ABSTRACT
Postharvest processing of coffee has been shown to impact cup quality. Yeasts are known to modulate the sensory traits of the final cup of coffee after controlled fermentation at the farm. Here, we enumerated native coffee yeasts in a Nicaraguan farm during dry and semidry postharvest processing of Arabica and Robusta beans. Subsequently, 90 endogenous yeast strains were selected from the collected endogenous isolates, identified, and subjected to high-throughput fermentation and biovolatile generation in a model system mimicking postharvesting conditions. Untargeted volatile analysis by SPME-GC-MS enabled the identification of key aroma compounds generated by the yeast pool and demonstrated differences among strains. Several genera, including Pichia, Candida, and Hanseniaspora, showed both strain- and species-level variability in volatile generation and profiles. This fermentation platform and biovolatile database could represent a versatile opportunity to accelerate the development of yeast starter cultures for generating specific and desired sensory attributes in the final cup of coffee.
