Discovery of Markers from Chromatogram Data by Machine Learning

Discovery of marker compounds for sample identification is a focus of interest in the fields of food products, clinical medicine, and polymers, among others. However, many customers feel uncertain about the analytical process from data acquisition to final determination of the marker. Using GC-MS data, this article presents an example of the workflow of marker discovery, that is, how markers are discovered from chromatogram data. Since various data analysis flows are conceivable, depending on the distinctive features of the sample and dataset, a trial-and-error process and original ideas of the analyst are necessary. The aim of this work is to provide assistance in deciding the proper direction of that process.