Relationships of Myoglobin Forms at Grading to Beef Longissimus Thoracis Lean Color, Color Stability, and Tenderness

Carcass sampling

Carcasses were sampled from 2 different processing plants for the present experiment. No screening based on lean color was conducted in either plant. Steers (n = 185) from the USMARC Germplasm Evaluation Project, described by Ahlberg et al. (2016), were harvested in Plant A. This plant utilized a 36-h chilling period. These carcasses represented marbling scores ranging from Traces to Slightly Abundant. Measurements of visible and near infrared reflectance (VIS-NIR) spectra were collected on the anterior surface of the longissimus thoracis exposed by ribbing at the 12^th–13^th rib interface (herein referred to as ribeye) on the left side of the carcass using the equipment described, in detail, by Shackelford et al. (2012) immediately after the carcasses were ribbed (unbloomed) and after the carcasses were graded and sorted (bloomed). Two identical custom-made VIS-NIR units (ASD Inc., Boulder, CO) were used to collect the bloomed and unbloomed spectra. These units had sampling heads designed to fit onto the ribbed carcass (Shackelford et al., 2012) with a 50 mm sampling area. These VIS-NIR spectra were collected within 1–2 min after ribbing for the unbloomed spectra and approximately 17 min later for the bloomed spectra. Carcasses were placed on a stationary rail for sample collection, as described below.

In plant B, VIS-NIR spectra were collected on the ribeye on the left side of the carcass immediately after the side was ribbed (within 1–3 min) for grading. A second spectrum was collected with an identical VIS-NIR system immediately after carcass grades were applied with an identical system. Carcasses included in the present experiment (n = 300) were sampled from typical production across many production lots and varied widely in carcass grade. Marbling scores ranged from Traces to Slightly Abundant. Chill time before grading was approximately 24 h. The interval between ribbing and grading was approximately 15 min.

Graded carcass sides were placed on a stationary rail, and a 2.54-cm-thick steak was removed from the 13^th rib region of the longissimus thoracis as described by King et al. (2012). Steaks were placed in black plastic trays with the newly exposed surface (posterior surface) facing upwards. The trays were deeper than the steak thickness, which were placed in insulated coolers with cardboard separators between layers. Care was taken to ensure that nothing contacted the newly exposed steak surface before packaging.

Steaks were transported to the USMARC abattoir and were placed on polystyrene trays and wrapped in oxygen-permeable polyvinylchloride (PVC) film (Stretchable meat film 55003815; Prime Source, St. Louis, MO; oxygen transmission rate = 1.4 mL · cm⁻² · 24 h⁻¹ at 23°C). Simulated retail display occurred in a refrigerated room (1°C). Lighting intensity of 2000 lux at the steak surface was produced by fluorescent bulbs (color temperature = 3,500 K; color rendering index = 86; 32 W T8 Ecolux bulb, model F32T8/SPX35 GE; GE Lighting, Cleveland, OH). Steaks were packaged and placed in a simulated retail display within 6 h of grading.

On days 0, 1, 4, 7, and 11 of simulated retail display, Commission Internationale de l’Éclairage (CIE) color space values were collected in duplicate on each steak with a Hunter Miniscan EZ (HunterLab, Reston, VA). The colorimeter used Illuminant A with a 10° observer and a 2.5 cm aperture. Redness (a*) and yellowness (b*) were used to calculate chroma [(a*²+ b*²)^0.5] and hue angle [Arctangent(b*/a*) × 180/3.142] (King et al., 2023).

Day 11 of the simulated retail display coincided with 13- and 12-d postmortem for plants A and B, respectively. After color measurement on day 11, steaks were trimmed free of subcutaneous fat and held until day 14 postmortem for slice shear force determination. Steaks were held at 5°C overnight before slice shear force determination. Steaks were cooked to a final internal temperature of 70°C on a TBG-60 Magigrill belt grill (MagiKitch’n Inc., Quakertown, PA, USA) as described by Wheeler et al. (1998). Slice shear force was determined as described by Shackelford et al. (1999) on a TMS-PRO Texture Measurement System (Food Technology Corp, Sterling, VA, USA). Peak force recorded as described by Shackelford et al. (1999).

Calculation of myoglobin forms from VIS-NIR spectra

The spectra obtained from the VIS-NIR measurements were used to obtain reflectance values at 474, 525, 572, 610, and 730 nm. Reflectance at 525 nm is isobestic (equal reflectance) for all 3 myoglobin forms (King et al., 2023). Reflectance at 610 nm is isobestic (equal reflectance) for met- and deoxymyoglobin. Oxymyoglobin and metmyoglobin are isobestic at 474 nm, and deoxymyoglobin and oxymyoglobin are isobestic at 572 nm. Thus, reflectance at 474, 572, and 610 nm can be used in conjunction with reflectance at 525 nm to indicate relative abundance of deoxy-, met-, and oxymyoglobin, respectively, present on the meat surface.

These values were used to calculate the percentage of surface oxy-, met-, and deoxymyoglobin according to the spectral absorbance approach described by (Piao et al., 2025) and the reflectance attenuance method as described by (Krzywicki, 1979). The American Meat Science Association Guidelines for Meat Color Measurement recommends that oxymyoglobin be calculated by difference (i.e., subtract the sum of the values obtained for deoxy- and metmyoglobin from 100) rather than using the oxymyoglobin equations directly (King et al., 2023). The authors of Piao et al. (2025) made similar recommendations for determining oxymyoglobin using their approach and equations. Thus, for the present experiment, oxymyoglobin was determined by difference in both approaches.

Ratios of these reflectance values after K/S transformation can also be used to indicate the relative abundance of myoglobin forms (Francis and Clydesdale, 1975). The percentage of the myoglobin forms can also be calculated using equations described by King et al. (2023). However, these calculations require the creation of 100% myoglobin forms, which are technically difficult to create and must be created using samples that have been exposed to the same conditions as experimental samples. Such standards were not feasible for this experiment. Thus, the ratios of the K/S transformed reflectance values were used in the present experiment.

Statistical analysis

Estimates of oxy-, deoxy-, and metmyoglobin calculated using the Piao and Krzywicki methods, and K/S transformed reflectance ratios were compared across bloomed status and plant using the lm() function of R (R Development Core Team, 2025). The model included fixed effects of bloom status, plant, and their interaction. Carcass was included as a random effect. Additionally, density plots were constructed to evaluate the distribution of each myoglobin form in each bloom status relative to the mean using the geom_density() option within the ggplot() function (Wickham, 2009). Density plots were constructed in a manner that the density estimates were separate for each plant × bloomed status combination so that the variability of the estimates of myoglobin forms of bloomed and unbloomed muscles could be compared. Pearson correlation coefficients were generated separately for estimates obtained from bloomed and unbloomed ribeye muscles using the rcorr() function in the Hmisc package (Harrel et al., 2020)

Data generated from simulated retail displays were decomposed via principal component analysis using the PCA function of the FactoMineR package (Le et al., 2008). The first principal component was strongly related to the initial color of steaks in a simulated retail display. The second principal component was strongly related to loss of redness (i.e., lability) during simulated retail display. Because higher values were indicative of less stable (more labile) lean color, this value is referred to throughout the results and discussion as color lability.

Linear regression was used to test the relationship between oxymyoglobin of bloomed muscles to the observed values for principal component 1 (initial color), principal component 2 (color lability), and slice shear force. The model included bloomed oxymyoglobin, plant, and their interaction. When the interaction term was deemed significant (P < 0.05), beta coefficients were generated for each plant, and the slopes and intercepts were compared with contrasts. A second regression including only oxymyoglobin was run to compare the R-squared values between the single and two-plant models. A level of P < 0.05 was used for all judgements of statistical significance.

Estimation of myoglobin forms of unbloomed and bloomed longissimus thoracis

The distributions of oxymyoglobin estimates given by each of the approaches to calculation on reflectance spectra obtained from the longissimus thoracis muscle exposed by ribbing the carcass before grading (ribeye) are depicted in Figure 1. The estimates provided by all 3 methods differentiated between bloomed and unbloomed ribeyes, although the ratio of K/S transformed reflectance at 610 and 525 nm presented substantial overlap between bloomed and unbloomed ribeyes. Moreover, all of the calculation approaches detected differences between unbloomed data from the 2 plants at which data were collected. The Krzywicki and Piao methods differentiated data from the 2 plants with regard to oxymyoglobin content of bloomed longissimus thoracis. However, ratios of K/S transformed reflectance did not differentiate oxymyoglobin content across the 2 plants.

Figure 1.

Least-squares means (vertical lines) and density plots of estimates of oxymyoglobin calculated from spectra collected on unbloomed and bloomed longissimus thoracis exposed by ribbing at the 12^th-13^th rib interface (ribeye) in 2 plants using 3 methods. Panel A depicts oxymyoglobin calculated using the Piao method. Panel B depicts oxymyoglobin calculated with the Krzywicki method. Panel C depicts K/S₆₁₀/K/S₅₂₅ ratios. Due to a lack of correlation between K/S ratios and the Piao and Krzwicki estimates, the range of values in Panel C does not correspond to the ranges of Panels A and B.

^a–dLeast-squares means, within a panel, lacking common superscripts differ (P < 0.05).

The spectral absorption algorithm approach described by Piao et al. (2025) resulted in estimates of oxymyoglobin that were lower with respect to unbloomed measurements compared to estimates by the reflex attenuance method reported by Krzywicki (1979). Moreover, estimates of oxymyoglobin of bloomed samples via the Piao method were higher than those obtained by the Krzywicki method. Moreover, the distribution of estimates provided by the Piao method was wider than the distribution of estimates from the other methods. This indicates that the estimates obtained from the Piao approach will provide greater discrimination among samples. However, the Piao method resulted in a significant number of samples with estimates of the proportion of myoglobin in the oxymyoglobin form greater than 100 percent or lower than 0 percent for bloomed and unbloomed ribeye muscles, respectively.

Figure 2 depicts the distribution and mean values for surface metmyoglobin estimates of bloomed and unbloomed ribeye muscles in 2 different plants. The difference (P < 0.05) between bloomed and unbloomed ribeyes was much greater in carcasses from Plant B than in carcasses from Plant A. In carcasses from Plant A, surface metmyoglobin did not differ (P > 0.05) between unbloomed and bloomed longissimus thoracis muscles. In carcasses from Plant B, bloomed ribeye muscles exhibited greater (P < 0.05) surface metmyoglobin than unbloomed ribeye muscles.

Figure 2.

Least-squares means (vertical lines) and density plots of estimates of metmyoglobin calculated from spectra collected on unbloomed and bloomed longissimus thoracis exposed by ribbing at the 12^th-13^th rib interface (ribeye) in 2 plants using 3 methods. Panel A depicts metmyoglobin calculated using the Piao method. Panel B depicts metmyoglobin calculated with the Krzywicki method. Panel C depicts K/S₅₇₂/K/S₅₂₅ ratios. The range in Panel C corresponds to the ranges in Panels A and B.

^a–dLeast-squares means, within a panel, lacking common superscripts differ (P < 0.05).

When estimated by the Piao method, the distribution of metmyoglobin values was centered around zero, and a wide distribution was observed with metmyoglobin estimates ranging from −20 to 20%. The Krzywicki method for metmyoglobin percentage produced estimates ranging from 10 to 30 percent surface metmyoglobin with mean values ranging from 19.0 to 20.6 percent metmyoglobin.

The distribution and mean values of surface deoxymyoglobin estimated by these equations are shown in Figure 3. All 3 calculation approaches detected (P < 0.05) differences between bloomed and unbloomed muscles, as well as differences between plants at both stages of oxygenation. For all 3 calculation methods, unbloomed samples from Plant A had lower (P < 0.05) deoxymyoglobin levels than unbloomed samples from Plant B. However, Plant A’s estimates of deoxymyglobin in bloomed longissimus thoracis muscles were greater (P < 0.05) than bloomed samples from Plant B.

Figure 3.

Least-squares means (vertical lines) and density plots of estimates of deoxymyoglobin calculated from spectra collected on unbloomed and bloomed longissimus thoracis exposed by ribbing at the 12^th-13^th rib interface (ribeye) in 2 plants using 3 methods. Panel A depicts deoxymyoglobin calculated using the Piao method. Panel B depicts deoxymyoglobin calculated with the Krzywicki method. Panel C depicts K/S₄₇₄/K/S₅₂₅ ratios. The range in Panel C corresponds to the ranges in Panels A and B.

^a–dLeast-squares means, within a panel, lacking common superscripts differ (P < 0.05).

The magnitude of the deoxymyoglobin estimates calculated using the Piao method for unbloomed ribeye muscles was much larger than those calculated using the Krzywicki method. Hernández et al. (2015) compared estimates of the myoglobin forms by the Krzywicki (1979) reflectance attenuance approach and the K/S transformation approach suggested by Francis and Clydesdale (1975) and noted that the methods resulted in markedly different estimates for the 3 myoglobin forms. Ledward (1970) indicated that estimates of myoglobin forms using the K/S ratio method were accurate to within 6 or 7%. However, that report also indicated that factors such as total myoglobin concentration, pH, and intramuscular fat could affect the reflectance spectra, and the effects of those factors were not corrected for by the K/S transformation of the reflectance values. King et al. (2023) indicated that separate standards should be created for treatments differing in reflectance properties. Ramanathan et al. (2010) demonstrated the impact that enhancement treatments can have on the K/S ratios of 100 percent standards for myoglobin redox forms. It is likely that the other methods may be affected by factors other than myoglobin concentration.

Pearson correlation coefficients of estimates of the 3 myoglobin forms produced using the 3 different algorithms are presented in Table 1. Correlation coefficients were produced separately for estimates generated for unbloomed and bloomed ribeye muscles. The Krzywicki and Piao methods produced estimates of oxymyoglobin that were highly correlated (P < 0.05). Similarly, these 2 approaches produced estimates of metmyoglobin and deoxymyoglobin that were also highly correlated (P < 0.05). The Piao method estimates of metmyoglobin and deoxymyoglobin in unbloomed ribeye muscles were very highly (P < 0.05) correlated to K/S ratios indicative of those myoglobin forms. Similarly, results of the Krzywicki method for deoxymyoglobin and metmyoglobin were also very strongly related to K/S ratios. However, neither approach produced oxymyoglobin estimates that were correlated to K/S ratios for oxymyoglobin for unbloomed muscles. However, all 3 methods produced estimates of myoglobin forms in bloomed muscles that were highly correlated (P < 0.05) with estimates for the same myoglobin form resulting from the other methods. Hernández et al. (2015) reported a high degree of correlation between the reflectance attenuance (Krzywicki, 1979) and K/S (Francis and Clydesdale, 1975) approaches to estimating myoglobin forms despite differences in the absolute values of the estimates. Those authors concluded that statistical analyses would yield similar results regarding relative differences across treatments.

In agreement with those results, the high degree of correlation among the 3 approaches to calculating myoglobin forms in bloomed muscles indicated that all 3 approaches would yield substantially the same relative differences in samples that have been allowed to oxygenate. However, the lack of correlation between K/S₆₁₀/K/S₅₂₅ of unbloomed muscles to estimates of oxymyoglobin obtained from either the Piao or Krzywicki methods is concerning. Moreover, the extent to which the distributions of K/S₆₁₀/K/S₅₂₅ of bloomed and unbloomed muscles overlapped is troubling. One of the most common uses for K/S ratios or estimates of oxymyoglobin calculated from K/S ratios is the measurement of oxygen consumption rate as described by the AMSA Guidelines for Meat Color Measurement (King et al., 2023). That method uses the conversion of oxymyoglobin to deoxymyoglobin under vacuum in controlled temperature conditions. The results of the present experiment suggest that using K/S₄₇₄/K/S₅₂₅ to estimate deoxymyoglobin might better reflect that conversion. Alternatively, using either of the Piao (2025) or Krzywicki (1979) approaches could offer greater differentiation of the degree of oxygenated and deoxygenated samples in that assay.

The magnitude and distribution of values for myoglobin forms differed considerably between the Piao et al. (2025) and Krzywicki (1979) approaches. Both methods differentiated between unbloomed and bloomed ribeye muscles as well as plant differences. However, estimates of deoxymyoglobin from the Piao were higher than estimates of deoxymyoglobin from the Krzywicki method. This is particularly true for unbloomed muscles, which would be expected to have myoglobin predominantly in the deoxymyoglobin form. Both methods produced a number of estimates for the percentage of deoxymyoglobin in bloomed muscles that were less than zero. Additionally, the Piao method produced a number of estimates for the proportion of myoglobin in the deoxygenated form present on the surface of unbloomed ribeye muscles greater than 100 percent.

Estimates for metmyoglobin present on the surface of the ribeye were also quite different between the Piao et al. (2025) and Krzywicki (1979) equations. The mean values for metmyoglobin determined via the Piao method were close to zero. However, a substantial number of these estimates were well below zero, which is troubling. Conversely, the mean for surface metmyoglobin produced by the Krzywicki equations was near 20 percent. Whereas 20 percent surface metmyoglobin (as determined by the K/S method) has been reported as a point where consumers start to discriminate against discoloring products (Lybarger et al., 2023), these estimates are high for freshly cut, early postmortem muscles.

Both the Piao and Krzywicki methods produced estimates of one or more of the myoglobin forms outside the plausible range of 0 to 100. By this criterion, the Krzywicki method appears to outperform the Piao method with a comparatively few number of bloomed ribeyes having deoxymyoglobin estimates below 0. However, as noted, the Krzywicki estimates for metmyoglobin are higher than expected for early postmortem, freshly cut muscles. Thus, oxymyoglobin values by the Krzywicki method are likely under-estimated. Cooper et al. (2022) used the Krzywicki equations on longissimus lumborum steaks that were either 48 h or 14 d postmortem, which had been allowed to bloom for 20 min. Those investigators reported metmyoglobin estimates that were higher (34 and 40% for 48 h and 14 d, respectively) than those observed in the present study. Those authors also reported relatively low oxymyoglobin estimates (46 and 49% for 48 h and 14 d, respectively) in those freshly cut steak surfaces that were only slightly higher than the metmyoglobin estimates.

Current recommendations for implementing the Krzywicki and Piao equations prescribe determining oxymyoglobin by difference rather than using direct equations for oxymyoglobin (King et al., 2023; Piao et al., 2025). Thus, the accuracy of oxymyoglobin estimates is dependent on the accuracy of the estimates for both deoxymyoglobin and metmyoglobin. As a result, any error associated with individual myoglobin form estimation is compounded when estimating oxymyoglobin. In the present experiment, metmyoglobin estimates by the Piao method were centered around zero percent metmyoglobin, but a substantial number of estimates were below zero, resulting in oxymyoglobin estimates greater than 100 percent.

These equations use the reflectance at 730 nm (700 with most colorimeters) to correct for non-myoglobin effects on reflectance. The authors of Piao et al. (2025) investigated the sensitivity of both the Krzywicki approach and the Piao approach to the absorbance at 700 nm, with the assumption that absorbance at 700 nm was greater than the expected absorbance at 730 nm. The instrument used in the present experiment collected absorbance at both 700 and 730 nm. Thus, we had the opportunity to compare reflectance at these wavelengths. Using linear regression to predict absorbance at 730 nm (abs₇₃₀) with absorbance at 700 nm (abs₇₀₀) resulted in the equation: Abs₇₃₀ = 0.95 Abs₇₀₀ + 0.03 (R² = 0.99). These results indicate that using reflectance at 700 nm for the correction factor is appropriate.

These equations are predicated on the correction of diffuse reflectance data for the scattering and absorptive properties of muscle. This includes the assumption that the scattering and absorption properties and subsequent correction are equal across all of the relevant wavelengths. Moreover, these equations assume that the absorbance and scattering properties are equivalent regardless of structural and biochemical properties known to vary across muscles. Ledward (1985) reported that K/S ratios could not correct the effects of factors such as muscle pH, intramuscular fat, and total myoglobin concentration. In the present experiment, all of these factors (and likely others of importance) varied across the muscles evaluated. The American Meat Science Guidelines for Meat Color Measurement (King et al., 2023) recommends creating separate standards to account for these factors. However, the Krzywicki and Piao methods do not provide a simple method to address such variation. The development of methods to address accounting for these factors warrants further research effort.

Additionally, other practical considerations likely added to the variation observed in these factors. All of the spectra collected in the present experiment were collected on the production line at normal production speeds. Thus, carcasses were moving during measurement. The operators were highly experienced in collecting VIS/NIR spectra under these conditions, but error could have been introduced by movement during measurement. Additionally, the carcasses assessed in the present experiment were ribbed according to normal plant operations. Our experience in instrument grading research indicates that significant variation exists in the workmanship associated with ribbing beef carcasses, which affects instrumental measurements made on the ribeye surface. No effort was made to avoid ribbing workmanship defects in the present experiment. Including such normal sources of error was important in an assessment of the utility of these measurements in predicting meat quality outcomes during typical production. However, interpretation of absolute values of instrumental measures requires recognizing the existence of such sources of variation.

For bloomed ribeye muscles, estimates of oxymyoglobin from all 3 methods were highly correlated with the oxymyoglobin estimates from the other methods. The same was true for metmyoglobin estimates. Thus, we felt that all of the methods could be used to evaluate the relationship between oxymyoglobin present in bloomed ribeye muscles to lean color, lean color stability, and slice shear force of longissimus thoracis steaks. Despite the number of estimates greater than 100% for oxymyoglobin, we conducted this evaluation using estimates from the Piao method because of the greater range and distribution of estimates from this method relative to the Krzywicki method. For these analyses, we used the estimate provided by the equation even if it fell outside the range of 0 to 100. Additionally, because such a large range of metmyoglobin estimates was observed in bloomed ribeye muscles in the present experiment, we evaluated the relationship between the proportion of the metmyoglobin present in the bloomed ribeye muscles to the meat quality end points.

Principal component analysis to reduce color stability data

To simplify the interpretation of the relationships between oxymyoglobin forms and lean color and color stability, principal component analysis was used to decompose the color stability data into components explaining variation in all of the traits. The first principal component, which explained 45.3% of the variation in color traits during simulated retail display, was highly related to color, particularly early in the display period. Positive values for Component 1 were related to higher values for L*, a*, b*, and chroma, particularly on d 0, 1, and 4 of the simulated retail display. Component 2 explained 18.2% of the variation in the simulated retail display data. Positive values for Component 2 were related to higher values for hue angle on days 4, 7, and 11 of the simulated retail display. Negative values for Component 2 were associated with higher values for a* and Chroma on day 11 of the simulated retail display. Thus, Component 2 is indicative of labile lean color, particularly with losses in redness were indicated by higher values for Component 2.

The individual score plot of the first 2 components of the principal component analysis indicated that the 2 plants sampled for this experiment differed regarding the 2 components describing color and color stability of the samples included in the present experiment. For Component 1, samples from Plant B exhibited greater variability, particularly in negative values for Component 1. Carcasses from Plant B generally had greater values for the principal component, indicating color lability. Thus, steaks from carcasses from Plant B exhibited less stable lean color than steaks from carcasses from Plant A.

Oxymyoglobin at grading as a predictor of meat quality

Plant interacted (P < 0.001) with surface oxymyoglobin to affect initial color (Figure 5A). As the percentage of oxymyoglobin on the ribeye surface increased at grading, the values for initial color increased. This indicates that increased bloom (and lesser oxygen consumption) resulted in higher values for instrumental color measures, particularly early in simulated retail display. Both the slope and intercept of the regression equations using oxymyoglobin in the ribeye to predict initial color differed (P < 0.05) across plants. The regression line for Plant B had a lower (P < 0.05) intercept but also had a greater slope (P < 0.05).

Figure 4.

Results of principal component analysis of instrumental color data collected on longissimus thoracis steaks during simulated retail display. Panel A depicts variable loadings of instrumental color variables collected during a simulated retail display. Panel B depicts individual loadings for longissimus lumborum steaks stratified by processing plant.

Figure 5.

Relationships between oxymyoglobin in the ribeye at grading to meat quality traits. Panel A depicts the regression equations for the relationship between oxymyoglobin in the ribeye at grading and the principal component explaining initial color in 2 plants. Plant effect slope P = 0.02, Plant effect intercept P < 0.01. Panel B depicts the relationship between oxymyoglobin in the ribeye at grading and the principal component explaining variation in color lability in 2 plants. Plant effect slope P = 0.06, Plant effect intercept P < 0.05. Panel C depicts the relationship between oxymyoglobin in the ribeye at grading and slice shear force int 2 plants. Plant effect slope P < 0.001, Plant effect intercept P < 0.001.

Plant also impacted the relationship of oxymyoglobin on the surface of the ribeye at grading to the stability of lean color of longissimus thoracis steaks in simulated retail display (Figure 5B). Increased oxymyoglobin in the ribeye surface at grading was associated with greater Component 2 values. The slope of the regression equations indicated that increases in oxymyoglobin had a similar (P = 0.06) impact on color stability across the 2 plants. However, the intercepts of the regression equation indicated steaks from carcasses sampled in Plant B had greater (P < 0.05) values for the principal component, indicating greater loss of redness and color change during simulated retail display. The model, including oxymyoglobin in the ribeye at grading, plant, and their interaction, explained 41% of the variation in color lability. However, a model with only oxymyoglobin in the model explained 3% of the variation in color lability (not shown). Thus, the plant effect explained much more variation in color lability than the amount of oxymyoglobin present in the ribeye at grading.

Increased bloom in the ribeye at grading was associated with lower slice shear force values in aged longissimus thoracis steaks, although this relationship differed across plant (P < 0.001). At low levels of bloom, steaks from carcasses sampled from Plant B had higher (P < 0.001) slice shear force values than steaks from carcasses sampled in Plant A. However, the decrease in slice shear force associated with increased bloom was greater (P < 0.001) in carcasses from Plant B than in carcasses from Plant A. Thus, at higher levels of bloom, the slice shear force difference across plants was minimal.

Of the meat quality end points evaluated, the trait most affected by the proportion of oxymyoglobin was initial color, followed by slice shear force. Color stability was influenced by oxymyoglobin in the ribeye at grading, but oxymyoglobin explained little of the variation in color stability. Similarly, Calnan et al. (2019) found that bloom, as measured by the ratio of reflectance at 630 to 589 nm, was not predictive of color stability in retail lamb packages.

Mitochondrial oxygen consumption competes with myoglobin for available oxygen. Thus, the extent of oxygenation of myoglobin at grading is inversely related to mitochondrial oxygen consumption. In agreement with the present experiment, previous results from our laboratory indicated that longissimus thoracis steaks classified as having unstable lean color had greater oxygen consumption, and higher initial values for instrumental color measurements in addition to greater color change during simulated retail display (King et al., 2011). Moreover, oxygen consumption has been reported to be negatively related to tenderness (Cooper et al., 2025).

Instrumental measures of beef longissimus color, specifically at grading, have previously been reported to be associated with tenderness (Cooper et al., 2022; Wulf et al., 1997; Wulf and Page, 2000). Cooper et al. (2022) found increased surface metmyoglobin in longissimus steaks bloomed for 20 min at 48 h postmortem to be related to increased abundance of intact calpain-1 and suggested oxidative conditions might have inhibited autolysis and consequent activation of calpain-1.

We speculate that the relationship of oxygen consumption as measured by bloom to tenderness is indicative of the impact of mitochondrial function on tenderness rather than the direct impact of oxygen consumption on tenderness. We have used oxygen consumption and nitric oxide metmyoglobin reducing ability as indicators of mitochondrial functionality to explain variation in the oxidative capacity of muscle (King et al., 2024; King et al., 2025a). Mitochondrial activity contributes to variation in both tenderness and color stability (Cooper et al., 2025; Mancini and Ramanathan, 2014; Ramanathan et al., 2021). Ramos et al. (2020) found that increased Brahman inheritance was associated with decreased rate of pH decline and delayed autolysis of calpain-1 and calpastatin, and, consequently, less proteolytic tenderization of longissimus muscles. Those authors suggested that greater mitochondrial function delayed pH decline, which delayed calcium release by the mitochondria and possibly sarcoplasmic reticulum in Brahman-influenced muscles, which delayed autolysis. In support of this finding, Wright et al. (2018) detected increased citrate synthase activity, indicating increased mitochondrial function with increased Brahman inheritance. Ramos et al. (2025) reported increased mitochondrial enzyme abundance in association with decreased tenderization in muscles from cattle with excitable temperaments. As noted, we have used bloom and initial metmyoglobin formation as indicators of mitochondrial function and found that both traits were involved in complex interactions involving glycolytic and mitochondrial metabolism, oxidative damage, and chaperone proteins affecting tenderness of beef longissimus and gluteus medius muscles (King et al., 2025b)

The extent of oxymyoglobin at grading might be used to provide insight into variation in mitochondrial function, where direct measurements of oxygen consumption or mitochondrial properties are cost and labor-prohibitive. For example, genomic studies targeting meat quality traits required large contemporary groups of animals to be phenotyped at once, and measuring oxygen consumption on the whole group is often not feasible.

Metmyoglobin at grading as a predictor of meat quality

The variation in the estimated metmyoglobin present in the bloomed ribeye observed for all of the methods suggested that substantial variation existed in the oxidative status of the ribeye muscle early postmortem. We tested the extent to which this variation was indicative of differences in lean color, color stability, and tenderness. The relationship between color, color lability, and slice shear force to metmyoglobin present in the bloomed ribeye as it was presented for grading is presented in Figure 6.

Figure 6.

Relationships between metmyoglobin in the ribeye at grading to meat quality traits. Panel A depicts the regression equations for the relationship between metmyoglobin in the ribeye at grading and the principal component explaining initial color in 2 plants. Plant effect slope P = 0.06. Plant effect intercept P < 0.001. Panel B depicts the relationship between metmyoglobin in the ribeye at grading and the principal component explaining variation in color lability in 2 plants. Plant effect slope P = 0.03. Plant effect intercept P < 0.001. Panel C depicts the relationship between metmyoglobin in the ribeye at grading and slice shear force int 2 plants. Plant effect slope P = 0.96. Plant effect intercept P < 0.001.

Increases in metmyoglobin in the bloomed ribeye resulted in decreased (P = 0.002) values for the principal component related to initial color (Figure 6A). This indicates that carcasses with ribeyes exhibiting less metmyoglobin at grading also had greater values for instrumental color measurements during simulated retail display, particularly early in the display period. The slope of the regression equations depicting the relationship between metmyoglobin in the bloomed ribeye surface differed across plants, as did the intercepts of those equations. The intercept was higher (P < 0.001; indicating greater color values at low metmyoglobin levels) for steaks from carcasses sampled in Plant B than for steaks sampled from Plant A. However, the slope was more negative (P = 0.03) in samples from Plant B, indicating that increases in metmyoglobin in the bloomed ribeye had a greater impact on lean color early in the display period. Despite statistical differences in the coefficients of the regression equations due to plant, the actual differences in the regression lines were quite small. The regression model that included plant as a fixed effect explained 50 percent of the variation in initial lean color. A simpler model that omitted the fixed effect of the plant also explained 50 percent of the variation in initial lean color. Thus, the relationship of metmyoglobin in the bloomed ribeye at grading to lean color early in the display period was similar in the samples from the 2 plants.

Increases in metmyoglobin were associated with slight decreases (P < 0.05) in the principal component related to the lability of lean color during simulated retail display. The slope for Plant B was greater (P = 0.03) than the slope for Plant A. The higher (P < 0.001) intercept in the equation for Plant B indicated that steaks from carcasses sampled in Plant B had less stable lean color. The model fitting separate equations for the relationship between metmyoglobin in the ribeye at grading to color stability of longissimus thoracis steaks explained 44 percent of the variation in the principal component related to loss of redness (R² = 0.44; Figure 6B). However, a model that omitted the plant had a coefficient of regression (R²) of 0.01, which suggests most of the variation explained in the two-plant model was due to the plant effect.

The relationship between metmyoglobin present in the ribeye surface at grading and slice shear force is depicted in Figure 6C. The slope for this relationship was not different (P = 0.16) from zero and did not differ (P = 0.96) across plants. The intercept was higher (P <0.001) for the equation for Plant B than for the equation for Plant A, indicating that longissimus thoracis steaks from Plant B had greater slice shear force values than longissimus thoracis steaks from Plant A. These data indicate that surface metmyoglobin of the ribeye at grading explained little of the variation in meat quality attributes