Multi-layered analyses using directed partitioning to identify and discriminate between biogenic amines.

Multiple layers of statistical analyses were used to decipher the response from a single, cross-reactive conjugated polymer (1) providing enhanced classification accuracies over traditional multivariate statistical approaches. This analysis was demonstrated by classifying a series of seven biologically relevant, nonvolatile amines (i.e. biogenic amines). If only a single layer of analysis was employed (linear discriminant analysis), 89% classification accuracy was achieved lacking any concentration information. However, using this multi-layered, group-ungroup method, the analytes were first categorized based on general class of molecule (directed partitioning), i.e. aromatic, aliphatic, polyamines, with 98% accuracy. In a second analysis layer, these sub-groups were broken down into the individual molecular components, with the aliphatic and aromatic amines classifying near 99%, while the polyamine identification accuracy approached 90%. In the third layer of analysis, the concentration of the analytes in question was determined in the biologically relevant range within approximately 10% accuracy by following trends in the principle component analysis output.