Databases predictive models

The Office of Pollution Prevention and Toxics (OPPT) has developed a series of methods, databases, and predictive models to help in evaluating what happens to chemicals when they are used and released into the environment. These tools are intended to be used by scientists and engineers famihar with exposure assessment principles. 

The evaluation for aquatic toxicity on daphnids and fish is reported in Tables 12 and 13. Bold values indicate that compounds are out of the model applicability domain (ECOSAR) or that the prediction is not reliable. ECOSAR and ToxSuite are able to predict all the selected compounds while T.E.S.T. fails in prediction for the daphnia toxicity of perfluorinated compounds (PFOS and PFOA). Tables 12 and 13 include also a limited number of experimental results provided by the model training dataset (some data are extracted from USEPA Ecotox database). Predicted results are in agreement for five compounds only (2, 3, 5, 13 and 14) for both endpoints while the predictions for the other compounds are highly variable. 

The simulated C02 fugacity matches the initial reservoir C02 content and indicates that the pH is buffered by C02-calcite equilibrium. Further modelling was carried out using the Geochemists Workbench React and Tact modules with the thermodynamic database modified to reflect the elevated P conditions and kinetic rate parameters consistent with the Waarre C mineralogy. The Waarre C shows low reactivity and short-term predictive modelling of the system under elevated C02 content changes little with time (Fig. 1). 

There is, as always, a need for good quality data. Most of this is now available in electronic form and Chapter 11 lists some of the databases available. In spite of proclaimed good intentions, there is little systematic documentation of the successful application of plastics and their lifetimes, only examples of unexpected failure. There is a need for medium-term, lightly accelerated tests under intermediate conditions to validate the predictive models. While inspection of components at end-of-life is more prevalent than expected, there is a need for coupling it to predictive techniques to validate these techniques and to close the loop of life prediction. 

The advantage of using QSAR models for database mining is that it affords not only the compounds selection but also quantitative prediction of their activity. For illustration, we shall discuss our recent success in developing validated predictive models of anticonvulsants [51] and their application to the discovery of novel potent compounds by the means of database mining [10]. 

Create a centralized database that would lead to more robust statistical evaluations and improved prediction models (PMs). 

PG-ROUT is a deterministic river model applicable throughout the United States (57). Predictions are based on more than 500,000 United States river miles. This model also predicts concentrations under 7Q10 and mean-flow conditions. The model is driven by several large EPA databases. Predictions are made below each of the 11,500 POTWs, at drinking water intakes and at any desired mile points in the river systems. The model output includes a frequency distribution by river mile and a detailed PC database. 

One obvious application of these databases is to use them as sources of training data when developing predictive models. For example, Novotarskyi et al. (46) employed PubChem Bioassay data to develop models to predict CYP450 1A2 inhibition, and Shen et al. (47) employed the database to develop a support vector... 

The EPA uses QSARs to predict a large number of ecological effects, as well as for environmental fate within the PMN process. The EPA s website (www.epa.gov) provides a valuable source of further information on all these predictive methods, as well as a database and aquatic toxicity values and detailed information on how the models have been validated. Many of the predictive models have been brought together into the EPISUITE software (see Table 19.2 for a listing of the models available). This includes the OPPT s models used for the prediction of physical and chemical properties for new chemical substances. The EPISUITE software is downloadable free of charge (www.epa.gov/oppt/exposure/docs/episuitedl.htm). This provides not only an excellent resource for the development of QSARs, but also a transparent mechanism for the assessment of PMNs. 

New experiments may not be required if, for instance, the optimum is located at a position in the parameter space where an experiment has already been performed. If this is not the case, then the location of one or more additional experiments will be the result of the calculation step. Subsequently, a new set of experiments is run and added to the existing database. The model can then be refined using all the available data, and a new optimum can be predicted. 

---