The detection and identification of metal and organic pollutants in potable water using enzyme assays suitable for sensor development

A panel of five pollutants have been identified as a model system to study: the herbicide atrazine, phenol, pentachlorophenol, cadmium and chromium. The non-specific inhibition of the activity of 23 enzymes was studied for these five pollutants at 1000 times the statutory limits. An array of six enzymes was chosen to develop four computer models from which the resultant pattern of inhibition could be interpreted using artificial neural nets. The models clearly demonstrated that the identification of pollutants using this novel approach is a feasible proposition. No false negatives were demonstrated at the concentrations tested by any of the models studied and the false positive rate was acceptably minimal. The models further demonstrated their ability to be semi-quantitative with regard to the level of pollutant present.