Systematic search for surrogate chromatographic models of biopartitioning processes

Abstract

A systematic approach for identifing surrogate chromatographic models for biopartitioning processes is described. The method is based on a comparison of the system constant ratios of the solvation parameter model for biopartitioning processes and a database of system constant ratios for reversed-phase liquid chromatographic and micellar electrokinetic chromatographic systems compiled from literature sources. An acceptance filter of ⩽0.2 is applied for each difference in system constant ratio for the compared systems to provide a reasonable probability of success without outputting too many systems with limited predictive properties. Surrogate chromatographic models identified for the non-specific toxicity of neutral organic compounds to the fathead minnow and the soil–water distribution constant are tested by construction of a correlation model for the characteristic property of the biological process and the chromatographic retention factors for a structurally varied group of compounds. Although these models are not the best that could be obtained based on ranking of the differences in system constant ratios the predictive ability of the correlation models is suitable for typical applications and similar to the accepted uncertainty in the measurements of the biological property. Retention factors on the immobilized artificial membrane column (IAM PC DD 2) with 10% (v/v) methanol–water as mobile phase are able to estimate non-specific toxicity to the fathead minnow with a standard error (SE) of 0.22 log units and coefficient of determination (r²) of 0.97 for 31 compounds. Retention factors on a Bakerbond DIOL column with 20% (v/v) acetonitrile–water as mobile phase are able to estimate the soil–water distribution constant with an SE of 0.38 log units and r² = 0.88 for 59 compounds. Other potential surrogate chromatographic models are identified for non-specific toxicity to the guppy, tadpole, Vibrio fischeri and Terahymena pyriformis as well as the plant cuticle matrix–water distribution constant. On the other hand reversed-phase chromatographic systems seem poorly suited for estimating intestinal absorption and the blood–brain distribution constant.