Toxicity to bacteria and protozoa

Model Species Duration Units r n s Descriptor range 

The slopes (0.8-1) are similar to those of the respective fish baseline models. However, conclusions from the intercepts of the QSARs about the differences in the test systems sensitivity may be misleading because of the different test procedures - especially with regard to the individual species tested - and the differences in the test durations. Nevertheless, the derivation of conforming models by different investigators with diverse sets of chemicals in the various test systems supports the principal validity of these baseline QSARs for estimating toxicity of non-polar non-reactive chemicals towards bacteria and protozoa. Any compound is expected to be at least as toxic as predicted from these models. Similar relationships have also been derived using other descriptors that are generally collinear with log P for non-polar non-specific toxicants (e.g. connectivity indices), which may be applied if the log Pq cannot be estimated, and may also be used to cross-check the obtained predictions, especially if there is reasonable doubt about the correctness of the respective log P values. 

The apparent disadvantages of the low-complexity test systems have turned into one of their major applications in connection with QSAR models, they can be used to identify outliers, which are more (or less) toxic than calculated from their physico-chemical properties. Comparisons of measured toxicity data with calculated baseline predictions discriminate non-specific toxicants from those revealing specific effects. The latter group may be further categorized by comparing the actual effects (in a series of assays 

Non-polar non-specific toxicants 1 Bacillus TL81 dehydrogenase 30 min mmol/1 log I/EC50 = 0.84 log 2.54 0.92 24 0.36 1.5-5 

---