Representation, comparison, and interpretation of metabolome fingerprint data for total composition analysis and quality trait investigation in potato cultivars

Abstract

Understanding attributes of crop varieties and food raw materials underlying desirable characteristics is a significant challenge. Metabolomics technology based on flow infusion electrospray ionization mass spectrometry (FIE-MS) has been used to investigate the chemical composition of potato cultivars associated with quality traits in harvested tubers. Through the combination of metabolite fingerprinting with random forest data modeling, a subset of metabolome signals explanatory of compositional differences between individual genotypes were ranked for importance. Interpretative analysis of highlighted signals based on ranking behavior, intensity correlations, and mathematical relationships of ion masses correctly predicted metabolites associated with flavor and pigmentation traits in potato tubers. GC-MS profiling was used to further validate proposed compositional differences. The potential for the development of a database strategy for large scale, long-term projects requiring comparison of chemical composition in plant breeding, mutant population analysis in functional genomics experiments, or food raw material analysis is described.