Development and application of a data processing method for food metabolomics analysis

Abstract

Food metabolomics is described as the implementation of metabolomics to food systems such as food materials, food processing, and food nutrition. These applications generally create large amounts of data, and although technologies exist to analyze these data and different tools exist for various ecosystems, downstream analysis is still a challenge and the tools are not integrated into a single method. In this article, we developed a data processing method for untargeted LC-MS data in metabolomics, derived from the integration of computational MS tools from OpenMS into the workflow system Konstanz Information Miner (KNIME). This method can analyze raw MS data and produce high-quality visualization. A MS1 spectra-based identification, two MS2 spectra-based identification workflows and a GNPSExport-GNPS workflow are included in this method. Compared with conventional approaches, the results of MS1&MS2 spectra-based identification workflows are combined in this approach via the tolerance of retention times and mass to charge ratios (m/z), which can greatly reduce the rate of false positives in metabolomics datasets. In our example, filtering with the tolerance removed more than 50% of the possible identifications while retaining 90% of the correct identification. The results demonstrated that the developed method is a rapid and reliable method for food metabolomics data processing.