METHOD FOR ANALYTE CLASSIFICATION BY SC PROFILES

The precision of identification of analyte composition in a sample, where th e possible analytes each provide a series of va- lues for characteristic parameters; in particular where the parameters are generated by cross-reaction with specific binding rea- gents, is enhanced by applying pattern recognition techniques. Samples to be tested are evaluated with respect to each survey par- ameter to obtain a pattern of parameter values with respect to each analyte at a given concentration. In the case of the use of a panel of specific binding reagents, the samples to be tested are reacted wit h this panel and the affinities at various analyte concen- trations are determined. This results in a databank of "SC profiles" for kno wn concentrations of each analyte. This databank is stored in a computationally accessible form, which then can be matched again st SC profiles obtained by testing unknown sam- ples. 2107474 9217784 PCTABS00016 The precision of identification of analyte composition in a sample, where the possible analytes each provide a series of values for characteristic parameters; in particular where the parameters are generated by cross-reaction with specific binding reagents, is enhanced by applying pattern recognition techniques. Samples to be tested are evaluated with respect to each survey parameter to obtain a pattern of parameter values with respect to each analyte at a given concentration. In the case of the use of a panel of specific binding reagents, the samples to be tested are reacted with this panel and the affinities at various analyte concentrations are determined. This results in a databank of "SC profiles" for known concentrations of each analyte. This databank is stored in a computationally accessible form, which then can be matched against SC profiles obtained by testing unknown samples.