Aroma is one of the most important quality attributes of bread or any food. It will determine whether the product will be tasted and eaten in the first instance and is a major factor in establishing acceptability and preference. The dominant preference by consumers of bread made from refined flour in contrast to whole wheat flour is at least in part due to the strong and different aroma of whole wheat bread. White wheats may have an advantage over red wheat in this regard according to some industry reports, but the science is extremely limited. The goal of this research was to add more science-based knowledge to this topic via the use of machine olfaction technology, specifically electronic (E) nose and gas chromatography combined with mass spectrometry (GC-MS). A state-of-the-art E-nose system (AlphaMOS FOX 3000) with metal oxide sensors (MOS) was used to capture aroma volatiles from crumb, crust and whole slices of breads made from sound Canadian Western Red Spring (CWRS) wheat as well as representative samples of two hard white wheats, viz. Snowbird, a cultivar belonging to the Canada Western Hard White Spring (CWHWS) class of wheat, and Platte, a U.S. Hard White Winter (HWW) wheat. The same CWRS wheat provided the base flour for all the breads. A commercial formula and size format was used to produce breads from four flours for the study, i.e. refined CWRS wheat, and three whole wheat flours comprising blends of 85% CWRS flour and 15% bran from CWRS wheat, Snowbird and Platte. As there was no established protocol in the literature to evaluate bread aroma by E-nose, one was developed. Five temperatures (35, 40, 45, 50 and 60°C) were tested along with two incubation times (5 and 10 min) and four sample sizes (0.05, 0.1, 0.25 and 0.50 g) of ground bread crumbs. Through optimization using E-nose software including principal component analysis, a procedure was adopted using 40 °C, 5 min incubation time and 0.05 g of sample to acquire MOS data from 12 sensors for crust, crumb and whole slices of refined and whole wheat breads. Multivariate analysis methods were used to evaluate the capabilities of the E-nose system to discriminate and correctly classify samples according to bread type. Data for analysis comprised approximately 24 samples each of crust, crumb and whole slices randomly selected from three loaves each of refined and the three whole wheat breads. Results varied according to the nature of the sample, i.e. crust, crumb or whole slices. For crusts, the greatest distinction in aroma was found between refined and whole wheat breads. Refined bread crust was correctly classified 67% of the time. When refined bread crust was misclassified, samples were confused with whole white wheat crust predominantly from Platte bread. For whole wheat bread crusts, the pattern of classification depended mainly on bran colour. Whole wheat bread crust samples had correct classification scores in the range 54-58%. When misclassified, whole wheat CWRS crust was equally confused with the aroma of crust of the white whole wheat breads, Platte and Snowbird. Whole wheat Platte crust tended to be misclassified with the counterpart white whole wheat Snowbird or refined bread crust. In contrast, Snowbird whole wheat crust tended to be misclassified as either its counterpart HW wheat Platte or whole CWRS wheat. Accordingly, Platte bread crust appeared to possess an aroma more in line with refined wheat bread as opposed to whole wheat bread. For bread crumb, the pattern of E-nose differentiation of samples was different. In this case, CWRS whole wheat bread aroma was clearly and perfectly distinguished from the crumb aroma of all the other breads, either whole white wheat or refined CWRS. The latter tended to cluster on its own, as might be expected, and had a correct classification score of 75%, with the balance of samples largely misclassified as Snowbird crumb. Whole wheat Platte and Snowbird bread crumb had identical correct classification scores of 42%, and were similarly confused with the other’s aroma (average 3 4% classified) or the aroma of refined wheat bread (average 21% classified). E-nose results for crumb indicated a clear distinction in aroma between the hard red and white wheats in this study. E-nose analysis of bread samples representing whole slices produced results that provided unsatisfactory discrimination among bread types likely due to the blending of the different aromas of constituent crust and crumb. For whole slices, discrimination between refined and whole wheat breads was substantially lower than that for either crust and crumb samples. Based on this result, analysis of samples that combine both crust and crumb is not recommended for sensory analysis of bread, whether by instruments or human sensory panel. Further understanding of the differences between different types of bread made from refined wheat flour and whole wheat, and how the inclusion of bran from red and white-grained wheats modifies the composition and content of volatile and non-volatile compounds in crust and crumb was determined by gas chromatography-mass spectrometry (GC-MS). In total, 50 compounds were found, the greater majority of which have been previously reported in bread. Major Maillard reaction compounds like furfural, 2-furanmethanol, pyranone, maltol and 5-hydroxymethyl- 2-furancarboxaldehyde were present in highest concentration in whole CWRS bread. Significantly fewer compounds were found in the crust and crumb of CWRS refined wheat bread compared to the other whole wheat breads. In contrast, whole CWRS bread crumb and crust had the highest number of compounds, and in considerably higher total concentration compared to the other two whole white wheat breads, Snowbird and Platte. The higher concentration and number of compounds in whole CWRS bread was attributed to the wheat bran fraction. White whole wheat breads, Snowbird and Platte, had a total number of compounds in crust and crumb approximately intermediate between refined and whole CWRS bread, although Platte whole wheat bread crust was closer to refined bread crust in compound numbers. In terms of total compound concentrations, crust and crumb samples of the whole white breads were clearly more similar to refined CWRS bread, and in the case of whole wheat Platte bread crust, compound concentrations were much lower. On the whole, these aggregate totals of compound numbers and concentrations by GC-MS mirrored the discrimination and classification results obtained by E-nose, and supported the contention that whole wheat bread made with white wheat bran was milder in aroma compared to bread formulated using red wheat bran. While the number of samples of red and white wheats were very few in this study, results support the contention that different wheat genotypes and specifically, the bran tissue of these genotypes, contain differences in compound composition and/or concentration which when processed by breadmaking, manifest volatiles characteristic of those genotypes even between genotypes possessing the same colour of bran. E-nose instrumentation appears to be very capable of accommodating these sorts of complex tasks on fresh bread. It would be highly beneficial in future research to carry out similar studies in parallel with a human sensory panel, and ideally with many more genotypes of red and white grained wheat with an aim to firmly establish the relative superiority of particular genotypes to produce whole wheat bread with aroma profiles more similar to those of white pan bread. The long term goal of such studies would be to foster increased consumption of whole wheat products and constituent bioactive compounds which confer favourable health benefits in the general population.

Aroma is one of the most important quality attributes of bread or any food. It will determine whether the product will be tasted and eaten in the first instance and is a major factor in establishing acceptability and preference. The dominant preference by consumers of bread made from refined flour in contrast to whole wheat flour is at least in part due to the strong and different aroma of whole wheat bread. White wheats may have an advantage over red wheat in this regard according to some industry reports, but the science is extremely limited. The goal of this research was to add more science-based knowledge to this topic via the use of machine olfaction technology, specifically electronic (E) nose and gas chromatography combined with mass spectrometry (GC-MS). A state-of-the-art E-nose system (AlphaMOS FOX 3000) with metal oxide sensors (MOS) was used to capture aroma volatiles from crumb, crust and whole slices of breads made from sound Canadian Western Red Spring (CWRS) wheat as well as representative samples of two hard white wheats, viz. Snowbird, a cultivar belonging to the Canada Western Hard White Spring (CWHWS) class of wheat, and Platte, a U.S. Hard White Winter (HWW) wheat. The same CWRS wheat provided the base flour for all the breads. A commercial formula and size format was used to produce breads from four flours for the study, i.e. refined CWRS wheat, and three whole wheat flours comprising blends of 85% CWRS flour and 15% bran from CWRS wheat, Snowbird and Platte. As there was no established protocol in the literature to evaluate bread aroma by E-nose, one was developed. Five temperatures (35, 40, 45, 50 and 60°C) were tested along with two incubation times (5 and 10 min) and four sample sizes (0.05, 0.1, 0.25 and 0.50 g) of ground bread crumbs. Through optimization using E-nose software including principal component analysis, a procedure was adopted using 40 °C, 5 min incubation time and 0.05 g of sample to acquire MOS data from 12 sensors for crust, crumb and whole slices of refined and whole wheat breads. Multivariate analysis methods were used to evaluate the capabilities of the E-nose system to discriminate and correctly classify samples according to bread type. Data for analysis comprised approximately 24 samples each of crust, crumb and whole slices randomly selected from three loaves each of refined and the three whole wheat breads. Results varied according to the nature of the sample, i.e. crust, crumb or whole slices. For crusts, the greatest distinction in aroma was found between refined and whole wheat breads. Refined bread crust was correctly classified 67% of the time. When refined bread crust was misclassified, samples were confused with whole white wheat crust predominantly from Platte bread. For whole wheat bread crusts, the pattern of classification depended mainly on bran colour. Whole wheat bread crust samples had correct classification scores in the range 54-58%. When misclassified, whole wheat CWRS crust was equally confused with the aroma of crust of the white whole wheat breads, Platte and Snowbird. Whole wheat Platte crust tended to be misclassified with the counterpart white whole wheat Snowbird or refined bread crust. In contrast, Snowbird whole wheat crust tended to be misclassified as either its counterpart HW wheat Platte or whole CWRS wheat. Accordingly, Platte bread crust appeared to possess an aroma more in line with refined wheat bread as opposed to whole wheat bread. For bread crumb, the pattern of E-nose differentiation of samples was different. In this case, CWRS whole wheat bread aroma was clearly and perfectly distinguished from the crumb aroma of all the other breads, either whole white wheat or refined CWRS. The latter tended to cluster on its own, as might be expected, and had a correct classification score of 75%, with the balance of samples largely misclassified as Snowbird crumb. Whole wheat Platte and Snowbird bread crumb had identical correct classification scores of 42%, and were similarly confused with the other’s aroma (average 3 4% classified) or the aroma of refined wheat bread (average 21% classified). E-nose results for crumb indicated a clear distinction in aroma between the hard red and white wheats in this study. E-nose analysis of bread samples representing whole slices produced results that provided unsatisfactory discrimination among bread types likely due to the blending of the different aromas of constituent crust and crumb. For whole slices, discrimination between refined and whole wheat breads was substantially lower than that for either crust and crumb samples. Based on this result, analysis of samples that combine both crust and crumb is not recommended for sensory analysis of bread, whether by instruments or human sensory panel. Further understanding of the differences between different types of bread made from refined wheat flour and whole wheat, and how the inclusion of bran from red and white-grained wheats modifies the composition and content of volatile and non-volatile compounds in crust and crumb was determined by gas chromatography-mass spectrometry (GC-MS). In total, 50 compounds were found, the greater majority of which have been previously reported in bread. Major Maillard reaction compounds like furfural, 2-furanmethanol, pyranone, maltol and 5-hydroxymethyl- 2-furancarboxaldehyde were present in highest concentration in whole CWRS bread. Significantly fewer compounds were found in the crust and crumb of CWRS refined wheat bread compared to the other whole wheat breads. In contrast, whole CWRS bread crumb and crust had the highest number of compounds, and in considerably higher total concentration compared to the other two whole white wheat breads, Snowbird and Platte. The higher concentration and number of compounds in whole CWRS bread was attributed to the wheat bran fraction. White whole wheat breads, Snowbird and Platte, had a total number of compounds in crust and crumb approximately intermediate between refined and whole CWRS bread, although Platte whole wheat bread crust was closer to refined bread crust in compound numbers. In terms of total compound concentrations, crust and crumb samples of the whole white breads were clearly more similar to refined CWRS bread, and in the case of whole wheat Platte bread crust, compound concentrations were much lower. On the whole, these aggregate totals of compound numbers and concentrations by GC-MS mirrored the discrimination and classification results obtained by E-nose, and supported the contention that whole wheat bread made with white wheat bran was milder in aroma compared to bread formulated using red wheat bran. While the number of samples of red and white wheats were very few in this study, results support the contention that different wheat genotypes and specifically, the bran tissue of these genotypes, contain differences in compound composition and/or concentration which when processed by breadmaking, manifest volatiles characteristic of those genotypes even between genotypes possessing the same colour of bran. E-nose instrumentation appears to be very capable of accommodating these sorts of complex tasks on fresh bread. It would be highly beneficial in future research to carry out similar studies in parallel with a human sensory panel, and ideally with many more genotypes of red and white grained wheat with an aim to firmly establish the relative superiority of particular genotypes to produce whole wheat bread with aroma profiles more similar to those of white pan bread. The long term goal of such studies would be to foster increased consumption of whole wheat products and constituent bioactive compounds which confer favourable health benefits in the general population.