Intensities of Sensory Attributes in High- and Normal-Oleic Cultivars in the Uniform Peanut Performance Test

The high-oleic fatty acid seed oil trait in peanut (Arachis hypogaea L.) was discovered in the 1980s by researchers at the Univ. of Florida (Norden et al., 1987). Its inheritance was deduced (Moore and Knauft, 1989; Knauft et al., 1993; López et al. 2001; Isleib et al. 2006d), and the trait was transferred by backcrossing into the runner-type Sunrunner cultivar (Norden et al., 1985). The first high-oleic cultivars were released in the 1990s (Gorbet and Knauft, 1997, 2000). Sunrunner was highly susceptible to the tomato spotted wilt (TSW) caused by Tomato spotted wilt tospovirus, and the new SunOleic cultivars were as well. The shortcomings of the first high-oleic cultivars and the need to pay a royalty to the Univ. of Florida due to their ownership of a US Utility Patent on the high-oleic trait made peanut growers reluctant to adopt high-oleic cultivars in spite of their clear advantages in terms of finished peanut products, particularly with respect to shelf life (Mozingo, 2004). Processors prefer high-oleic peanuts for some products, and most US peanut breeders have adopted elevated oleic acid level as an objective in their programs.

As more high-oleic peanut cultivars are released in the USA, processors and shellers have inquired whether or not flavor differences exist between high-oleic cultivars and normal ones. The Uniform Peanut Performance Test (UPPT) program provides a mechanism whereby advanced breeding lines of the runner or virginia market type from all three of the major US production regions are compared directly (Branch et al., 2014). Many UPPT entries are released as cultivars subsequent to their testing in the UPPT. In the years from 2001 through 2013, many high- and normal-oleic cultivars have been UPPT entries that were tested for flavor profile by the USDA-ARS Market Quality and Handling Research Unit (MQHRU) in Raleigh, NC. Even though not all cultivars were tested every year, there was sufficient overlap among entries between years, and check cultivars Florunner (Norden et al., 1969) and NC 7 (Wynne et al., 1979) were included in most individual UPPT tests, allowing comparison among the larger group of cultivars.

There has been some review of the relative flavor profiles of high-and normal-oleic breeding lines (Isleib et al., 2006a) including a direct comparison of the near-isogenic cultivars, NC 7 and Brantley (Isleib et al., 2006c). Based on that limited data, there were not large differences in flavor between the two groups of breeding lines and cultivars. More data are available now. It was the objective of this work to use the UPPT flavor data to determine whether or not systematic differences exist between normal- and high-oleic peanut cultivars with respect to their flavor profiles.

The UPPT is a trial series that has continued for decades (Branch et al., 2014) although collection of composition and sensory data did not begin until the 2001 trials. Each year, participants in the UPPT program grow a replicated field trial of breeding lines contributed by programs in southern US states including state-funded programs in Alabama, Florida, Georgia, North Carolina, Texas and Virginia and USDA-ARS programs in Georgia and Oklahoma. Test sites include three in the Virginia-Carolina production area (Suffolk, VA, Lewiston, NC, and Blackville, SC), three in the Southeastern area (Tifton, GA, Marianna, FL, and Headland, AL), and six in the Southwestern area (Brownfield, TX, La Mesa, TX, Seminole, TX, Pearsall, TX, Stephenville, TX, and Fort Cobb, OK). No more than three locations in the Southwest were used in a single year. After pods from the field reps were graded by the participants, a sample of pods was composited across reps and sent to the USDA-ARS National Peanut Research Laboratory at Dawson, GA, where an extensive array of physical properties were measured on pods and seeds. Samples of the predominating fractions of the shelled seeds (extra-large kernels or “ELK” for virginia-type breeding lines and jumbo or medium kernels for runner-type lines) were sent to the USDA-ARS MQHRU in Raleigh, NC.

Personnel in the USDA-ARS MQHRU roasted the peanuts to a Hunter L value of 49±1, ground the roasted peanuts to paste to eliminate kernel size and texture as sensory criteria, and presented the paste samples to a trained descriptive sensory panel that scored an array of sensory attributes or “flavor notes” using the lexicon developed by Johnsen et al. (2007). This lexicon has a continuous scale of intensity from 0 (no perception of the flavor) to 15 (most intense flavor). Paste color (brightness) was measured on the Hunter L scale on room-temperature paste using a Hunter Laboratories DP-9000 colorimeter equipped with a D25 L optical sensor (Hunter Associates Laboratories, Reston, VA).

Since the inception of the sensory evaluation of UPPT samples in 2001, sensory intensity scores for samples from the UPPT have been accumulated in a database maintained by the NCSU peanut breeding program. This database includes data on the fatty acid profiles of the various samples as well as the names assigned to breeding lines that were released as cultivars after inclusion in the program either as “official” entries, i.e., ones grown in all UPPT tests that year, or as “local options,” i.e., entries grown at one or a few locations at the discretion of the individual participant. Local options often include cultivars commonly grown commercially in the area. If a cultivar was tested in the UPPT prior to inception of the flavor assessment program, it may appear in the database as a local option.

Thirty-two normal- and 27 high-oleic cultivars were identified on the basis of fatty acid concentrations measured as part of the UPPT quality assessment, their data extracted from the database, and the data subjected to an unbalanced analysis of variance using the general linear models procedure (PROC GLM) of SAS Ver. 9.3 (SAS Inst., Cary, NC). For the roasted peanut, sweet aromatic, sweet, and bitter sensory attributes, roast color linear and quadratic effects and the intensity of the fruity attribute were tested as covariates. The model selected for final use included any covariates found to be simultaneously significant (P ≤ 0.05). Means adjusted to a common environmental effect were computed for the oleic acid groups and for individual lines within groups using the “least squares means” (LSMEAN) option. Means were separated using t-tests (P ≤ 0.05).

There was no indication that normal- and high-oleic cultivars differed in any sensory trait except stale/cardboard (Tables 1, 2). The difference for that trait was very small (0.28 vs. −0.16 flavor intensity units (fiu), P = 0.0008) and favored high-oleics. Although statistically different, the intensity scores for both groups were close to zero, i.e., most often imperceptible, so it is questionable whether or not the statistically significant difference rises to the level of biological or economic significance. For UPPT samples, it is customary to assess flavor as quickly as possible after processing, and if paste samples must be stored, they are usually held at −15 C, a temperature at which oxidation and generation of the stale/cardboard flavor are negligible (Pattee et al., 2002). Another experiment might be designed to assess flavor in samples of normal- and high-oleic cultivars maintained over longer periods of time at higher temperatures, but the effect of the high-oleic trait on retardation of the onset of rancid flavor is well known (Mozingo, 2004), and the apparent effect of the high-oleic trait on the other sensory attributes was not detected. Several of the negative sensory attributes such as raw beany or dark roast or so-called “off flavors” such as wood-hulls-skins, astringent, earthy, painty, metallic, and sour exhibited no genotypic variation whatever after accounting for the average effects of specific tests.

It is of interest that there was statistically significant variation among lines within oleic acid groups for the generally positive sensory attributes roasted peanutty, sweet aromatic, and sweet and for the generally negative ones bitter, fruity/fermented, stale/cardboard, and plastic/chemical. Roasted peanutty intensities ranged from 3.92 to 5.15 fiu for high-oleics and 4.26 to 4.89 fiu for normal-oleics. Sweet aromatic intensities ranged from 2.41 to 3.24 fiu for high-oleics and 2.71 to 3.24 fiu for normal-oleics. Sweet intensities ranged from 1.91 to 2.70 fiu for high-oleics and 2.02 to 2.70 fiu for normal-oleics. Bitter intensities ranged from 2.35 to 3.05 fiu for high-oleics and 2.30 to 3.05 fiu for normal-oleics. Wood/hulls/skins intensities ranged from 2.96 to 3.21 fiu for high-oleics and 2.95 to 3.21 fiu for normal-oleics. Stale/cardboardy intensities ranged from 0.13 to 0.38 fiu for high-oleics and −0.07 to 0.38 fiu for normal-oleics. Fruity/fermented intensities ranged from 0.11 to 0.62 fiu for high-oleics and −0.44 to 0.62 fiu for normal-oleics. Plastic/chemical intensities ranged from 0.17 to 0.44 fiu for high-oleics and −0.04 to 0.44 fiu for normal-oleics.

In each case, if it was a positive sensory attribute, the upper limit for the high-oleic cultivars was greater than or statistically equivalent to that for the normal-oleics; if it was a negative attribute, then the lower limit for the high-oleics was less than or statistically equivalent to the limit for normal-oleics. This suggests that it is possible to identify high-oleic cultivars with superior flavor profiles, at least as good as profiles of normal-oleic cultivars.