Identification and exploration of the components of a desirable pecan flavor

Abstract

The pecan, [Carya illinoinensis (Wangenh) K. Koch], has a long history of cultivation and economic value. Knowledge of the compositional differences that exist between cultivars is important to the marketing of pecan varieties. The objectives of this study were to A) profile flavors for various pecans, B) determine flavor differences attributed to preparation method, C) find characteristics of acceptable pecan flavor, and D) evaluate sources of pecan flavor variation through chemical profiling. The flavor profiles of eight pecan cultivars ('Chetopa,' 'Giles,' 'Kanza,' 'Lakota,' 'Major,' 'Maramec,' 'Pawnee,' and 'Witte') were evaluated using descriptive sensory analysis under raw, roasted, and candied preparation methods. A trained panel evaluated samples for 21 flavor attributes. Five of these attributes differed significantly (p ≤ 0.05) between cultivars, while the preparation method significantly affected 17 attributes. Unique profiles were exhibited for each sample, with the 'Pawnee' and 'Lakota' samples displaying outlying characteristics for certain attributes. These results were used to select cultivars with varied but desirable pecan flavor. 102 nut consumers evaluated 'Kanza,' 'Maramec,' 'Pawnee,' and 'Witte' pecans under raw and roasted conditions for liking and flavor intensity. All samples were met with generally positive consumer acceptance, but three consumer segments were formed based on Overall Flavor Liking scores. Segment 1 was driven by cultivar differences, segment 2 by preparation method, and segment 3 by a combination of these factors. The largest drivers of consumer liking related to the roasting process. Chemical differences between cultivars under raw and roasted preparation methods were explored through fatty acid profiling (8 cultivars) and volatile olfactory compound profiles ('Kanza,' 'Maramec,' 'Pawnee,' and 'Witte'). Fatty acid profile variation could generally be attributed to cultivar differences, not changing much with the roasting process. Linoleic, palmitic, and stearic acids were correlated with more roasted-type attributes while linolenic acid was associated with dry, unfavorable attributes. 51 compounds with olfactory contribution were tentatively identified, 33 of which were found in all samples. Chemical profiles were unique to each sample, but some trends were apparent. The roasted 'Pawnee' sample, having many desirable flavor attributes, being met with great consumer acceptance, and having a composition that is associated with preferential attributes, may serve as a good standard for flavor.