Potential markers of coffee genotypes grown in different Brazilian regions: A metabolomics approach

- Metabolite profiling of coffee genotypes grown in different Brazilian regions
- We used chemometric tools PCA and PLS-DA to assess the metabolite profiles.
- GC-Q/MS ally to chemometric analysis is suitable to discriminate coffee samples.
- Different regions have different metabolite profiles of coffee seeds.
- Different genotypes have different metabolite profiles.

Seeds from different coffee species and coffee from different continents or countries have very distinct chemical composition. However, the differences between genotypes grown at micro-regional levels with similar geographical characteristics are still unclear. In this study, we highlighted the need of using metabolite profiling instead of the usual targeted analysis as a more powerful tool to describe the slight differences between coffees of the same species grown in close origins. Thus, our study focused on finding potential metabolite markers to describe differences of Coffea arabica L. genotypes (Mundo Novo and Bourbons) grown in Brazilian coffee producing municipalities (Lavras, Santo Antônio do Amparo-SAA, and São Sebastião da Grama-SSG). Using the metabolomics approach, 44 metabolites were identified, and some showed great potential for origin and genotype differentiation. The Partial Least Square Discriminant Analysis - PLS-DA model showed that the SAA coffee samples had the most differentiated metabolite profile (approximately 95% accuracy) compared to the other municipalities. The samples from Lavras and SGG had similar profiles (model accuracy of approximately 50%). Potential metabolite markers for the SAA samples included galactinol, fructose, malic acid, oxalic acid, d-glucose, d-sorbitol, galactinol, and myo-inositol. The model used to differentiate the Bourbon and the MN genotype showed 100% accuracy indicating very different metabolite profiles. The features that were most influential in differentiating genotype were: 5-CQA, oxalic acid, galactinol, nicotinic acid, caffeine, and caffeic acid (Bourbon) and myo-inositol, quinic acid, malic acid, fructose, and d-glucose (MN). Enhancing subtle differences in the data by combining information from GC-Q/MS and multivariate analysis resulted in the identification of coffee origins and genotypes as well as the identification of potential markers.