Reliable discriminant analysis tool for controlling the roast degree of coffee samples through chemical markers approach
Roasting is one of the most influencing stages of coffee processing. Accordingly, a discriminant analysis (DA) was carried out with the objective of identifying key compounds (chemical markers) that enable a differentiation of coffee samples according to their roasting degree. For this, chromatographic data of the volatile fraction of 21 coffee samples submitted to distinct roasting treatments (Light, Medium, Dark, and French Roasts) were employed. Using three discriminant functions that rely on only ten chemical markers, it was possible to explain 100 % of the variance of the data points. If two functions are used, the surprisingly high value of 99.4 % is achieved. The model was cross-validated, and the main function successfully passed a permutation test using two statistical indicators. It was found that half of the markers belong to the pyrazines family, known to grant sensorial notes related to roasted hazelnut and peanuts. In the whole, this essay demonstrates the usefulness of DA as a tool to control the quality of roasting treatment of coffee and can be further extended with advantage to the eight roasting degrees of the AGTRON Roasting Classification as soon as larger databases become available.
KeywordsChemical markers Coffee quality Discriminant analysis Roasting Volatiles composition
The authors thank the Brazilian National Research Council (CNPq) and the Coordination for the Improvement of Higher Level Personnel (CAPES) for financial support. This work was developed within the scope of the project CICECO-Aveiro Institute of Materials, POCI-01-0145-FEDER-007679 (FCT Ref. UID/CTM/50011/2013), financed by national funds through the FCT/MEC and when appropriate co-financed by FEDER under the PT2020 Partnership Agreement.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflicts of interest.
Compliance with ethics requirements
This article does not contain any studies with human or animal subjects.
- 4.Costa LL, Toci AT, Silveira CLP, Herszkowicz N, Pinto M, Farah A (2010) Discrimination of Brazilian C. Canephora by region using mineral composition. In: Proceedings of the 23rd International Coll on the Chemistry of Coffee, BaliGoogle Scholar
- 5.Damoradam S (2007) Amino acids, peptides and proteins. In: Fennema OR (ed) Food chemistry. Marcel Dekker, New York, pp 412–416Google Scholar
- 8.Eggers R (2005) Roasting techniques. In: Illy A, Viani R (eds) Espresso coffee: the science of quality, 2nd edn. Elsevier Academic Press, London, pp 184–191Google Scholar
- 13.McLachlan G (2004) Discriminant analysis and statistical pattern recognition, vol 544. Wiley, Chicago, pp 168–211Google Scholar
- 16.Morais SAL, Aquino FJT, Chang R, Nascimento EA, Oliveira GA, Santos NC (2007) Chemical analysis of Arabica coffee (Coffea arabica L.) and defective beans submitted to different degrees of roasting. Coffee Sci 2:97–111Google Scholar
- 25.Toci AT, Silva CM, Fernandes F (2014) Effect of roasting speed on the volatile composition of different quality coffee blends roasted in an industrial semi-fluidized bed roaster and in a small scale fluidized bed roaster. In: 25nd international conference on coffee scienceGoogle Scholar
- 28.Westerhuis JA, Hoefsloot HCJ, Smit S, Vis DJ, Smilde AK, van Velzen EJJ, van Duijnhoven JPM, van Dorsten FA (2008) Assessment of PLSDA cross validation. Metabolomics 4:81–89Google Scholar
- 29.Whistler RL, BeMiller JN (1997) Carbohydrate chemistry for food scientists. Marcel Dekker, New York, pp 171–174Google Scholar