Performance of the ALOGPS 2.1 program for octanol-water partition coefficient prediction with organic chemicals on the Canadian Domestic Substances List

摘要

The KOWWIN and ALOGPS octanol-water partition coefficient (K~ow~) estimation software programs were compared for their capacity to accurately predict log K~ow~ values of 1596 organic compounds on the publicly available Domestic Substances List (DSL) from Environment Canada for which experimental data is available. KOWWIN contained a significantly lower number and magnitude of prediction errors compared to ALOGPS, particularly at experimental log K~ow~ values <2. Substantial predictive differences were observed between the two programs for 9093 compounds not having experimental K~ow~ data on the Canadian DSL. Predictive differences of up to 40 log K~ow~ units were found between KOWWIN and ALOGPS, and in some cases, the discrepancies were sufficiently large that strongly opposing hydrophobicity classifications were obtained.