Introduction

Beef carcases in the abattoir are currently valued based on carcase weight, EUROP conformation grid and 1–5 fat class. In order to breed cattle with improved carcase attributes, the pedigree sector select animals based upon Estimated Breeding Values (EBV) for 400 day live weight (a proxy for carcase weight), ultrasound scanned muscle depth (a proxy for conformation) and ultrasound fat depth (a proxy for fat). These traits are recorded on pedigree animals at 400 days of age.

New technologies and methodologies have been developed recently which can be adopted by the beef industry to improve genetic evaluations for carcase traits. Imaging technologies have developed allowing for Video Image Analysis (VIA) traits of carcases from the abattoir to be routinely collected on all animals slaughtered. Recent research has shown that carcase traits measured by VIA in the abattoir can be used for genetic evaluations to identify genetically superior animals for carcase traits (Pabiou et al., 2010; 2011). With carcase traits being measured late in life, in high volume and on commercial animals, they are an ideal subset of traits to benefit from genomic selection. Genomic selection was first proposed in 2001 (Meuwissen et al., 2001) but was not practical until the bovine genome was available in 2006. Since then it has been implemented in the dairy industry across the world, including the UK, but is yet to be widely implemented in beef.

Motivated by the developments in genomic selection methods and the use of VIA to record carcase traits, the British Limousin Cattle Society (BLCS), Anglo Beef Processors (ABP) and Scotland’s Rural College (SRUC) have embarked upon a three year project to produce VIA carcase trait genomic breeding values for UK Limousin cattle. The project is co-funded by the government-backed Technology Strategy Board and BBSRC. This project will combine abattoir VIA carcase information on slaughter animals and genotypes from Limousin sires to produce a UK Limousin SNP key for carcase traits. The SNP key can be thought of as a library containing the different DNA signatures in the population and how the carcase traits are influenced by these signatures. Once developed, Limousin genomic breeding values will then be accessed though Genesure – a subsidiary company formed by the project partners – which will assist breeders to identify which animals to genotype, facilitate the collection of DNA samples, and coordinate the transfer of genotype information and the resulting genomic breeding values.

Using Video Image Analysis (VIA) to assess abattoir carcase yields

VIA machines are integrated into the abattoir processing chain and mechanically grade carcases for EUROP conformation grid and 1–5 fat class, as well as automatically measuring individual primal cut yields all at line speed. As part of this project VBS2000 (developed by E+V in Germany) VIA machines measure 9 different primal cuts: 5th rib forequarter, 8th rib hindquarter, flank, fillet, knuckle, rump, silverside, striploin, and topside. In order to measure the primal cut yields the VIA system uses a holding frame to position the carcase in front of a mounted camera and 2D and 3D images are taken using specific lighting arrangements. These images are then analysed using software to assess carcases for EUROP conformation grid and 1–5 fat class and individual primal cut yields. VIA has been shown to be an accurate assessment of carcase characteristics (Allen and Finnerty, 2000) with individual primal cut yields being heritable and with sufficient genetic variation for genetic selection. In the Irish cross bred population, Pabiou et al., 2011 estimated that the heritabilities ranged between 0.13 (0.018) and 0.47 (0.051), depending on the carcase cut and sex of the animal.

Installing VIA machines into abattoirs produces large amounts of accurate information on carcases in the slaughter population. In approximately 18 months, over 80,000 VIA records from over 60 breeds have been collected from just one abattoir. This represents all animals slaughtered from prime slaughter steers to cull cows. About a quarter of animals slaughtered were prime steers and this data was extracted out and the VIA traits analysed. Figure 1 shows that for all VIA measures and carcase weight there is substantial variation across the beef industry.

Figure 2 looks closer at one of the primal cuts (striploin) and it can be seen that there is a strong relationship between increasing striploin weight with increasing carcase weight and increasing carcase conformation. However, more importantly, it can also be seen that at any given carcase weight or conformation grid there is significant variation in striploin weight. The variation within breeds was also observed to be as great as the variation between breeds.

The phenotypic variation observed for given carcase weights and conformation also translates into differences in financial value. The retail prices of individual primal cuts were taken from www.meatprices.co.uk and the retail value of the 9 VIA cuts computed for the 20,000 steers in the dataset. A EUROP purchase price was also computed based on the carcase weight, EUROP conformation grid and 1–5 fat class. A strong relationship was observed with animals that were paid a higher EUROP price being those with a greater retail value for the 9 VIA primal cuts. There was however large variation in retail value for any given EUROP price paid. Closer examination of the data showed that for any given EUROP conformation grid value, animals varied in retail value of VIA primal cuts by as much as £750. For any given carcase weight the variation in retail value of VIA primal cuts was £300 and within animals that received the same EUROP purchase price (same weight and EUROP confirmation grid and 1–5 fat class) the variation was £100. This variation represents an opportunity to better differentiate the value of carcases and reward accordingly.

Genetic improvement of VIA beef carcase traits

BLUP (Best Linear Unbiased Prediction) EBVs are a very effective tool to undertake genetic improvement, which is a permanent, cumulative and cost effective method to improve carcase traits. Preliminary analysis of the VIA carcase traits has shown them to be both heritable and have genetic variation. Our analysis has estimated moderate heritabilities ranging from 0.19 (fillet) to 0.37 (striploin) for the primal cuts, which are comparable to the heritabilities of live weight. Preliminary EBVs have been produced for carcase weight and individual VIA primal cut yields adjusted for carcase weight. The primal cut EBVs are adjusted for carcase weight so that selecting animals with higher EBVs selects animals with a larger primal cut for a given weight and not just the animals with heavier carcase weights.

In the current dataset there were 30 Limousin sires that had 10 or more progeny with VIA carcase traits recorded. The 9 VIA primal cut EBVs were weighted based on their retail prices to construct a VIA retail value index. These 30 sires are only a very small subset of the Limousin population and their conventional EBVs do not represent the extremes of the population. Nonetheless, large differences were observed for these 30 sires. When the raw VIA phenotypes and the resulting retail value are considered, a difference in retail value of £350 was observed between the top and bottom 4 sires. Preliminary BLUP VIA EBVs were computed and showed that while there were general agreement with the ranking based on the VIA phenotype there were also some large differences. For example, taking the top 4 sires based on the raw phenotypes, three of these sires also had high EBVs indicating that their progeny performed well because they had good genetics. One sire however, despite having one of the highest average phenotypes had one of the lowest EBVs. This indicates that the high phenotypes are the result of environmental effects (i.e. amount of feed) and not genetics, demonstrating the value of EBVs when making breeding decisions. The difference between the best and worst sire, ranked according to their EBVs using an economic retail index, was £27.50. This might not seem to be a big difference but comparing the top and bottom 4 sires – ranked on their genetic merit for VIA carcase traits – showed a difference of £463 in the retail value of their progeny.

Conventional EBVs for VIA carcase traits allow sires to be selected based on their genetic merit for carcase traits. However, these EBVs rely on carcase information from progeny – which is available after selection decisions have already been made. This means that at the time of selection EBVs based on parent averages will only be available, and while EBVs based on parent averages are still informative they are less accurate. The late availability of phenotypes makes carcase traits ideal for genomic selection. Through genomic selection, accurate GEBVs would be available at birth allowing for more accurate selection decisions improving the genetic gain of VIA carcase traits. Once a genomic evaluation system has been established, breeders will be able to take a tissue sample (via a nasal swab or hair sample) which will then be genotyped in a laboratory to provide information about the animal’s DNA. The DNA will then be compared to a ‘SNP key’ specific to the population (in this case UK Limousin) to provide a GEBV. This GEBV can then be used in exactly the same way as the current EBVs to make selection decisions. The difficult and time consuming part of genomic selection is developing the ‘SNP key’. The ‘SNP key’ is developed using a reference population of animals that are genotyped and have high quality phenotypes. This project is amassing high numbers of VIA phenotypes from the ABP abattoirs and EBVs are being produced to use as the high quality phenotypes. We are now identifying animals with accurate VIA EBVs to obtain tissue samples to genotype and include in a UK Limousin reference population to produce the ‘SNP key’ and deliver GEBVs to industry.

Benefits to the UK beef industry

There are many benefits of this project to the UK beef industry. Limousin breeders will be able to genotype animals and receive genomic breeding values for VIA carcase traits. This will not only accelerate genetic progress for carcase traits but will also be the first UK breeding values based on actual carcase traits rather than proxy measures as well as the first genomic breeding values for the UK beef industry. With the methodology established it will be easier for other breeds to establish a reference population and obtain GEBVs for VIA carcase traits. Likewise, the methodology can also be extended to include more traits in the future by expanding the phenotypes included in the reference population.

Not only does the introduction of GEBVs for VIA carcase trait provide the potential for substantial genetic progress of carcase traits but it also provides opportunities to strengthen market signals between the individual sectors of the beef supply chain. Currently the pedigree sector breeds cattle with improved EBVs for 400 day live weight and ultrasound fat and muscle depth of animals at 400 days. These EBVs also provide information for the commercial sector when choosing bulls to purchase and produce offspring for slaughter. However, when the commercial sector sell cattle to the abattoir they are paid according to the EUROP grid based on carcase weight and EUROP confirmation grid and 1–5 fat class. The different methods of assessing carcase along the supply chain can mean that market signals can be diluted along the chain. Using the same method of assessing carcases – like VIA carcase traits – along the whole of the supply chain, from pedigree breeders to the abattoirs, will enable clearer market signals to move along the supply chain and enable the beef industry to produce animals with improved carcases that consistently meet market specification.

Conclusions

The implementation of VIA carcase trait genomic selection is an exciting time and will provide many benefits and opportunities for the UK Limousin beef population as well as the wider beef industry.

Acknowledgements

This project is a collaboration involving Anglo Beef Processors, British Limousin Cattle Society and Scotland’s Rural College and is co-funded by the government backed Technology Strategy Board and BBSRC.

References

Allen, P. and Finnerty, N. (2000). Objective beef carcass classification. A report of a trial of three VIA classification systems. Ashtown National Food Center, Dublin 15, Teagasc publication 34 pages.

Meuwissen, T.H.E., Hayes, B.J. and Goddard, M.E. (2001). Prediction of total genetic value using genome-wide dense marker maps. Genetics, 157, 1819–1829.

Pabiou, T., Fikse, W.F., Cromie, A.R., Keane, M.G., Näsholm, A. and Berry, D.P. (2010). Use of digital images to predict carcass cut yields in cattle. Livestock Science, 137: 130–140.

Pabiou, T., Fikse, W.F., Amer, P.R., Cromie, A.R., Näsholm, A. and Berry, D.P. (2011). Genetic variation in wholesale carcass cuts predicted from digital images in cattle. Animal, 5(11): 1720–1727.