Processing of chromatographic data for chemometric analysis of peptide profiles from cheese extracts: A novel approach

Chemometric analysis of chromatograms plays a fundamental role in characterization of foods or in detection of adulteration. Data for multivariate analysis of chromatographic profiles are usually obtained by visual matching (VM) of peaks, the identities of which, as for peptide profiles from cheese extracts, are often unknown. To avoid the main disadvantages of VM, which is subjective and time-consuming, a novel approach was developed. Fuzzy logic was employed to handle in a systematic way uncertainty in the position of peptide peaks, and chromatograms were processed by a rule-based membership function. Processed data consisted of classes of retention time wherein peak heights were accumulated by using the distance from the center of the class as a weight. The novel approach (fuzzy approach, FA) was compared with VM by using a real data set and by performing multivariate descriptive statistical techniques (principal component analysis, multidimensional scaling, and nonhierarchical cluster analysis). FA provided a fast, reliable, and objective alternative to VM and could be successfully applied for chemometric analysis of chromatographic profiles whenever knowledge of the identity of peaks is lacking or unnecessary.