In real environment, various chemicals are always existing as mixtures. The joint effect of a chemical mixture has different features from the components alone and it is not always predicted by the concentration-response information of pure components in the mixture. At present, there is no general approach and procedure for the estimation on the joint effect of a chemical mixture and neither credible fundamental data to develop novel methodology. In recent years, we developed the microplate toxicity analysis to improve the precision in the determination of toxicity data, created the direct equipartition ray to characterize the concentration composition of various binary mixtures and the uniform design ray to represent the concentration diversity in the real mixtures, built the support vector regression technique to model various shapes of different concentration-response curves such as sigmoid curve and hormetic J-shaped curve, proposed the comparison of the whole concentration-response curve of the mixture observed with that predicted by the additive reference model to analyze the toxicity interaction such as synergism or antagonism taken place in a mixture, and designed a computer program named as the assessment and prediction of toxicity of chemical mixture, simply APTox, by integrating with many techniques such as microplate concentration design, best subset regression, iterative algorithm, and so on.

Key words: chemical mixtures, microplate toxicity analysis, direct equipartition ray, uniform design ray, support vector regression, APTox