Despite efforts by the industry to control the eating quality of beef, there remains a high level of variability in palatability, which is one reason for consumer dissatisfaction. In Europe, there is still no reliable on-line tool to predict beef quality and deliver consistent quality beef to consumers. Beef quality traits depend in part on the physical and chemical properties of the muscles. The determination of these properties (known as muscle profiling) will allow for more informed decisions to be made in the selection of individual muscles for the production of value-added products. Therefore, scientists and professional partners of the ProSafeBeef project have brought together all the data they have accumulated over 20 years. The resulting BIF-Beef (Integrated and Functional Biology of Beef) data warehouse contains available data of animal growth, carcass composition, muscle tissue characteristics and beef quality traits. This database is useful to determine the most important muscle characteristics associated with a high tenderness, a high flavour or generally a high quality. Another more consumer driven modelling tool was developed in Australia: the Meat Standards Australia (MSA) grading scheme that predicts beef quality for each individual muscle x specific cooking method combination using various information on the corresponding animals and post-slaughter processing factors. This system has also the potential to detect variability in quality within muscles. The MSA system proved to be effective in predicting beef palatability not only in A C C E P T E D M A N U S C R I P T ACCEPTED MANUSCRIPTAustralia but also in many other countries. The results of the work conducted in Europe within the ProSafeBeef project indicate that it would be possible to manage a grading system in Europe similar to the MSA system. The combination of the different modelling approaches (namely muscle biochemistry and a MSA-like meat grading system adapted to the European market) is a promising area of research to improve the prediction of beef quality. In both approaches, the volume of data available not only provides statistically sound correlations between various factors and beef quality traits but also a better understanding of the variability of beef quality according to various criteria (breed, age, sex, pH, marbling etc).

An experiment was set up for (i) comparing Australian and French consumer preferences to beef and to (ii) quantify how well the Meat Standards Australia (MSA) grading model could predict the eating quality of beef in France. Six muscles from 18 Australian and 18 French cattle were tested as paired samples. In France, steaks were grilled ‘medium’ or ‘rare’, whereas in Australia ‘medium’ cooking was used. In total, 360 French consumers took part in the ‘medium’ cooking test, with each eating half Australian beef and half French beef and 180 French consumers tested the ‘rare’ beef. Consumers scored steaks for tenderness (tn), juiciness (ju), flavour liking (fl) and overall liking (ov). They also assigned a quality rating to each sample: ‘unsatisfactory’, ‘satisfactory everyday quality’ (3*), ‘better than everyday quality’ (4*) or ‘premium quality’ (5*). The prediction of the final ratings (3*, 4*, 5*) by the French consumers using the MSA-weighted eating quality score (0.3 tn + 0.1 ju + 0.3 fl + 0.3 ov) was over 70%, which is at least similar to the Australian experience. The boundaries between ‘unsatisfactory’, 3*, 4* and 5* were found to be ca. 38, 61 and 80, respectively. The differences between extreme classes are therefore slightly more important in France than in Australia. On average, even though it does not have predictive equations for bull meat, the mean predicted scores calculated by the MSA model deviated from observed values by a maximum of 5 points on a 0 to 100 scale except for the Australian oyster blade and the French topside, rump and outside (deviating by <15). Overall, the data indicate that it would be possible to manage a grading system in France as there is high agreement and consistency across consumers. The ‘rare’ and ‘medium’ results are also very similar, indicating that a common set of weightings and cut-offs can be employed.

The beef industry must become more responsive to the changing market place and consumer demands. An essential part of this is quantifying a consumer's perception of the eating quality of beef and their willingness to pay for that quality, across a broad range of demographics. Over 19 000 consumers from Northern Ireland, Poland, Ireland and France each tasted seven beef samples and scored them for tenderness, juiciness, flavour liking and overall liking. These scores were weighted and combined to create a fifth score, termed the Meat Quality 4 score (MQ4) (0.3 × tenderness, 0.1 × juiciness, 0.3 × flavour liking and 0.3 × overall liking). They also allocated the beef samples into one of four quality grades that best described the sample; unsatisfactory, good-every-day, better-than-every-day or premium. After the completion of the tasting panel, consumers were then asked to detail, in their own currency, their willingness to pay for these four categories which was subsequently converted to a proportion relative to the goodevery-day category (P-WTP). Consumers also answered a short demographic questionnaire. The four sensory scores, the MQ4 score and the P-WTP were analysed separately, as dependant variables in linear mixed effects models. The answers from the demographic questionnaire were included in the model as fixed effects. Overall, there were only small differences in consumer scores and P-WTP between demographic groups. Consumers who preferred their beef cooked medium or well-done scored beef higher, except in Poland, where the opposite trend was found. This may be because Polish consumers were more likely to prefer their beef cooked well-done, but samples were cooked medium for this group. There was a small positive relationship with the importance of beef in the diet, increasing sensory scores by about 4% in Poland and Northern Ireland. Men also scored beef about 2% higher than women for most sensory scores in most countries. In most countries, consumers were willing to pay between 150 and 200% more for premium beef, and there was a 50% penalty in value for unsatisfactory beef. After quality grade, by far the greatest influence on P-WTP was country of origin. Consumer age also had a small negative relationship with P-WTP. The results indicate that a single quality score could reliably describe the eating quality experienced by all consumers. In addition, if reliable quality information is delivered to consumers they will pay more for better quality beef, which would add value to the beef industry and encourage improvements in quality.Keywords: consumer testing, beef, quality, demographics, Europe Implications A single quality descriptor of beef eating quality will likely be applicable to the entire European market due to the small impact of demographics on consumer scores. This descriptor could form the basis of an eating quality-based grading system for beef which would allow consumers to select beef of a desired quality when purchasing beef. As European consumers are also willing to pay more for better q...

This review is aimed at providing an overview of recent advances made in the field of meat quality prediction, particularly in Europe. The different methods used in research labs or by the production sectors for the development of equations and tools based on different types of biological (genomic or phenotypic) or physical (spectroscopy) markers are discussed. Through the various examples, it appears that although biological markers have been identified, quality parameters go through a complex determinism process. This makes the development of generic molecular tests even more difficult. However, in recent years, progress in the development of predictive tools has benefited from technological breakthroughs in genomics, proteomics, and metabolomics. Concerning spectroscopy, the most significant progress was achieved using near-infrared spectroscopy (NIRS) to predict the composition and nutritional value of meats. However, predicting the functional properties of meats using this method—mainly, the sensorial quality—is more difficult. Finally, the example of the MSA (Meat Standards Australia) phenotypic model, which predicts the eating quality of beef based on a combination of upstream and downstream data, is described. Its benefit for the beef industry has been extensively demonstrated in Australia, and its generic performance has already been proven in several countries.

No abstract