BACKGROUND: Although the evaluation of beer is conducted by sensory experts, we cannot neglect the influence of human factors and subjectivity. This problem could be solved by the chemical analysis of the volatile part of beer aroma and, from this, we can build a database for the construction of a model that classifies samples in a comparable manner to sensory assessment. RESULTS: Twenty-two batches of the same beer brand were assessed by sensory evaluation and described chemically in terms of the contents of alcohols and esters (n = 9), hop essential oil compounds (n = 15) using gas chromatography (GC) and other aroma volatiles (n = 33) as analysed by head space solid-phase microextraction (SPME)-GC/mass spectrometry. The best match of 91% with respect to sample classification on the basis of chemical analyses to sensory scores was achieved with a dataset of results from headspace-SPME and analyses of higher alcohols and esters by regularized discriminant analysis. CONCLUSION: The results of the present study show that deviations in beer aroma are not a consequence of a permanent repeatable error in brewing process, nor are they influenced by raw materials, but, instead, they are a consequence of alcoholic fermentation. Sensory analysis could be replaced with chemical/statistical analysis on an appropriate data set and for a distinct beer brand. The good results achieved confirm our approach; however, for different beer brands or types, this method should be optimised.
BACKGROUND: Although the evaluation of beer is conducted by sensory experts, we cannot neglect the influence of human factors and subjectivity. This problem could be solved by the chemical analysis of the volatile part of beer aroma and, from this, we can build a database for the construction of a model that classifies samples in a comparable manner to sensory assessment. RESULTS: Twenty-two batches of the same beer brand were assessed by sensory evaluation and described chemically in terms of the contents of alcohols and esters (n = 9), hopessential oil compounds (n = 15) using gas chromatography (GC) and other aroma volatiles (n = 33) as analysed by head space solid-phase microextraction (SPME)-GC/mass spectrometry. The best match of 91% with respect to sample classification on the basis of chemical analyses to sensory scores was achieved with a dataset of results from headspace-SPME and analyses of higher alcohols and esters by regularized discriminant analysis. CONCLUSION: The results of the present study show that deviations in beer aroma are not a consequence of a permanent repeatable error in brewing process, nor are they influenced by raw materials, but, instead, they are a consequence of alcoholic fermentation. Sensory analysis could be replaced with chemical/statistical analysis on an appropriate data set and for a distinct beer brand. The good results achieved confirm our approach; however, for different beer brands or types, this method should be optimised.