An overview of OSPAR priority compounds and selection of a representative training set.

The aim of our study was to chemically characterize a set of priority substances in order to systematically select a representative set of heterogeneous substances for experimental studies of their persistence, bioaccumulation and toxicity (PBT). By including structurally diverse compounds in a PBT-screening protocol, structure-based knowledge would be achieved as a basis for more detailed studies of specific compounds from certain regions of the chemical domain. Initially compounds on the Draft Preliminary List of Substances of Possible Concern presented by the Oslo & Paris Commission (OSPAR) were subjected to broad chemical characterization using Principal Component Analysis (PCA). Molecular weight and seven physico-chemical descriptors collected from the literature, five calculated properties and 25 constitutional descriptors were used in this characterization. The OSPAR list represents a broad spectrum of substances, and thus provides a good basis for selecting heterogeneous substances. D-optimal design was applied to the first four principal components derived in the PCA. In total, 19 structurally diverse substances were chosen as representatives of the chemical domain including 353 substances.