This study (purchase required, GBP£29.95) set out to improve the confidence in discriminatory models between Arabica and Rustica ground coffee.
The researchers pooled Arabica (n = 41) and Robusta (n = 21) samples from different origins and prepared in-house “calibration” mixtures.
They compared alternative approaches for processing data obtained by nontargeted ultra-high-performance liquid chromatography coupled to high-resolution tandem mass spectrometry analysis of in-house control Arabica/Robusta mixtures
They report that the most accurate results were obtained by targeted extraction of a set of five ions corresponding to kahweol palmitate, kahweol linoleate, 16-O-methylcafestol palmitate, 16-O-methylcafestol linoleate, and 16-O-methylcafestol stearate. Tthe first two compounds were identified as markers of Arabica and the latter three of Robusta using multivariate statistical models constructed in the first phase of the study on an authentic coffee sample set. This compared favourably with the most common approach currently used by laboratories of classification based on 16-O-methylcafestol quantification. Using this latter approach, more than 60% of the in-house control coffee mixtures did not meet the preset accuracy criteria.
Photo by Andrea Tummons on Unsplash
Comments