Prediction of Intramuscular Fat Content of Beef Longissimus Muscle | UNL Beef

Authors: Kyra L. Elliott, Graduate Student, Animal Science, Lincoln; Willy J. Horne, Bucky L. Gwartney, USDA Agricultural Marketing Service; Jessie C. Morrill, Assitant Professor, Animal Science, Lincoln.

Summary with Implications

In this study, we report the percentage of extractable intramuscular fat that is associated with beef longissimus muscle from a range of USDA marbling scores. The marbling scores were determined by two beef grading camera systems. We also assessed the utility of an open-source imaging software (ImageJ) for objectively quantifying the percentage of intramuscular fat, represented by pixels, in digital images of beef longissimus muscle.

Introduction

Beef grading camera systems are widely utilized to assist with the assignment of USDA yield and quality grades to beef carcasses in commercial packing facilities. Instrument grading systems offer the advantage of rapid, non-invasive analysis; however, questions remain regarding the alignment of grading camera marbling score calls and chemically quantifiable intramuscular fat content of beef muscle.

The most widely used camera grading system has been the e+v VBG 2000, also referred to as the GigE, grading camera (Marel Oranienburg formerly e+v Technology GmbH & Co.KG). However, updated and new instrument grading systems have recently entered the marketplace. Among these are the VBG 7L (Marel Oranienburg), which is an upgraded version of the VBG 2000, that is designed to correct for ribbing and instrument placement errors. Other compact, cost-effective, cell phone camera systems such as the VBG Smart (Marel Oranienburg), MIJ (MIJ Labo Inc.) and MEQ (MEQ Solutions) have also entered the marketplace, making grading more accessible to small- and mid-sized packing facilities. As instrument grading technologies continue to evolve, it will be important to validate their accuracy with objective laboratory methods to ensure reliable carcass evaluation that supports both producer returns and industry integrity.

In 2024, the USDA Agricultural Marketing Service launched a remote grading pilot program, which utilizes a cell phone camera to assist with image capture for USDA grading. There is a timely need to understand how marbling scores assigned by cellphone camera systems, like the MEQ, align with the VBG 2000 grading camera and measures of chemically extractable intramuscular fat.

In this study, our objective was to quantify the amount of chemically extractable intramuscular fat in beef longissimus muscle with varying levels of marbling. An additional objective was to determine if an open-source imaging processing software could be utilized to objectively quantify percentage intramuscular fat in digital images of beef longissimus muscle.

Procedure

Institutional Animal Care and Use Committee approval was not sought for this study as no measurements were taken from live animals.

Sample Procurement and Handling

Untrimmed USDA Choice Beef Ribs, Oven-Prepared (Institutional Meat Purchase Specification 107; n=79) were procured from a commercial beef processing facility (Omaha, NE). The ribs were obtained approximately 36-48 hours after harvest, vacuum packaged, and transported to the Loeffel Meat Laboratory at the University of Nebraska – Lincoln. Upon arrival at the laboratory, beef ribs were stored and handled within a walk-in refrigerator. Beef ribs were intentionally not aged prior to the start of the trial to minimize changes in lean color and texture properties between procurement at the commercial plant and laboratory analyses.
Prior to analysis, ribs were unpackaged and were then refaced by removing approximately 0.5 inch from the posterior end using a band saw. Subsequently, a 1.5-inch-thick steak was cut from the posterior end of each rib and the steak was scraped to be practically free of bone dust and was allowed to bloom for 30 minutes.

Color, Temperature, and Camera Grading of Steaks

After blooming, instrumental color and lean surface temperature were assessed. Color was assessed using a calibrated Konica Minolta colorimeter (CR-400 Chroma Meter; Konica Minolta; Ramsey, NJ) and Commission Internationale de l’Éclairage (CIE) color space values (L*, a*, and b*) were recorded. Color and temperature data were not remarkable and were thus not reported in this short summary. Subsequently, surface temperature of the longissimus thoracis was measured with a laboratory grade infrared thermometer. Immediately following these measurements, the steaks were imaged with the VBG 2000 and the MEQ grading cameras. The personnel collecting the data at subsequent steps were kept blind to all grading camera measurements until data was transferred to the University of Nebraska at the conclusion of the experiment.

ImageJ Analysis

Immediately after grading camera measurements were taken, photographs were taken of steaks in a photo studio light box using a Canon Rebel T3 digital camera. Photographs were saved in .CR2 format for high-quality RAW image data. Prior to analysis .CR2 files were converted to .tif files using Adobe Photoshop. The .tif files were then imported into ImageJ and analyzed to count lean and intramuscular fat pixels in the longissimus muscle.

Sample Powdering

After image capture, steaks were frozen at 24.8°F until powdering. Steaks were removed from the freezer 2 days prior to dissection and powdering and were placed in a refrigerated cooler for thawing. The longissimus muscle was dissected to be practically free of subcutaneous and intermuscular fat, surrounding secondary muscles, and residual connective tissue. The longissimus thoracis muscle was then cubed, flash-frozen in liquid nitrogen, and pulverized using a Waring blender. Powdered samples were transferred to Whirl-Pak bags and stored at -112°F until analysis.

Fat and Moisture Content

For determination of fat content, total lipids of powdered beef samples were extracted using a modified Folch method. The weight of the lipid residue was recorded as the total extracted lipid, and lipid content was expressed as a percentage of wet sample weight. For determination of moisture content, powdered samples were dried at 221°F for 24 h.

Statistical Analysis

Data were analyzed using R (version 4.0) and R Studio (tidyverse version 1.3.0, nlme, version 3.0).

Results

As expected, chemically extractable fat was positively correlated with VBG 2000 and MEQ marbling scores (P < 0.001; r > 0.73), and regression equations had R2 values of 0.66 and 0.53, respectively. The ImageJ model was also highly significant (P < 0.001; r = 0.84) and had an R2 value of 0.7. When the VBG 2000 marbling score is known, the equation for determining percentage extractable fat is equal to: (0.0083 x VBG 2000 numeric marbling score) – 0.43. When MEQ marbling score is known, the equation for determining percentage extractable fat is equal to: (0.011 x MEQ numeric marbling score) – 2.4. Likewise, the equation for determining percentage extractable fat when percentage of intramuscular fat pixels is determined using ImageJ is equal to: (0.55 x ImageJ percent marbling pixels) + 1.4. The MEQ grading camera also estimates percentage intramuscular fat, which was correlated to extractable fat (P < 0.001; r = 0.66), but estimates of percentage intramuscular fat were numerically higher than what was measured chemically. This relationship is reflected in the equation: (0.36 x MEQ percentage intramuscular fat) + 1.6. The R2 value for this equation was 0.53.

Conclusions

Beef grading camera marbling scores and ImageJ estimates of marbling can be used to estimate chemically extractable fat content, but users should be aware that marbling estimates from each system are calculated through different mechanisms, which are often proprietary. In this study, even when the same steaks were imaged, different marbling scores resulted. In this study, the range of marbling scores for the VBG 2000 camera were 281 – 992, whereas the range in marbling scores for the MEQ camera was 383 – 904. When accurate measurements of intramuscular fat content are needed, chemical extraction should still be utilized.

Table 1. Mean chloroform:methanol extractable fat and moisture contents of beef longissimus muscle by marbling level
		CM extractable fat¹		Moisture
Marbling score	N	Mean	SD	Mean	SD
Slight	24	2.87	0.578	73.35	0.674
Small	24	3.74	0.605	72.61	0.599
Modest	10	4.80	0.763	71.63	0.590
Moderate	10	5.78	0.383	70.62	0.807
Slightly Abundant²	10	7.39	0.995	69.34	1.000
¹ CM extractable fat = total lipid, as a percetage of wet tissue weight, extractedin chloroform:methanol (2:1, v/v) ² Equal to or greater than slightly abundant

Figure 1. (A) Correlation between MEQ camera IMF (%) and ImageJ IMF pixels (%). Linear regression shows a moderate positive relationship (R2 = 0.66). (B) Correlation between chemically measured (CM %) extractable lipids and MEQ Camera Intramuscular fat (IMF %). Linear regression shows a moderate positive relationship (R2 = 0.53). (C) Correlation between intramuscular fat (IMF) % estimated by ImageJ analysis and chemically measured (CM) % extractable lipids. Linear regression shows a strong positive relation — **Figure 1.** (A) Correlation between MEQ camera IMF (%) and ImageJ IMF pixels (%). Linear regression shows a moderate positive relationship (R2 = 0.66). (B) Correlation between chemically measured (CM %) extractable lipids and MEQ Camera Intramuscular fat (IMF %). Linear regression shows a moderate positive relationship (R2 = 0.53). (C) Correlation between intramuscular fat (IMF) % estimated by ImageJ analysis and chemically measured (CM) % extractable lipids. Linear regression shows a strong positive relationship (R2 = 0.70).

Acknowledgment

This research was funded by the USDA Agricultural Marketing Service. The authors also wish to express gratitude to Greater Omaha Packing for assisting with and supporting beef collection.