Background

Rapid Evaporative Ionization Mass Spectrometry (REIMS) is a relatively new technology that is emerging in many areas of science, including human medicine and biological sciences. REIMS‐based tissue analysis generally takes only a few seconds and can provide histological tissue identification with 90−98% correct classification performance. Recently, utilization of REIMS in meat products has provided very promising results across various classification scenarios. Using time‐of‐flight (TOF) mass spectrometry, REIMS profiling provides in situ, real‐time molecularly‐resolved information by ionizing biological samples in real‐time without any sample preparation. Waters Corporation (Wilmslow, UK) has developed this technology coupled to a hand‐held iKnife sampling device, allowing for tremendous mobility in the sampling procedure. For the first time, this technology would allow for meat quality attributes, such as flavor profile and tenderness, to be predicted and characterized in real‐time via broad molecular profiling of tissue samples. Unlike other molecular approaches that require tedious sample preparation and analysis times, this technology could be further developed as an on‐line system in the processing environment to enable meaningful sorting of beef products into categories reflecting tangible differences in eating characteristics.  

The objective of this study was to evaluate the ability of rapid evaporative ionization mass spectrometry as a novel method to predict various components of beef quality including: carcass type, sensory attributes, and objective tenderness measurements.

Methodology

Striploin sections (N = 292) from 7 beef carcass types (Select, Low Choice, Top Choice, Prime, Dark Cutter, Grass‐ fed, and Wagyu) were collected to achieve variation in fat content, sensory attributes, tenderness, and production background. A 5 cm section was collected from each side of the carcass, aged for 14 d, fabricated into 2.54 cm thick steaks, and frozen until analysis. A trained descriptive panel was used to determine tenderness, flavor, and juiciness attributes for sensory prediction models. Slice shear force (SSF) and Warner‐Bratzler shear force (WBS) values were collected to predict tenderness classifications. A molecular fingerprint of each sample was collected via REIMS to build prediction models. Prediction models were built to predict carcass type, objective tenderness, and sensory attributes. Models were built using 80% of samples and tested for prediction accuracy using the remaining 20%. Partial least squares (PLS) discriminant analysis was used as a dimension reduction technique before building a linear discriminant analysis (LDA) model for classification.

Findings

The results of this study show that collecting molecular profile data from beef strip steaks using REIMS is a rapid approach to predicting meaningful beef carcass characteristics. When carcass types were combined to create the 5 classes of Select/Low Choice, Top Choice/Prime, Dark Cutter, Grass‐fed and Wagyu, carcass type could be predicted with 83.93% accuracy (Table 1). Figure 1 shows a plot of the first 3 LDA discriminants, clearly presenting unique components between each carcass type. Although USDA graders and grading cameras are able to visually determine marbling score and quality grade, they do not have the ability to visually assess differences in specific breed type or production background. In the current study, even though Grass‐fed qualified for USDA Choice and Wagyu carcasses to USDA Prime, REIMS was able to very clearly differentiate these samples based on their unique production history and distinct breed type, respectively.

Principal component analysis (PCA) was used to calculate an overall sensory, as well as, a combined flavor score using trained sensory panel data. From each PCA, the top 1/3 of samples were designated as positive and the bottom 2/3 as negative. Although samples were categorized as positive or negative independent of carcass type, both the overall sensory and combined flavor prediction models provided 80.70% prediction accuracy (Table 2). This would allow REIMS the ability to apply further premiums to highly palatable samples or exclude undesirable samples, even within the same USDA quality grade. Finally, samples were correctly categorized into tender (SSF < 15.4 kg) and tough 75.44% of the time based on predetermined SSF values (Table 3). Whereas, samples categorized as tender (WBS < 3.9 kg) or tough by WBS were correctly identified with 71.93% accuracy (Table 4). Similar to sensory prediction models, the ability to accurately predict tender or tough samples beyond what is predicted by quality grade would provide the opportunity to add further value to beef products.

Implications

The results of this study show potential in REIMS to be used to predict various beef quality attributes. Due to its lack of a sample preparation step coupled with its rapid analysis time, REIMS provides the opportunity to evaluate molecular components at chain speeds to be able to incorporate this information into the grading process. Although it may never replace the current USDA grading system, it could complement this system by selecting samples with high and low palatability ratings from within a grade. Also, as shown by its ability to segregate Grass‐fed and Wagyu samples, REIMS could be utilized in branded programs to verify certain production backgrounds. Further investigation into data processing and classification techniques is needed to improve the prediction capabilities of REIMS data to deal with class variation and sample overlap.