 529 compounds selected from two lists of substances previously identified by Brown and Wania (2008) and Howard and Muir (2010). Our review of the Supplementary materials in Zhang et al. (2010) shows that 18 compounds are duplicate pair structures (given as the following CAS numbers): 2276906 and 13087531; 117964 and 131497 (diatrizoic acid); 20566352 and 77098078 (1,2-benzenedicarboxylic acid, 3,4,5,6-tetrabromo-, 2-(2-hydroxyethoxy)ethyl 2-hydroxypropyl ester); 88722 and 1321126 (1-methyl-2-nitrobenzene); 58899 and 319846 (1,2,3,4,5,6-hexachlorocyclohexane); 2176627 and 68412408 (pentachloropyridine); 335671 and 72968388 (octanoic acid, pentadecafluoro- [n-PFOA]); 68140181 and 68140192; and 678397 and 68391082 (1-decanol, 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro- [8:2-fiuorotelomer alcohol]). One of each of these pairs should be removed from the database. As we noted in our previous comparative assessment of various K_(ow) prediction methods for compounds on the Canadian Domestic Substances List (DSL) (Rayne and Forest, 2009), and our investigation into the potential for corresponding false positivesegatives that are dependent on the software employed in regulatory screening frameworks, such databases contain substantial numbers of structural errors and duplication, necessitating rigorous screening.