|
(346) The identification and quantification of biologically relevant effect patterns in a simple mixture data set. Jonker, Martijs1, Bedaux, Jacques2, Kammenga, Jan1, 1 2 ABSTRACT- The expected toxicity of a mixture is usually deduced from the toxicity of the individual compounds. Hence, detailed mechanistic information is required about how the individual compounds interact. Due to the difficulty of obtaining this information, predictions are traditionally based on either similar (additive) or independent action. For analyzing mixture toxicity data, it is inevitable to adopt one of these axioms. However, formulating scientifically sound selection criteria is unfeasible. Because of the complexity of the studied systems and the lack of mechanistic insight, both similar and independent action are reference situations relative to which the data can be analyzed. A complete description of the mixture properties is obtained when the toxicity is compared to both references. This comparison enables the characterization of the actual mixture effect, which is biologically most interesting if a certain pattern is observed. Therefore, tools are developed to identify and quantify the most relevant effect patterns: 1) no deviation, 2) absolute deviation (synergism/antagonism), 3) toxicant ratio dependent deviation, and 4) effect level dependent deviation from one of the chosen references. The data analysis is based on non-linear regression routines and is statistically robust. It can be applied generally to all kinds of mixtures, without extrapolation problems. The models can deal with the, sometimes large, differences in toxicity and slope between the individual compounds. Moreover, the parameters have a biological meaning. After implementation in an approach for data analysis in soil systems, the models are used to analyze data sets in practice. It is experienced that, although toxicity data from soil systems are far from perfect, the characteristics of the mixture effect can be identified and quantified adequately. Key words: mixture toxicity, data analysis, soil systems |
