| Synthetic organic chemicals are prevalent throughout the developed world. Understanding the disposition of these chemicals upon their release to the environment requires data or estimations of various chemical properties that drive environmental partitioning. Two properties of particular interest include the soil/water partition coefficient normalized to organic carbon (K oc) and biodegradability.; Many correlations are available for Koc that have been developed using the n-octanol/water partition coefficient (Kow) as the primary predictor. This work indicates that Kow may not be a strong predictor of Koc for persistent organic pollutants (POPs), which are defined in this work as chemicals with log Kow greater than 5.0. An alternative QSAR model was developed which provides more reliable estimates for the K oc for POPs. This model is based on a set of calculated molecular connectivity indices and critically evaluated Koc data for 18 POPs. The chemical's size and shape, quantified by the molecular connectivity indices, 1chi, 3chic, 4chi vc, are suggested, based on the statistical analysis, to have a dominant effect on the soil sorption process of POPs.; Two biodegradability models developed using an inductive machine learning artificial intelligence based methodology were evaluated as a means to assess the ability of the inductive machine learning artificial intelligence technique to account for environmental parameters, such as interface effects in biodegradability modeling. Validation results for these models, using an independent and quality reviewed database, demonstrate that the models perform well when compared to another commonly used biodegradability model. The ability of models induced by an inductive machine learning artificial intelligence methodology to accommodate complex interactions in detailed systems, and the demonstrated reliability of the approach evaluated by this study, indicate that the methodology may have application in broadening the scope of biodegradability models. Given adequate data for biodegradability of chemicals under environmental conditions, this may allow for the development of future models that include such things as surface interface impacts on biodegradability for example. |
