Correlation Comparisons of Early and Aged Quality Traits of Pork Aged Either as Intact Loins or Case-Ready Chops

Processing facility data collection

Pigs used for this study were from 2 separate production systems with different genetic sources representative of the genetic variation found in the industry today, as confidentially identified by the packer. Pigs were immobilized using carbon dioxide and terminated via exsanguination. Carcasses were blast-chilled for approximately 90 min and then held in an equilibration cooler until approximately 22 h postmortem, at which point they were fabricated into primal cuts. Loins were further cut into boneless strap-on (longissimus dorsi with iliocostarum attached) center cut pork loins [NAMP # 414, PSO 3; North American Meat Institution (NAMI), 2014] and imaged using a VQG pork loin grading camera on the ventral side of the loin immediately after boning. The VQG camera was used to select for variation in color and intramuscular fat content among loins as described by the Klehm et al. (2018). The VQG camera was used to select loins with variation in color (max NPPC score = 5.0, min NPPC score = 1.5) and intramuscular fat content (max NPPC marbling score = 4.0, min NPPC marbling score = 1.0; Klehm et al., 2018).

A target of 36 loins were selected from each production focus for a total of 72 loins from each collection day. Four collection days were each separated by a 3 to 4 wk period in the winter-spring season, totaling 288 loins overall. Within a production focus, an equal variation in color and marbling scores were targeted for each of the designated packaging treatments: intact loins aged (ILA; n = 144) until 12 d postmortem then sliced or case-ready aged (CRA; n = 144) loins sliced at 2 d postmortem and aged until 12 d postmortem. Selected loins were removed from the de-boning line and subsequently evaluated for CIE instrumental lightness (L*), redness (a*), and yellowness (b*; CIE, 1978), pH, visual color, visual marbling, subjective firmness, and weight. Ventral evaluations were measured near the area of the 10th rib. Instrumental lightness (L*), redness (a*), and yellowness (b*) scores were measured using a Konica Minolta CR-400 colorimeter (D65 light source, 2° observer, 8 mm aperture; Minolta Camera Company, Osaka, Japan) that was calibrated with a white tile prior to evaluations. Loin pH was measured using a MPI pH meter (Meat Probes Inc. pH-Meter, Topeka, KS) with a glass electrode probe that was calibrated prior to evaluations using pH 4 and pH 7 calibration buffers stored at 4°C. A trained technician evaluated visual color and visual marbling, using National Pork Producers Council (NPPC, 1999) color and marbling standards and subjective firmness using National Pork Producers Council (NPPC, 1991) scale. Visual color was assigned using a 6-point scale (1 = pale, grayish-pink; 6 = dark, purplish-red); visual marbling was assigned using a 10 point scale (1 = 1% intramuscular fat; 10 = 10% intramuscular fat), and subjective firmness was assigned using a 5 point scale (1 = soft; 5 = firm). All evaluations were determined in half score increments. Boneless loins were vacuum-packaged and transported in coolers with ice packs to the University of Illinois Meat Science Laboratory for aging and further evaluations.

Loins aged intact

Loins selected to be ILA remained in vacuum packages in boxes and stored in a dark cooler (4°C) until 12 d postmortem. At 12 d postmortem, loins were removed from their packaging and allowed a minimum of 30 min for oxygenation of myoglobin prior to ventral loin quality evaluation. Ventral loin evaluations at 12 d postmortem were performed by the same trained technician who conducted quality evaluations on 1 d postmortem, followed the same procedures, and used the same equipment as previously described. Ventral aged evaluations consisted of CIE instrumental lightness (L*), redness (a*), and yellowness (b*), pH, visual color, visual marbling, subjective firmness, and weight. Loins were then sliced into 28 mm thick chops [NAMP # 1413, PSO 3; North American Meat Institution (NAMI), 2014] using a PUMA slicer (TREIF USA Inc., Shelton, CT). Chops were collected for meat quality evaluation starting from the anterior end of the loin, with the first selected chop containing a posterior portion of the spinalis dorsi (Fig. 1; Klehm et al., 2018). Chops were assigned to meat quality evaluations in the following order: Chop 1) Warner-Bratzler shear force (WBSF), Chop 2) proximate analyses, Chops 3 through 10) chop face surface quality evaluations. Chop 1 was vacuum-packaged and stored at 4°C until 14 d postmortem for evaluation of instrumental tenderness. Chop 2 was trimmed of subcutaneous fat, connective tissue, and accessory muscles, vacuum-packaged, and stored at –20°C until proximate analyses. Chops 3 through 10 were allowed at least 30 min for oxygenation of myoglobin before surface evaluations were conducted on 1 chop selected randomly. Evaluations for instrumental CIE lightness, redness, and yellowness and visual color, marbling, and subjective firmness used the same equipment and procedures described for 1 d postmortem evaluations.

Case-ready aged chops

Loins selected to be CRA were removed from their vacuum-sealed packaging at 2 d postmortem. Loins were sliced into 28-mm thick chops using a PUMA slicer (TREIF USA Inc., Shelton, CT). Chops were collected and quality traits were assessed using the same protocol as whole loin aged chops. Chops were packaged with an absorbent pad (Ultra ZAP Soaker, The Brenmar Company, Omaha, NE) in a 13.65cm² tray (polystyrene Cascades Evok) and overwrapped in polyvinyl chloride (PVC) film (O₂ transmission = 23,250 mL × m² × d–1, 72 gauge; Resinite Packaging Films, Borden, Inc., North Andover, MA). Individual overwrap packaged chops originating from the same loin were placed in a 35.56 cm × 50.80 cm vacuum sealable bag gas-flushed with 600 millibars of a gas mixture (0.2 to 0.4% carbon monoxide, 23 to 30% carbon dioxide, and 60 to 80% nitrogen). Packages were boxed and aged in a dark cooler at 4°C until 9 d postmortem. At 9 d postmortem, chops were removed from boxes and bulk packaging, and placed in simulated retail display under fluorescent lighting (General Electric 32W, 122cm fluorescent Kitchen/Bath bulb in an Utilitech 121 cm 2-light fixture suspended 30.5 cm above the chops, color rendering index = 78, light color temperature = 3000 K, and 2450 mean lumens) in the individual overwrapped packages, to simulate retail case conditions. Procedures for bulk MAP packaging and retail display followed generally recognized as safe (GRAS) notice No. GRN 000083 MAP system for packaging fresh cuts of case-ready muscle meat. At 12 d postmortem (after 3 d of simulated retail display), chops were evaluated for instrumental CIE lightness, redness, and yellowness with the same equipment and procedures described at 1 d postmortem (Fig. 2).

Warner-Bratzler Shear Force

The furthest anterior chops from ILA loins were selected for Warner-Bratzler shear force analysis. Chop 1 from ILA loins when sliced at 12 d PM were vacuum sealed in individual packages, boxed, and aged to 14 d postmortem in a refrigerated cooler at 4°C. At the end of the aging period, chops were removed from their packaging and an individual raw weight was recorded. A copper-constantan thermocouple (Type T, Omega Engineering, Stamford, CT) was inserted into the geometric center of each chop to accurately monitor internal cooking temperature on a digital thermometer (model 92000–00, Barnat Co, Barrington, IL). A Farberware Open Hearth grill (model 455N, Walter Kidde, Bronx, NY) was used to cook chops. Chops were initially cooked to an initial internal temperature of 36°C and then flipped. Chops were further cooked until a final internal temperature of 68°C was reached, at which point they were immediately removed and placed on an aluminum tray for cooling. After chops cooled to ambient temperature (approximately 22°C), they were weighed. Cooking loss was calculated using the following equation: Cook Loss, % = [(raw weight– cooked weight)/raw weight] × 100. Four 1.25-cm diameter cores were collected from each chop, parallel with the muscle fiber. Cores were sheared using a Texture Analyzer HD Plus (Texture Technologies Corp., Scarsdale, NY, and Stable Microsystems, Godalming, UK) equipped with a Warner-Bratzler shear force blade at a speed of 3.3 mm/s and load cell capacity of 100 kg. The force of each of the 4 cores were averaged and reported as Warner-Bratzler shear force.

Proximate analysis

Chops selected for proximate analysis were trimmed of epimysium and accessory muscles and the remaining denuded longissimus chops were vacuum sealed in individual packages, boxed and placed in a freezer at –20°C. At 12 d postmortem, chops were shipped in coolers with ice packs to the USDA Meat Animal Research Center for determination of moisture and extractable lipid. Chops were tempered and diced into cubes that were approximately 5 mm × 5 mm × 5 mm. Cubes were mixed and powdered with liquid nitrogen using a waring commercial blender and micro-mini (15 to 37 mL) blender cups. Powdered samples were maintained at –20°C until weighed. Duplicate 1.5-g samples from each chop were weighed into pre-weighed filter bags (Ankom Technology XT4 Filter Bag). Filter bags were heat sealed (Ankom model HS), placed on a drying rack in a single layer and dried for 24 h at 102°C in a drying oven (VWR 89511–414). The filter bags were placed into a weigh tin desiccator (Ankom X49) and cooled at room temperature for 15 min and then weighed. Ether-extractable fat was removed with an Ankom XT15 system. Samples were dried for 15 min at 102°C, returned to the weigh tin desiccator, cooled at room temperature for 15 min and then weighed. Moisture and fat percentage were calculated, agreement of duplicate samples was checked, and the average value was calculated for each loin.

Statistical analyses

Loin quality parameters measured at 1 d postmortem after being grouped by packaging treatment were compared using a one-way ANOVA in the MIXED procedure of SAS 9.4 (SAS Inst. Inc., Cary, NC). The model included the fixed effect of aging treatment and the random effect of slaughter date. Each individual loin served as the experimental unit for all dependent variables. Least squares means and standard errors were reported for all measured attributes and were considered different at P ≤ 0.05 for all analyses. Chop quality parameters from CRA chops at 2 d, 9 d, 12 d postmortem and ILA chops at 12 d postmortem were compared using a one-way ANOVA in the MIXED procedure of SAS 9.4. The model included the fixed effect of day postmortem and the random effect of slaughter date. Each individual chop served as the experimental unit for all dependent variables. Least squares means and standard errors were reported for all measured attributes and were considered different at P ≤ 0.05 for all analyses.

Comparison between 1 d postmortem and 12 d postmortem ventral surface loin quality (intact loin aged treatment only) using a paired t test by way of PROC T TEST in SAS. Means were considered different at P ≤ 0.05.

A z-test was used to compare two independent correlations as described in Kenny (1987). First, correlation coefficients (r) were determined between early and aged quality traits within each packaging treatment using the CORR procedure of SAS. Then, correlation coefficients were transformed with the FISHER option in SAS. The Fisher’s r to z transformation equation was defined as:

Equation 1.

Where r is the Pearson correlation coefficient, and z is the transformed correlation coefficient. The z here calculated should not be confused with the z-score associated with approximations of the normal distribution. Fisher’s transformed z values were then merged into a single data set for comparison of independent correlation coefficients. Data were compared based on the following equation:

Equation 2.

Where Z_calculated is the calculated z-score, the test statistic which follows a normal distribution, z_ILA is the transformed correlation coefficient of the ILA variables, z_CRA is the transformed correlation coefficient of the CRA variables, n_ILA is the number of observations for the ILA variables, and n_CRA is the number of observations for the CRA variables. The correlations between 1 d postmortem ventral loin traits with 14 d WBSF and cooking loss (r₁₃) was compared with the correlations between 12 d postmortem loin traits with 14 d WBSF and cooking loss (r₂₃). Because the correlations being calculated were not derived from different, independent populations and the correlations being compared shared a common trait (14 d WBSF or cooking loss), they were considered dependent. That is, r₁₂ is correlated with r₁₃, therefore the influence of the relationship of between 1 d postmortem and 12 d postmortem ventral loin quality traits (r₁₂) must be accounted for when comparing r₁₃ with r₂₃ (Kenny, 1987). To achieve this, 3 sets of Pearson correlation coefficients were calculated (r₁₃, r₂₃, and r₁₂) and the correlation coefficients transformed using Fishers r to z equation (Eq. [1]). Then, the test statistic, t was calculated using the following equation from Zou (2007):

Equation 3.

Where t is the calculated test statistic, r₁₃ is the correlation between 1 d quality traits and 14 d WBSF and cook loss, r₂₃ is the correlation between 12 d quality traits and 14 d WBSF and cook loss, and r₁₂ is the influence of the relationship between 1 d and 12 d postmortem. The resulting test statistic (t) was then compared with tcritical from a t distribution.

Correlations of the z- or t-value were considered different between packaging treatments at P ≤ 0.05 (Kenny, 1987). Correlations were considered weak at |r| < 0.35, moderate at 0.36 ≥ |r| < 0.67, and strong at |r| ≥ 0.68 (Taylor, 1990). Analysis in data sets exceeding 100 observations may result in correlation coefficients of 0.20 that have little practical importance despite being statistically different from 0 (ɑ = 0.05). Therefore, comparisons between correlations of early and aged postmortem loin quality by packaging treatment were considered significantly different at P ≤ 0.05 with a corresponding correlation coefficient of |r| ≥ 0.36 to ensure practical significance (Taylor, 1990) .