Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Comparison of two live-animal ultrasound systems for genetic evaluation of carcass traits in Angus cattle

Comparison of two live-animal ultrasound systems for genetic evaluation of carcass traits in... Comparison of two live-animal ultrasound systems for genetic evaluation of carcass traits in Angus cattle †,1 ‡ † ‡ C.J. Duff, J.H.J. van der Werf, P.F. Parnell, and S.A. Clark † ‡ Angus Australia, Armidale, New South Wales, 2350, Australia; and School of Environmental and Rural Science, University of New England, Armidale, New South Wales, 2351, Australia ABSTRACT:  The improvement of carcass traits for CUP compared to 0.37 for PIE) and after 100 is an important breeding objective in beef cattle d feeding (0.54 for CUP compared to 0.45 PIE). breeding programs. The most common way of CUP predicted IMF also tended to have stronger selecting for improvement in carcass traits is via correlations with the breeding objective traits of indirect selection using ultrasound scanning of se- carcass IMF and marbling traits, both genetically lection candidates which are submitted to genetic (ranging from 0.59 to 0.75 for CUP compared to evaluation programs. Two systems used to ana- 0.45–0.63 for PIE) and phenotypically (ranging lyze ultrasound images to predict carcass traits from 0.27 to 0.43 for CUP compared to 0.19–0.28 are the Pie Medical Esaote Aquila (PIE) and for PIE). Ultrasound scan EMA was the only Central Ultrasound Processing (CUP). This study group of traits in which the heritabilities were compared the ability of the two systems to predict higher for PIE (0.52 for PIE compared to 0.40 for carcass traits for genetic evaluation in Australian CUP at feedlot intake and 0.46 for PIE compared Angus cattle. Genetic and phenotypic param- to 0.43 for CUP at 100 d of feeding), however with eters were estimated using data from 1,648 Angus similar relationships to the breeding objective car- steers which were ultrasound scanned twice with cass EMA observed. For subcutaneous fat traits both systems, first at feedlot entry and then fol- of ultrasound RIB and RUMP, the heritabilites lowing 100 d in the feedlot. The traits interpreted and genetic correlations to the related carcass from ultrasound scanning included eye muscle traits were similar, with the exception being the area (EMA), rib fat (RIB) rump fat (RUMP), higher heritability observed for CUP predicted and intramuscular fat (IMF). Abattoir carcass RUMP at feedlot intake at 0.52 compared to 0.38 data were collected on all steers following the full for PIE. The results from this study indicates that feedlot feeding period of 285 d. For all ultrasound the CUP system, compared to PIE, provides an scan traits, CUP resulted in higher phenotypic and advantage for genetic evaluation of carcass traits genetic variances compared to the PIE. For IMF, in Angus cattle, particularly for the IMF and as- CUP had higher heritability at feedlot intake (0.51 sociated marbling traits. Key words: Angus, beef cattle, carcass, genetic parameters, phenotypic parameters ultrasound © The Author(s) 2021. Published by Oxford University Press on behalf of the American Society of Animal Science. This is an Open Access article distributed under the terms of the Creative Commons Attribution- NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Transl. Anim. Sci. 2021.5:1-11 doi: 10.1093/tas/txab011 INTRODUCTION Corresponding author: christian@angusaustralia.com. au A common breeding objective for beef produ- Received November 9, 2020. cers is to improve carcass traits of animals used Accepted January 21, 2021. 1 Duff et al. in breeding programs. Traditionally, carcass traits MATERIALS AND METHODS have proven expensive and difficult to measure and they cannot be measured on selection candidates. Animal Care Due to this limitation, breeders use correlated Records collected during the feedlot feed- ultrasound scan measurements on the live animal, ing period were subject to animal ethics approval submitted to genetic evaluation programs, to in- AEC12-082. Data for carcass traits were collected crease selection accuracy for carcass traits related as part of routine commercial animal manage- to meat quantity and quality, including eye muscle ment and, therefore, not subject to animal care and area (EMA), rib fat (RIB) rump fat (RUMP) and animal ethics committee approval. intramuscular fat (IMF). Since becoming avail- able in the mid-1990s, ultrasound scanning for car- cass traits has been widely adopted in beef cattle Animals, Phenotypes, and Pedigree breeding programs. During this period, ultrasound scan records on over 640,000 animals have been re- All phenotypic data, associated fixed effects corded in the Angus Australia performance data- and pedigree data used in this study were supplied base and included in genetic evaluation and the by Angus Australia and generated from the Angus production of Estimated Breeding Values for car- Sire Benchmarking Program, also known as the cass traits (A Byrne, Angus Australia, pers. comm., Angus Beef Information Nucleus (BIN), described August 31, 2020). by Parnell et  al. (2019). The animals in the study The most common ultrasound scanning tech- (n  =  1,648), born across 2011–2015 calving years, nology used to predict carcass traits in Australian were straightbred steer progeny of Angus sires Angus herds is the Esaote Aquila system produced (n = 173) and Angus dams (n = 1,448) from seven by Pie Medical (PIE). This technology facilitates different co-operator herds located in New South crush-side and real-time image capture, inter- Wales and Victoria, Australia. Of the dams, the pretation and analysis using inbuilt software and majority had a single progeny represented, while algorithms. An alternative approach, which is com- 190 had two or more progeny included in the study. monly used in the United States of America, is the In contemporary groups, the progeny were Central Ultrasound Processing (CUP) system. The ultrasound scanned twice, first at feedlot entry at CUP system uses different software, algorithms and an average age of 511 d (SD 72.4), then following processes to predict carcass traits through a central- an average of 103 d in a feedlot at an average age ized image analysis laboratory based on images that of 614 d (SD 78.4). The steers were then harvested are also captured crush-side through ultrasound for slaughter, staying in their contemporary groups scanning. (i.e., no selective harvesting), at an average age of Previous studies have published estimates of 795 d (SD 70.0) following the full feeding period of genetic and phenotypic parameters using Angus approximately 285 d. cattle for carcass traits based on ultrasound scan The first feedlot phase (initial 103 d on average) records (Reverter et  al., 2000; Kemp et  al., 2002; was undertaken at Tullimba, Kingstown, NSW, Boerner et al., 2013), but none of these compared Australia, where the steers had ad libitum access to different ultrasound scan systems. Herring et  al. a ration composed of 74.8% tempered barley, 4.6% (1998) compared four ultrasound scan systems, cotton hulls, 6% cottonseed, 5% mill run, 4.6% but focused solely on the phenotypic prediction chopped hay, and 5% liquid mineral supplement. of carcass IMF. No other comparisons have been Here, they were fed utilizing the GrowSafe feeding published describing the precision of different live system (GrowSafe Systems Ltd., Airdrie, Alberta, animal ultrasound systems for predicting carcass Canada). Following this first feedlot phase, the traits, or the genetic parameters between the dif- steers were relocated for phase two (final 182 d on ferent systems, including their relationships to the average) to Rangers Valley Feedlot, Rangers Valley, direct carcass traits. NSW to finish the feeding program. Here, they were The objective of this study was to estimate fed a similar ration under a normal, controlled phenotypic and genetic parameters for the ultra- commercial feeding program. sound scan measured traits (IMF, EMA, RIB, and All animals were ultrasound scanned at the RUMP) to compare the two live-animal ultrasound 12th and 13th rib site and P8 site by one ac- systems (PIE and CUP) and determine their gen- credited and experienced technician (Upton etic relationships with the direct carcass breeding et  al., 1999). All steers within a contemporary objective traits for genetic evaluation programs. group were scanned on the same day with the Translate basic science to industry innovation Comparison of ultrasound to predict carcass Esaote Aquila system (Pie Medical, Maastricht, chine centered on the crest of the spinous process The Netherlands) equipped with a 3.5-MHz, of the third sacral vertebrae (AUS-MEAT, 2020). 18-cm transducer. EMA, RIB, RUMP, and Additionally, all animals in this study with IMF were measured using the PIE software CUP phenotypes also had the matching PIE providing a real-time prediction at feedlot in- phenotype recorded. Some steers with PIE pheno- take (InP-EMA InP-RIB, InP-RUMP, and InP- types did not have the matching CUP phenotype, IMF, respectively) and repeated after 103 d on mainly due to the CUP image capture system not average in the feedlot (100dP-EMA 100dP-RIB, being available at three scanning events accounting 100dP-RUMP, and 100dP-IMF, respectively). for 173 steers at feedlot intake and 66 steers at 100 At the same time, images using the same ultra- d of feeding. A  smaller number of animals could sound hardware, from the same physiological lo- not have phenotypes provided by CUP due to the cations on the animal, were captured using CUP image quality not meeting the required standards image capture software and sent to the CUP la- for phenotype interpretation. This ranged from 10 boratory (Ames, Iowa, USA) for image interpret- to 22 steers across the ultrasound scan traits and ation and prediction of the same traits as the PIE events. The number of records and descriptive stat- system. Being EMA, RIB, RUMP, and IMF at istics for all traits are shown in Table 1. feedlot intake (InC-EMA InC-RIB, InC-RUMP, and InC-IMF, respectively) and after 103 d on Analysis Models average in the feedlot (100dC-EMA 100dC-RIB, 100dC-RUMP, and 100dC-IMF, respectively). ASReml software (Gilmour et  al., 2009)  was The phenotypes returned from CUP compared to used to model each trait and to estimate param- PIE were reported to an extra decimal place (e.g., eters based on univariate and bivariate mixed two compared to one) which may be considered model analysis including up to three generations additional precision. of pedigree. Maternal grandparents of the steers At the end of the feeding period, steers were were unknown from five of the seven co-operator harvested, on the same day, within contemporary herds. Fixed effects fitted in all models included the groups (i.e., no selective harvesting). On the day of contemporary group and dam age. Age at measure- harvest, hot standard carcass weight (C-WT) and ment was fitted by linear regression for ultrasound hot rump fat (C-RUMP) measured on the P8 site scan traits, while carcass weight was fitted by linear were collected. The following day the chilled car- regression for each of the other carcass traits. The casses were graded by experienced Meat Standards contemporary group included animals from the Australia (MSA) graders (Polkinghorne et al., 2008) same herd, year of birth, birth type (twin v single), for eye muscle area (C-EMA), rib fat (C-RIB), breeder-defined management group, and observa- MSA marbling score (C-MMBL) and AUS-MEAT tion date (ultrasound scan or harvest date). This marbling score (C-AMBL) (AUS-MEAT, 2020). resulted in 54 unique contemporary groups for the All carcass grade data was collected by the one ultrasound scan traits including an average of 30 grader on each steer carcase. Additionally, meat animals and 53 unique contemporary groups for samples were collected from the grading site, at the the carcass traits including an average of 26.4 ani- 12th and 13th rib, and assessed for IMF (C-IMF) mals. In all cases, contemporary group was a sig- using soxhlet calibrated near-infrared spectropho- nificant fixed effect (P  <  0.001), while the level of tometry (NIR), described by Perry et al. (2001). To significance varied for the other fixed effects. For ensure consistency and data quality, experienced consistency, the fixed effects as described above Angus Australia staff oversaw all collection on the were included in all models. The univariate animal live steers and their carcasses in the abattoir. models are expressed as The EMA trait measured in this study by PIE, y = Xb + Zu + e CUP, and on the carcass, is also commonly re- ferred to as rib eye area (REA). Furthermore, in where y is the vector of the trait phenotype; X is Australian abattoirs, hot carcasses are routinely the matrix which relates to the fixed effects; b is the measured for subcutaneous fat depth at the P8 site, vector of the fixed effect of the traits analysed; Z also referred to as rump fat, as an indicator of seal- is the matrix which relates to the animal effect; u is able meat yield and market suitability. The P8 site the vector of the random additive genetic effect of is defined as the point of intersection of a line from the animal; and e is the vector of residual effects for the dorsal tuberosity of the tripartite tuber ischii the traits analysed. The expectations and variance parallel with the chine, and a line at 90° to the sawn matrices for random vectors are described as Translate basic science to industry innovation Duff et al. Table 1. Descriptive statistics Trait Unit of measure n Mean SD Minimum Maximum CV (%) Ultrasound scan at feedlot intake* InP-IMF % 1,622 4.5 1.2 1.3 7.7 26.5 InC-IMF % 1,457 4.9 1.8 1.1 10.3 35.8 InP-EMA cm 1,647 59.7 5.6 41.0 79.0 9.4 InC-EMA cm 1,457 61.0 7.1 42.6 93.5 11.7 InP-RIB mm 1,648 4.4 1.8 1.0 11.0 40.9 InC-RIB mm 1,460 5.3 2.3 1.0 16.8 43.4 InP-RUMP mm 1,648 5.7 2.5 1.0 17.0 43.9 InC-RUMP mm 1,458 5.3 2.6 0.8 17.0 49.1 Ultrasound scan at 100 d feeding 100dP-IMF % 1,508 7.2 1.0 3.5 8.3 13.3 100dC-IMF % 1,432 6.0 1.8 1.3 11.9 29.8 100dP-EMA cm 1,508 80.7 8.0 46.0 104.0 9.9 100dC-EMA cm 1,420 83.5 8.6 58.7 115.5 10.3 100dP-RIB mm 1,508 10.5 2.1 5.0 22.0 20.0 100dC-RIB mm 1,429 13.6 3.3 5.3 26.4 24.3 100dP-RUMP mm 1,508 14.0 3.3 5.0 31.0 23.6 100dC-RUMP mm 1,432 14.0 3.5 4.6 30.5 25.0 Carcass C-IMF mm 1,475 10.1 3.3 3.2 25.1 32.6 C-AMBL score 1,473 2.7 1.2 0.0 8.0 46.4 C-MMBL score 1,474 514.4 120.2 160.0 1030.0 23.4 C-EMA cm 1,460 90.2 9.6 66.0 124.0 10.6 C-RIB mm 1,450 18.7 5.5 6.0 40.0 29.4 C-RUMP mm 1,462 23.2 6.3 10.0 50.0 27.2 C-WT kg 1,462 460.2 37.4 334.9 568.6 8.1 Steers ultrasound scanned at feedlot intake at an average age of 511 d (SD 72.4). Steers ultrasound scanned after an average of 103 d on feed, at an average age of 614 d (SD 78.4). Steers harvested and graded at an average age of 796 d (SD 70.0) following an average feedlot period of 285 d. is an identity matrix for the total number of obser-     ñ ô y Xb vations; and ⊗ is the Kronecker product.     u 0 E = ; V Heritability estimates from the univariate mod- e 0 els, as well as phenotypic and genetic correlations from the bivariate models, were calculated from the The bivariate animal models are expressed as resulting variance components. Y = Xb + Zu + e RESULTS AND DISCUSSION where Y is the vector of the trait phenotypes; X is the matrix which relates to the fixed effects; b is the Summary Statistics vector of the fixed effects of the traits analysed; Z is the matrix which relates to the animal effect; u is Summary statistics for the ultrasound scan the vector of the random animal effects; and e is measurements and carcass traits are shown in Table the vector of residual effects for the traits analyzed. 1. Comparing scanning systems, CUP consistently The expectations and variance matrices for random produced more variation as indicated by higher vectors are described as standard deviations and higher coefficients of vari-     ation. This was most noticeable for the IMF ultra- ñ ô ñ ô ñ ô y Xb u G A ⊗ G 0 sound scan trait, for example, 100dC-IMF had a     u 0 E = ; V = = e R 0 I ⊗ R 0.8 higher standard deviation and 16.5% higher co- e 0 efficient of variation compared to 100dP-IMF. Across all ultrasound scan trait traits, there were Where G and R denote the 2  × 2 matrices con- consistently fewer animals measured using the CUP taining additive genetic and residual variance com- system mainly due to the image capture technology ponents; A is the numerator relationship matrix; I not being available for some scanning events. There Translate basic science to industry innovation Comparison of ultrasound to predict carcass was also a decrease in the number of steers between For subcutaneous fat traits of ultrasound RIB scanning and harvest due to the normal attrition and RUMP, the heritabilities were similar across during the lot feeding and pre-harvest phase. This systems, with the exception being the higher herit- was also a function of the time between scanning ability observed for InC-RUMP at feedlot intake at events and harvest with the steers being an average 0.52 compared to 0.38 for InP-RUMP. of 511 d of age at feedlot intake, 614 d of age for The heritability estimates for InC-IMF and the second ultrasound scanning event, and 796 d of 100dC-IMF were noticeably higher than found in age at harvest. previous studies. For example, Walkom et al. (2015) obtained heritability estimates for ultrasound scan IMF of 0.28 in heifers and 0.20 for bulls, based on Variance Component and Heritability Estimates phenotypes collected mainly on the PIE system in Variance components and heritability estimates primarily Angus breeding animals. Similarly, Kelly for the intramuscular fat and marbling traits, eye et  al. (2019) estimated the heritability of ultra- muscle area, rib fat and rump fat are shown in sound scan IMF as 0.25 from a combined dataset Tables 2, 3, 4, and 5, respectively. of bulls, steers and heifers measured using the PIE All heritabilities were moderate to high, con- system. The heritability estimates of InC-IMF and firming that ultrasound scan and direct carcass 100dC-IMF from the current study were similar to traits provide valuable information for genetic the estimates from Kemp et al. (2002) of 0.51. This evaluation of beef cattle. This finding is consistent is a more comparable study as it was undertaken on with previous studies that have estimated vari- Angus steers, rather than bulls or heifers, and with ance components and heritabilities of ultrasound the ultrasound scan images interpreted in a labora- scan traits and direct carcass traits on Angus and tory setting rather than crush-side in real-time. Angus influenced beef cattle populations (Reverter The heritability estimates for InP-IMF and et al., 2000; Kemp et al., 2002; Boerner et al., 2013; 100dP-IMF in the current study were closer to Walkom et al., 2015; Kelly et al., 2019). most previous studies from Australian cattle popu- Comparing the two ultrasound scanning sys- lations, particularly heifers, which was expected tems, CUP resulted in higher phenotypic and gen- as most phenotypic data analyzed in those stud- etic variances compared to PIE for all ultrasound ies were based on the PIE ultrasound technology. scan traits of IMF, EMA, RIB, and RUMP at both Estimates of IMF from bull phenotypes from the steer feedlot intake and after 100 d of feeding. previous studies found lower heritability which is Heritability estimates of InC-IMF and likely to be the result of lower mean intramuscular 100dC-IMF were consistently higher than InP-IMF fat, and therefore genetic differences expressed to a and 100dP-IMF, with a 0.14 increase at steer feedlot lesser degree in bulls compared to heifers and steers intake and a 0.09 increase after 100 d of feeding. (Reverter et al., 2000; Boerner et al., 2013; Walkom Ultrasound EMA, at both feedlot intake and 100 et al., 2015). d of feeding, were the only group of traits in which For ultrasound scan EMA, this study showed the heritabilities were higher for PIE (0.52 for InP- higher heritability compared to previous stud- EMA compared to 0.40 for InC-EMA and 0.46 ies (Reverter et  al., 2000; Kemp et  al., 2002; for 100dP-EMA compared to 0.43 100dC-EMA). Boerner et  al., 2013; Kelly et  al., 2019). For the Table 2.  Heritabilities, additive genetic variances, phenotypic variances, genetic, and phenotypic correl- ations for IMF and carcass marbling traits (standard errors in parenthesis) Variance/trait InP-IMF InC-IMF 100dP-IMF 100dC-IMF C-IMF C-AMBL C-MMBL h 0.37 (0.08) 0.51 (0.09) 0.45 (0.09) 0.54 (0.09) 0.62 (0.09) 0.42 (0.09) 0.46 (0.09) σ a 0.25 0.73 0.13 1.16 5.91 0.57 5,872 σ p 0.68 1.45 0.29 2.13 9.46 1.35 12,794 InP-IMF – 0.79 (0.09) 0.73 (0.10) 0.71 (0.11) 0.64 (0.11) 0.45 (0.14) 0.46 (0.14) InC-IMF 0.34 (0.03) – 0.78 (0.10) 0.98 (0.06) 0.75 (0.09) 0.59 (0.12) 0.64 (0.12) 100dP-IMF 0.39 (0.02) 0.30 (0.02) – 0.76 (0.09) 0.59 (0.10) 0.62 (0.12) 0.63 (0.12) 100dC-IMF 0.30 (0.02) 0.49 (0.02) 0.43 (0.02) – 0.66 (0.09) 0.68 (0.10) 0.74 (0.09) C-IMF 0.27 (0.03) 0.36 (0.02) 0.28 (0.03) 0.43 (0.02) – 0.97 (0.04) 0.96 (0.03) C-AMBL 0.19 (0.03) 0.27 (0.03) 0.24 (0.03) 0.34 (0.03) 0.56 (0.02) – 0.99 (0.01) C-MMBL 0.21 (0.03) 0.30 (0.03) 0.25 (0.03) 0.38 (0.03) 0.62 (0.02) 0.94 (0.01) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Translate basic science to industry innovation Duff et al. Table 3.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for EMA traits (standard errors in parenthesis) Variance/trait InP-EMA InC-EMA 100dP-EMA 100dC-EMA C-EMA h 0.52 (0.09) 0.40 (0.09) 0.46 (0.09) 0.43 (0.08) 0.60 (0.10) σ a 10.01 12.27 13.48 24.55 37.91 σ p 19.43 30.82 29.41 56.85 62.92 InP-EMA – 0.94 (0.06) 0.80 (0.07) 0.92 (0.06) 0.83 (0.07) InC-EMA 0.76 (0.01) – 0.84 (0.08) 0.94 (0.07) 0.90 (0.07) 100dP-EMA 0.52 (0.02) 0.49 (0.02) – 0.94 (0.04) 0.86 (0.07) 100dC-EMA 0.51 (0.02) 0.50 (0.02) 0.71 (0.01) – 0.78 (0.08) C-EMA 0.35 (0.03) 0.36 (0.02) 0.38 (0.03) 0.38 (0.03) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Table 4.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for Rib Fat traits (standard errors in parenthesis) Variance/trait InP-RIB InC-RIB 100dP-RIB 100dC-RIB C-RIB h 0.42 (0.08) 0.44 (0.09) 0.50 (0.09) 0.59 (0.10) 0.40 (0.09) σ a 0.56 1.24 1.70 4.47 10.35 σ p 1.35 2.83 3.39 7.56 26.00 InP-RIB – 0.98 (0.03) 0.75 (0.08) 0.58 (0.10) 0.42 (0.08) InC-RIB 0.75 (0.01) – 0.75 (0.08) 0.58 (0.11) 0.33 (0.15) 100dP-RIB 0.48 (0.02) 0.49 (0.02) – 0.83 (0.04) 0.58 (0.11) 100dC-RIB 0.40 (0.02) 0.42 (0.03) 0.75 (0.01) – 0.60 (0.11) C-RIB 0.33 (0.03) 0.26 (0.03) 0.41 (0.02) 0.41 (0.03) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Table 5.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for Rump Fat traits (standard errors in parenthesis) Variance/trait InP-RUMP InC-RUMP 100dP-RUMP 100dC-RUMP C-RUMP h 0.38 (0.08) 0.52 (0.10) 0.61 (0.10) 0.61 (0.10) 0.50 (0.09) σ a 0.99 1.59 4.97 5.78 14.71 σ p 2.58 3.04 8.12 9.52 29.10 InP-RUMP – 0.80 (0.09) 0.85 (0.06) 0.84 (0.06) 0.55 (0.11) InC-RUMP 0.85 (0.01) – 0.86 (0.06) 0.87 (0.07) 0.55 (0.11) 100dP-RUMP 0.62 (0.02) 0.61 (0.02) – 0.99 (0.05) 0.75 (0.07) 100dC-RUMP 0.59 (0.02) 0.59 (0.02) 0.94 (0.01) – 0.71 (0.07) C-RUMP 0.41 (0.02) 0.41 (0.02) 0.55 (0.02) 0.55 (0.02) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. subcutaneous fat ultrasound scan traits of RIB of C-RIB and C-RUMP the heritability esti- and RUMP, the heritability results were like pre- mates where similar to the associated ultrasound vious studies. scan traits. The heritability estimates for the carcass marb- The heritability estimates for the carcass traits ling and EMA traits were higher than the scan were generally higher in the current study com- traits with estimates of 0.62, 0.42, 0.46, and 0.60 pared to some previous reports. For example, for C-IMF, C-AMBL, C-MMBL, and C-EMA, re- Borner et al. (2013) estimated heritabilities for car- spectively. For the marbling traits, the higher her- cass IMF, carcass rump fat, carcass rib fat and car- itability for C-IMF is expected given the objective cass eye muscle area of 0.33, 0.36, 0.23, and 0.39, NIR assay used to precisely measure this trait, respectively. In the current study, steers were killed compared to the subjective scoring by a human at an older age and higher carcass weight resulting grader and categorical nature of both C-AMBL in higher means and variances for all carcass traits. and C-MMBL. For the subcutaneous fat traits However, several other reports, based on similar Translate basic science to industry innovation Comparison of ultrasound to predict carcass cattle and production systems, showed comparable heritabilities to this study. For example, the Torres- Vázquez et al. (2018), Jeyaruban et al. (2017), and Kemp et  al. (2002) obtained heritability estimates for carcass IMF at 0.61, MSA marbling score at 0.48 and USDA marbling score at 0.40, respectively. In most genetic evaluation programs, it is more common for bulls and heifers to be ultrasound scanned for the correlated carcass traits, rather than steers as in the current study. For example, from the 593,376 ultrasound scan IMF records on the Angus Australia database, 49.8% are from bulls, 45.6% from heifers, and 4.6% from steers (A Byrne, Angus Australia 2020, pers. comm., January 28, 2020). While this should be a consideration in the interpretation and application of the results from this study, a similar study (Duff et  al., 2018) of combined steer and heifer data showed comparable results to this study, particularly the higher herit- ability for CUP IMF compared to PIE IMF. It is common practice to combine heifer and steer ultrasound scan data for parameter estima- tion and genetic evaluation. For example, Walkom et  al. (2015) observed substantially higher genetic Figure 1. Distribution of ultrasound scan intramuscular fat (IMF) steer phenotypes for the PIE (top) and CUP (bottom) systems at 100 variance and heritabilties for ultrasound scan IMF d feeding. from the combined heifer and steer phenotypes, compared to bull phenotypes. There are no known previous reports where Angus bulls have been ultra- significant. For example, the genetic correlation sound scanned for IMF, EMA, RIB, and RUMP with C-IMF was 0.11 higher for InC-IMF and with both the PIE and CUP systems. 0.07 higher for 100dC-IMF compared to the PIE estimates at the same event. A  possible explan- Genetic Correlations ation for this may be the narrower range of IMF prediction for PIE compared to CUP. The PIE Genetic correlations for the intramuscular system, and its in-built algorithm used to predict fat and marbling traits, eye muscle area, rib fat IMF, is known to be most effective between 2.0% and rump fat are shown in Tables 2, 3, 4, and 5, and 8.0% IMF range (R. Evans, Bovine Scanning respectively. Services Pty Ltd, pers. comm., January 6, 2021). The genetic correlations between ultrasound For this reason, we observed few records in this scan traits and the direct breeding objective car- study that are less than 2.0% or greater than 8.0% cass traits presented were positive and moderate from PIE, compared to CUP, particularly in the to strong. This is consistent with previous studies 100-d scan (Figure 1) where we expect to observe (Reverter et  al., 2000; Kemp et  al., 2002; Borner a higher proportion of IMF values greater than et  al., 2013; Walkom et  al., 2015), showing that 8%. The CUP system can predict IMF to a wider ultrasound scanning is a valuable indirect carcass range (Table 1) and can more precisely determine measurement for informing genetic evaluation pro- genetic merit by explaining greater genetic vari- grams of beef cattle. ance (Table 2). This is evident and consistent at Comparing the two ultrasound scanning sys- both steer intake and 100 d on feed. This finding tems, the genetic correlations of ultrasound scan also highlights the difference observed in the gen- IMF and the breeding objective carcass IMF etic correlations between the same ultrasound (C-IMF) and marbling traits (C-AMBL and scan system at steer intake and 100 d feeding, C-MMBL) tended to be higher for CUP compared being 0.98 for CUP and 0.73 for PIE. In con- to PIE, at both ultrasound scanning events, how- trast, the genetic correlations with the breeding ever, also acknowledging that the standard errors objective carcass traits tended to be similar when of these estimates indicate the differences are not Translate basic science to industry innovation Duff et al. comparing CUP to PIE for the ultrasound scan fat and rump fat are listed in Tables 2, 3, 4, and 5, traits of EMA, RIB, and RUMP. respectively. The genetic correlations for the ultrasound IMF Like the genetic parameters, the phenotypic traits with the carcass IMF and marbling traits were correlations for the CUP system for IMF to generally stronger than reported in previous studies. C-IMF, C-AMBL, and C-MMBL were higher For example, Reverter et al. (2000), obtained genetic than for the PIE system. While for ultrasound correlation estimates for carcass IMF to bull IMF scan EMA, RIB, and RUMP the phenotypic ultrasound of 0.47 and heifer IMF ultrasound of correlations to the associated breeding objective 0.46. Kemp et al. (2002), reported a much stronger traits tended to be similar between systems. An genetic correlation of 0.90 in steers, but the time exception being the 0.33 correlation for InP- interval in that study was much shorter between RIB with C-RIB, compared to 0.26 for InC-RIB ultrasound scanning and harvest of i.e. 52 d, com- with C-RIB. pared to a 285 and 182-d interval in the current study. Herring et al. (1998) reported stronger pheno- For EMA, the genetic correlations of PIE and typic correlations for the ultrasound predicted CUP ultrasound to carcass EMA were stronger in IMF traits with carcass IMF and marbling score this study than those reported in previous studies across four different ultrasound systems, including (Kemp et  al., 2002; Reverter et  al. 2000; Borner CUP (described as CVIS) and PIE in crossbred et al., 2013). The high correlations observed in the beef steers. They reported phenotypic correlations current study may have been a function of the use for CUP IMF with carcass IMF and marbling of highly experienced ultrasound scanning techni- score of 0.61 and 0.74, respectively. While for PIE cians on the live steers, experienced carcass grad- IMF to carcass IMF and marbling score, the re- ers in the abattoir and controlled data collection, ported estimates were 0.31 and 0.39, respectively. whereas field data from large scale bull breeding The contrasting results between studies is likely to herds were mostly utilized in other studies. be due to the different time intervals between ultra- In contrast, for the fat traits, genetic correlations sound scanning steers and their harvest followed of ultrasound rib and rump fat to the respective by carcass data collection. The interval was much carcass measures were weaker than those observed shorter in the Herring et  al. (1998) study ranging in previous studies. The possible reason being unin- from 8 to 14 d. tended abattoir effects, such as hide puller damage The results from the current study are more on the fat distribution on the long-fed steer car- likely to reflect industry practice, as phenotypic se- casses. With more subcutaneous fat observed on lection (e.g., drafting pre-harvest) with short time the long-fed steer carcass compared to shorter fed intervals between ultrasound scanning and harvest steers, there is higher probability of damage to the is unlikely due to the associated stressors having subcutaneous fat which may lead to reduced pre- negative impacts on meat quality through dark cut- cision of measurement in the chiller. For example, ting (Ponnampalam et al., 2017) or welfare implica- in this study steers averaged 18.7  mm for carcass tions of increased injury risk. It is more practical, rib fat at an average of 460.2 kg carcass weight. In and therefore more likely, to ultrasound scan ani- contrast, in the Reverter et al. (2000), Borner et al. mals on-farm or at feedlot induction, well before (2013) and Kemp et  al. (2002) studies the carcass harvest. rib fat measurements were leaner at 6.2, 9.0, and For the carcass traits, the correlations between 14.1 mm, respectively, and with lighter carcasses. C-IMF with C-AMBL and C-MMBL were mod- The genetic correlations between the three car- erate at 0.56 and 0.62, respectively. These esti- cass traits of C-IMF, C-AMBL, and C-MMBL mates were lower than those reported by Lee et al. were very strong and positive ranging from 0.96 to (2019) of 0.87 in a different breed and produc- 0.99. A study with temperate beef cattle by Johnston tion system having higher mean carcass IMF and (2001) showed similar genetic correlations between marbling scores with greater variability. Konarska carcass IMF to MSA Marbling and AUS-meat et al. (2017) reported a closer correlation between marbling score of 1.00 and 0.96, respectively, sup- MSA marbling score and carcass IMF by NIR in porting the findings in this study. M. longissimus thoracis, the same muscle as meas- ured in the current study, of 0.75. The phenotypic correlation of C-AMBL with C-MMBL was high Phenotypic Correlations at 0.94. This was expected as both scores were Phenotypic correlations for the intramus- assessed by the same grader, albeit on different cular fat and marbling traits, eye muscle area, rib scales. Translate basic science to industry innovation Comparison of ultrasound to predict carcass Breeding Program Design CUP, compared to the crush-side and real-time process of PIE. Comparing ultrasound scan methods to predict To benefit from the results of this study, beef carcass traits is an important step in understanding cattle genetic evaluation programs could consider strategies, particularly breeding program design, transitioning all live animal ultrasound phenotype aimed at increasing accuracy of selection and gen- recording to the CUP system or similar systems etic gain for carcass traits in breeding objectives. In using a centralized processing approach and predic- a companion study (Duff et al., 2019), we modeled tion algorithms. An alternative, but more complex, several phenotyping and genotyping scenarios fo- approach is to receive ultrasound phenotypes from cused on the breeding objective traits of C-IMF, a range of systems (e.g., both PIE and CUP) and C-AMBL, and C-MMBL. The study investigated model each specifically to recognize the differences how breeding programs may be enhanced by using in trait variances, heritabilities and genetic correl- genomic-based information as derived from a refer- ations to the breeding objective traits. ence population with direct carcass IMF and marb- Consistent with Kemp et al. (1998), this study ling score phenotypes coupled with genotypes, as confirmed that ultrasound scanning can be used described by Goddard et  al. (2010). This study to effectively predict carcass phenotypes in Angus found the highest rates of selection accuracy and steers, including those from progeny test programs, response would be achieved through a combination to inform genetic evaluation programs, particularly of CUP ultrasound scan phenotyping for IMF where collection of effective carcass data from the and genotyping with a reference population of re- abattoir is not possible or difficult. lated animals with carcass IMF and marbling score A unique feature of this study was the inclusion phenotypes. However, the value of ultrasound scan of three measurements of marbling traits on each phenotyping diminishes as the GBV prediction ac- carcass, being C-IMF, C-AMBL, and C-MMBL. curacy increases, which is mainly a function of the The results indicate that these measures are all reference population size. strongly and positively correlated, both phenotyp- ically and genetically. As a result, for beef cattle CONCLUSIONS genetic evaluation, the collection of just one of the marbling traits is likely to be sufficient. Additional This study compared the phenotypic and gen- benefit is attained from measuring C-IMF due to etic parameters for the ultrasound scan measured the higher heritability of this trait and stronger gen- traits (IMF, EMA, RIB, and RUMP) for two etic correlations with live animal ultrasound scan live-animal ultrasound systems (PIE and CUP) IMF, but there are added cost and sample collec- and estimated their relationship with the direct car- tion considerations associated with the C-IMF cass traits. The results showed substantial genetic phenotype. variation in carcass performance can be measured It is also recognized that the ultrasound scan- using either ultrasound scan system, even when ning hardware used in the study, which is still com- there is a considerable interval (e.g., 285 d) between monly used to phenotype live animals for carcass the ultrasound scanning event and harvest. This traits for genetic evaluation, was developed in the is based on the moderate to high heritabilities ob- 20th century. More sophisticated ultrasound scan served, coupled with moderate to strong relation- systems are available today which can capture ships with the related breeding objective carcass higher quality images and potentially predict more traits. A noticeable difference was the CUP system precise phenotypes, if coupled with appropriate explaining more variation, particularly for ultra- prediction algorithms. A  study to understand po- sound scan IMF, resulting in a higher heritability tential benefits of modern ultrasound scan sys- and stronger correlations with the carcass IMF and tems for beef cattle genetic evaluation programs is marbling traits in the breeding objective. This indi- recommended. cates that the CUP system, compared to PIE, pro- Further research is also warranted to under- vides an advantage for genetic evaluation of carcass stand genotyping and phenotyping strategies traits in Angus cattle. This advantage should be for beef herds with carcass traits included in the considered with knowledge of possible additional breeding objective. As it is more common for costs involved with interpreting ultrasound images breeding candidates to scanned for the correlated through a centralized laboratory. Furthermore, carcass traits, rather than steers as in this study, the there is also a turn-around time of 24–48  h in re- future research also needs to better understand the ceiving the phenotype measurement results from Translate basic science to industry innovation Duff et al. https://www.ausmeat.com.au/WebDocuments/Producer_ genetic and phenotypic relationship between bull, HAP_Beef_Small.pdf [accessed January 6, 2021]. heifer, and steer measurements. Börner,  V., D.  J.  Johnston, and H.  U.  Graser. 2013. Genetic relationships between live animal scan traits and carcass ABBREVIATIONS traits of Australian Angus bulls and heifers. Anim. Prod. Sci. 53: 1075–1082. doi:10.1071/AN12435 100dC-EMA, 100 days on feed eye muscle area Duff,  C., J.  H.  J.  van  der  Werf, and S.  A.  Clark. 2018. using CUP; 100dC-IMF, 100  days on feed intra- Comparison of two live-animal ultrasound systems muscular fat using CUP; 100dC-RIB, 100  days to predict carcass intramuscular fat and marbling in Australian Angus cattle. In: Proceedings of the 11th World on feed rib fat depth using CUP; 100dC-RUMP, Congress on Genetics Applied to Livestock Production 100  days on feed rump fat depth using CUP; No. Electronic Poster Session. Auckland, New Zealand. 100dP-EMA, 100  days on feed eye muscle area p. 262. http://www.wcgalp.org/system/files/proceed- using PIE; 100dP-IMF, 100 days on feed intramus- ings/2018/comparison-two-live-animal-ultrasound-sys- cular fat using PIE; 100dP-RIB, 100  days on feed tems-predict-carcass-intramuscular-fat-and-marbling.pdf rib fat depth using PIE; 100dP-RUMP, 100 days on [accessed February 29, 2020]. Duff, C., J. H J. van der Werf, and S. A. Clark. 2019. Should feed rump fat depth using PIE; ASBP, Angus Sire Angus breeders live-animal ultrasound scan for intramus- Benchmarking Program; BIN, Beef Information cular fat in the genomics era? Proc. Assoc. Adv. Anim. Nucleus; C-AMBL, carcass measured AUS-MEAT Breed. Genet. 23: 496–499. http://www.aaabg.org/aaab- marbling score; C-EMA, carcass measured eye ghome/AAABG23papers/122Duff23496.pdf [accessed muscle area; C-IMF, carcass measured intramus- March 1, 2020]. Gilmour, A. R., B. J. Gogel, B. R. Cullis, and R. Thompson. cular fat by near-infrared spectrophotometry; 2009. ASReml user guide release 3.0. Hemel Hempstead, C-MMBL, carcass measured Meat Standards UK: VSN Int. Ltd. Australia marbling score; C-RIB, carcass measured Goddard,  M.  E., B.  J.  Hayes, and T.  H.  Meuwissen. 2010. cold rib fat; C-RUMP, carcass measured hot P8 fat; Genomic selection in livestock populations. Genet. Res. CUP, Central Ultrasound Processing ultrasound (Camb). 92:413–421. doi:10.1017/S0016672310000613 scan system; C-WT, carcass measured hot carcass Herring,  W.  O., L.  A.  Kriese, J.  K.  Bertrand, and J.  Crouch. 1998. Comparison of four real-time ultrasound systems weight; EMA, eye muscle area; GBV, Genomic that predict intramuscular fat in beef cattle. J. Anim. Sci. Breeding Value; IMF, intramuscular fat; InC- 76:364–370. doi:10.2527/1998.762364x EMA, feedlot intake eye muscle area using CUP; Jeyaruban,  M.  G., D.  J.  Johnston, and B.  J.  Walmsley. 2017. InC-IMF, feedlot intake intramuscular fat using Genetic and phenotypic characterization of MSA index CUP; InC-RIB, feedlot intake rib fat depth using and its association with carcass and meat quality traits in Angus and Brahman cattle. Proc. Assoc. Adv. Anim. CUP; InC-RUMP, feedlot intake rump fat depth Breed. Genet. 22:313–316. http://www.aaabg.org/aaab- using CUP; InP-EMA, feedlot intake eye muscle ghome/AAABG22papers/71Jeyaruban22313.pdf [ac- area using PIE; InP-IMF, feedlot intake intramus- cessed August 10, 2019]. cular fat using PIE; InP-RIB, feedlot intake rib Johnston,  D.  J. 2001. Selecting for marbling and its relation- fat depth using PIE; InP-RUMP, feedlot intake ship with other important economic traits. What impact rump fat depth using PIE; MSA, Meat Standards does it have? In: Proceedings of Marbling Symposium, Coffs Harbour, Australia. p. 88–93. https://pdfs.seman- Australia; PIE, Pie Medical Esaote Aquila ultra- ticscholar.org/e8f9/70fa903a9ba929b83409db8acecb- sound Scan system; RIB, rib fat; RUMP, rump fat 2d2ac196.pdf [accessed March 11, 2020]. Kelly,  D.  N., M.  Murphy, R.  D.  Sleator, M.  M.  Judge, ACKNOWLEDGMENTS S. B. Conroy, and D. P. Berry. 2019. Feed efficiency and car - cass metrics in growing cattle. J. Anim. Sci. 97(11):4405– The research was supported by Meat and 4417. doi:10.1093/jas/skz316 Livestock Australia through MDC matching funds Kemp, D. J., W. O. Herring, and C. J. Kaiser. 2002. Genetic and environmental parameters for steer ultrasound and car- in project PSH.0528. Generous support has been cass traits. J. Anim. Sci. 80(6):1489–1496. Available from provided by numerous Angus bull breeders, co-op- http://search.proquest.com.ezproxy.une.edu.au/scholar- erator herds, supply chain partners, technicians, ly-journals/genetic-environmental-parameters-steer-ul- and research organizations. trasound/docview/218122729/se-2?accountid=17227 Conflicts of interest statement. The authors de- [accessed March 11, 2020]. Konarska,  M., K.  Kuchida, G.  Tarr, and R.  J.  Polkinghorne. clare that there are no conflicts of interest to disclose 2017. Relationships between marbling measures across that relate to the research described in this paper. principal muscles. Meat Sci. 123:67–78. doi:10.1016/j. meatsci.2016.09.005 LITERATURE CITED Lee, B., and Y. M. Choi. 2019. Correlation of marbling char- acteristics with meat quality and histochemical character- AUS-MEAT. 2020. Handbook of Australian Beef Processing, istics in longissimus thoracis muscle from hanwoo steers. Version 7. Murarrie, QLD: AUS-MEAT Ltd. Available from Translate basic science to industry innovation Comparison of ultrasound to predict carcass Food Sci. Anim. Resour. 39(1):151–161. doi:10.5851/ Reverter,  A., D.  J.  Johnston, H.  U.  Graser, M.  L.  Wolcott, kosfa.2019.e12 and W.  H.  Upton. 2000. Genetic analyses of live-ani- Parnell, P. F., C. J. Duff, A. I. Byrne, and N. M. Butcher. 2019. The mal ultrasound and abattoir carcass traits in Australian Angus sire benchmarking program—a major contributor to fu- Angus and Hereford cattle. J. Anim. Sci. 78:1786–1795. ture genetic improvement in the Australian beef industry. Proc. doi:10.2527/2000.7871786x Assoc. Adv. Anim. Breed. Genet. 23:492–495. http://www. Torres-Vázquez, J. A., J. H. J. van der Werf, and S. A. Clark. aaabg.org/aaabghome/AAABG23papers/121Parnell23492. 2018. Genetic and phenotypic associations of feed ef- pdf [accessed November 1, 2020]. ficiency with growth and carcass traits in Australian Perry,  D., W.  R.  Shorthose, D.  M.  Ferguson, and Angus cattle. J. Anim. Sci. 96:4521–4531. doi:10.1093/ J. M. Thompson. 2001. Methods used in the CRC program jas/sky325. for the determination of carcass yield and beef quality. Upton,  W.  H., K.  A.  Donoghue, H.  U.  Graser, and Aust. J. Exp. Agric. 41:953–957. doi:10.1071/EA00092 D.  J.  Johnston. 1999. Ultrasound proficiency testing. Polkinghorne,  R., J.  M.  Thompson, R.  Watson, A.  Gee, Proc. Assoc. Adv. Anim. Breed. Genet. 13:341–344. http:// and M.  Porter. 2008. Evolution of the Meat Standards www.aaabg.org/proceedings/1999/AB99079.pdf accessed Australia (MSA) beef grading system. Aust. J. Exp. Agric. August 10, 2019]. 48:1351–1359. doi:10.1071/EA07177 Walkom,  S.  F., M.  G.  Jeyaruban, and D.  J.  Johnston. 2015. Ponnampalam, E. N., D. L. Hopkins, H. Bruce, D. Li, G. Baldi, Impact of scanning lean cattle on the genetic correlation and A. E. Bekhit. 2017. Causes and contributing factors between scan and carcass intramuscular fat in Angus and to “Dark Cutting” meat: current trends and future direc- Hereford cattle. Proc. Assoc. Adv. Anim. Breed. Genet. tions: a review. Compr. Rev. Food Sci. Food Saf. 16:400– 21:253–256. http://www.aaabg.org/proceedings/2015/ 430. doi:10.1111/1541-4337.12258 Walkom21253.pdf [accessed August 10, 2019]. Translate basic science to industry innovation http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Translational Animal Science Oxford University Press

Comparison of two live-animal ultrasound systems for genetic evaluation of carcass traits in Angus cattle

Loading next page...
 
/lp/oxford-university-press/comparison-of-two-live-animal-ultrasound-systems-for-genetic-2QSmq9dB06
Copyright
© The Author(s) 2021. Published by Oxford University Press on behalf of the American Society of Animal Science.
eISSN
2573-2102
DOI
10.1093/tas/txab011
Publisher site
See Article on Publisher Site

Abstract

Comparison of two live-animal ultrasound systems for genetic evaluation of carcass traits in Angus cattle †,1 ‡ † ‡ C.J. Duff, J.H.J. van der Werf, P.F. Parnell, and S.A. Clark † ‡ Angus Australia, Armidale, New South Wales, 2350, Australia; and School of Environmental and Rural Science, University of New England, Armidale, New South Wales, 2351, Australia ABSTRACT:  The improvement of carcass traits for CUP compared to 0.37 for PIE) and after 100 is an important breeding objective in beef cattle d feeding (0.54 for CUP compared to 0.45 PIE). breeding programs. The most common way of CUP predicted IMF also tended to have stronger selecting for improvement in carcass traits is via correlations with the breeding objective traits of indirect selection using ultrasound scanning of se- carcass IMF and marbling traits, both genetically lection candidates which are submitted to genetic (ranging from 0.59 to 0.75 for CUP compared to evaluation programs. Two systems used to ana- 0.45–0.63 for PIE) and phenotypically (ranging lyze ultrasound images to predict carcass traits from 0.27 to 0.43 for CUP compared to 0.19–0.28 are the Pie Medical Esaote Aquila (PIE) and for PIE). Ultrasound scan EMA was the only Central Ultrasound Processing (CUP). This study group of traits in which the heritabilities were compared the ability of the two systems to predict higher for PIE (0.52 for PIE compared to 0.40 for carcass traits for genetic evaluation in Australian CUP at feedlot intake and 0.46 for PIE compared Angus cattle. Genetic and phenotypic param- to 0.43 for CUP at 100 d of feeding), however with eters were estimated using data from 1,648 Angus similar relationships to the breeding objective car- steers which were ultrasound scanned twice with cass EMA observed. For subcutaneous fat traits both systems, first at feedlot entry and then fol- of ultrasound RIB and RUMP, the heritabilites lowing 100 d in the feedlot. The traits interpreted and genetic correlations to the related carcass from ultrasound scanning included eye muscle traits were similar, with the exception being the area (EMA), rib fat (RIB) rump fat (RUMP), higher heritability observed for CUP predicted and intramuscular fat (IMF). Abattoir carcass RUMP at feedlot intake at 0.52 compared to 0.38 data were collected on all steers following the full for PIE. The results from this study indicates that feedlot feeding period of 285 d. For all ultrasound the CUP system, compared to PIE, provides an scan traits, CUP resulted in higher phenotypic and advantage for genetic evaluation of carcass traits genetic variances compared to the PIE. For IMF, in Angus cattle, particularly for the IMF and as- CUP had higher heritability at feedlot intake (0.51 sociated marbling traits. Key words: Angus, beef cattle, carcass, genetic parameters, phenotypic parameters ultrasound © The Author(s) 2021. Published by Oxford University Press on behalf of the American Society of Animal Science. This is an Open Access article distributed under the terms of the Creative Commons Attribution- NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Transl. Anim. Sci. 2021.5:1-11 doi: 10.1093/tas/txab011 INTRODUCTION Corresponding author: christian@angusaustralia.com. au A common breeding objective for beef produ- Received November 9, 2020. cers is to improve carcass traits of animals used Accepted January 21, 2021. 1 Duff et al. in breeding programs. Traditionally, carcass traits MATERIALS AND METHODS have proven expensive and difficult to measure and they cannot be measured on selection candidates. Animal Care Due to this limitation, breeders use correlated Records collected during the feedlot feed- ultrasound scan measurements on the live animal, ing period were subject to animal ethics approval submitted to genetic evaluation programs, to in- AEC12-082. Data for carcass traits were collected crease selection accuracy for carcass traits related as part of routine commercial animal manage- to meat quantity and quality, including eye muscle ment and, therefore, not subject to animal care and area (EMA), rib fat (RIB) rump fat (RUMP) and animal ethics committee approval. intramuscular fat (IMF). Since becoming avail- able in the mid-1990s, ultrasound scanning for car- cass traits has been widely adopted in beef cattle Animals, Phenotypes, and Pedigree breeding programs. During this period, ultrasound scan records on over 640,000 animals have been re- All phenotypic data, associated fixed effects corded in the Angus Australia performance data- and pedigree data used in this study were supplied base and included in genetic evaluation and the by Angus Australia and generated from the Angus production of Estimated Breeding Values for car- Sire Benchmarking Program, also known as the cass traits (A Byrne, Angus Australia, pers. comm., Angus Beef Information Nucleus (BIN), described August 31, 2020). by Parnell et  al. (2019). The animals in the study The most common ultrasound scanning tech- (n  =  1,648), born across 2011–2015 calving years, nology used to predict carcass traits in Australian were straightbred steer progeny of Angus sires Angus herds is the Esaote Aquila system produced (n = 173) and Angus dams (n = 1,448) from seven by Pie Medical (PIE). This technology facilitates different co-operator herds located in New South crush-side and real-time image capture, inter- Wales and Victoria, Australia. Of the dams, the pretation and analysis using inbuilt software and majority had a single progeny represented, while algorithms. An alternative approach, which is com- 190 had two or more progeny included in the study. monly used in the United States of America, is the In contemporary groups, the progeny were Central Ultrasound Processing (CUP) system. The ultrasound scanned twice, first at feedlot entry at CUP system uses different software, algorithms and an average age of 511 d (SD 72.4), then following processes to predict carcass traits through a central- an average of 103 d in a feedlot at an average age ized image analysis laboratory based on images that of 614 d (SD 78.4). The steers were then harvested are also captured crush-side through ultrasound for slaughter, staying in their contemporary groups scanning. (i.e., no selective harvesting), at an average age of Previous studies have published estimates of 795 d (SD 70.0) following the full feeding period of genetic and phenotypic parameters using Angus approximately 285 d. cattle for carcass traits based on ultrasound scan The first feedlot phase (initial 103 d on average) records (Reverter et  al., 2000; Kemp et  al., 2002; was undertaken at Tullimba, Kingstown, NSW, Boerner et al., 2013), but none of these compared Australia, where the steers had ad libitum access to different ultrasound scan systems. Herring et  al. a ration composed of 74.8% tempered barley, 4.6% (1998) compared four ultrasound scan systems, cotton hulls, 6% cottonseed, 5% mill run, 4.6% but focused solely on the phenotypic prediction chopped hay, and 5% liquid mineral supplement. of carcass IMF. No other comparisons have been Here, they were fed utilizing the GrowSafe feeding published describing the precision of different live system (GrowSafe Systems Ltd., Airdrie, Alberta, animal ultrasound systems for predicting carcass Canada). Following this first feedlot phase, the traits, or the genetic parameters between the dif- steers were relocated for phase two (final 182 d on ferent systems, including their relationships to the average) to Rangers Valley Feedlot, Rangers Valley, direct carcass traits. NSW to finish the feeding program. Here, they were The objective of this study was to estimate fed a similar ration under a normal, controlled phenotypic and genetic parameters for the ultra- commercial feeding program. sound scan measured traits (IMF, EMA, RIB, and All animals were ultrasound scanned at the RUMP) to compare the two live-animal ultrasound 12th and 13th rib site and P8 site by one ac- systems (PIE and CUP) and determine their gen- credited and experienced technician (Upton etic relationships with the direct carcass breeding et  al., 1999). All steers within a contemporary objective traits for genetic evaluation programs. group were scanned on the same day with the Translate basic science to industry innovation Comparison of ultrasound to predict carcass Esaote Aquila system (Pie Medical, Maastricht, chine centered on the crest of the spinous process The Netherlands) equipped with a 3.5-MHz, of the third sacral vertebrae (AUS-MEAT, 2020). 18-cm transducer. EMA, RIB, RUMP, and Additionally, all animals in this study with IMF were measured using the PIE software CUP phenotypes also had the matching PIE providing a real-time prediction at feedlot in- phenotype recorded. Some steers with PIE pheno- take (InP-EMA InP-RIB, InP-RUMP, and InP- types did not have the matching CUP phenotype, IMF, respectively) and repeated after 103 d on mainly due to the CUP image capture system not average in the feedlot (100dP-EMA 100dP-RIB, being available at three scanning events accounting 100dP-RUMP, and 100dP-IMF, respectively). for 173 steers at feedlot intake and 66 steers at 100 At the same time, images using the same ultra- d of feeding. A  smaller number of animals could sound hardware, from the same physiological lo- not have phenotypes provided by CUP due to the cations on the animal, were captured using CUP image quality not meeting the required standards image capture software and sent to the CUP la- for phenotype interpretation. This ranged from 10 boratory (Ames, Iowa, USA) for image interpret- to 22 steers across the ultrasound scan traits and ation and prediction of the same traits as the PIE events. The number of records and descriptive stat- system. Being EMA, RIB, RUMP, and IMF at istics for all traits are shown in Table 1. feedlot intake (InC-EMA InC-RIB, InC-RUMP, and InC-IMF, respectively) and after 103 d on Analysis Models average in the feedlot (100dC-EMA 100dC-RIB, 100dC-RUMP, and 100dC-IMF, respectively). ASReml software (Gilmour et  al., 2009)  was The phenotypes returned from CUP compared to used to model each trait and to estimate param- PIE were reported to an extra decimal place (e.g., eters based on univariate and bivariate mixed two compared to one) which may be considered model analysis including up to three generations additional precision. of pedigree. Maternal grandparents of the steers At the end of the feeding period, steers were were unknown from five of the seven co-operator harvested, on the same day, within contemporary herds. Fixed effects fitted in all models included the groups (i.e., no selective harvesting). On the day of contemporary group and dam age. Age at measure- harvest, hot standard carcass weight (C-WT) and ment was fitted by linear regression for ultrasound hot rump fat (C-RUMP) measured on the P8 site scan traits, while carcass weight was fitted by linear were collected. The following day the chilled car- regression for each of the other carcass traits. The casses were graded by experienced Meat Standards contemporary group included animals from the Australia (MSA) graders (Polkinghorne et al., 2008) same herd, year of birth, birth type (twin v single), for eye muscle area (C-EMA), rib fat (C-RIB), breeder-defined management group, and observa- MSA marbling score (C-MMBL) and AUS-MEAT tion date (ultrasound scan or harvest date). This marbling score (C-AMBL) (AUS-MEAT, 2020). resulted in 54 unique contemporary groups for the All carcass grade data was collected by the one ultrasound scan traits including an average of 30 grader on each steer carcase. Additionally, meat animals and 53 unique contemporary groups for samples were collected from the grading site, at the the carcass traits including an average of 26.4 ani- 12th and 13th rib, and assessed for IMF (C-IMF) mals. In all cases, contemporary group was a sig- using soxhlet calibrated near-infrared spectropho- nificant fixed effect (P  <  0.001), while the level of tometry (NIR), described by Perry et al. (2001). To significance varied for the other fixed effects. For ensure consistency and data quality, experienced consistency, the fixed effects as described above Angus Australia staff oversaw all collection on the were included in all models. The univariate animal live steers and their carcasses in the abattoir. models are expressed as The EMA trait measured in this study by PIE, y = Xb + Zu + e CUP, and on the carcass, is also commonly re- ferred to as rib eye area (REA). Furthermore, in where y is the vector of the trait phenotype; X is Australian abattoirs, hot carcasses are routinely the matrix which relates to the fixed effects; b is the measured for subcutaneous fat depth at the P8 site, vector of the fixed effect of the traits analysed; Z also referred to as rump fat, as an indicator of seal- is the matrix which relates to the animal effect; u is able meat yield and market suitability. The P8 site the vector of the random additive genetic effect of is defined as the point of intersection of a line from the animal; and e is the vector of residual effects for the dorsal tuberosity of the tripartite tuber ischii the traits analysed. The expectations and variance parallel with the chine, and a line at 90° to the sawn matrices for random vectors are described as Translate basic science to industry innovation Duff et al. Table 1. Descriptive statistics Trait Unit of measure n Mean SD Minimum Maximum CV (%) Ultrasound scan at feedlot intake* InP-IMF % 1,622 4.5 1.2 1.3 7.7 26.5 InC-IMF % 1,457 4.9 1.8 1.1 10.3 35.8 InP-EMA cm 1,647 59.7 5.6 41.0 79.0 9.4 InC-EMA cm 1,457 61.0 7.1 42.6 93.5 11.7 InP-RIB mm 1,648 4.4 1.8 1.0 11.0 40.9 InC-RIB mm 1,460 5.3 2.3 1.0 16.8 43.4 InP-RUMP mm 1,648 5.7 2.5 1.0 17.0 43.9 InC-RUMP mm 1,458 5.3 2.6 0.8 17.0 49.1 Ultrasound scan at 100 d feeding 100dP-IMF % 1,508 7.2 1.0 3.5 8.3 13.3 100dC-IMF % 1,432 6.0 1.8 1.3 11.9 29.8 100dP-EMA cm 1,508 80.7 8.0 46.0 104.0 9.9 100dC-EMA cm 1,420 83.5 8.6 58.7 115.5 10.3 100dP-RIB mm 1,508 10.5 2.1 5.0 22.0 20.0 100dC-RIB mm 1,429 13.6 3.3 5.3 26.4 24.3 100dP-RUMP mm 1,508 14.0 3.3 5.0 31.0 23.6 100dC-RUMP mm 1,432 14.0 3.5 4.6 30.5 25.0 Carcass C-IMF mm 1,475 10.1 3.3 3.2 25.1 32.6 C-AMBL score 1,473 2.7 1.2 0.0 8.0 46.4 C-MMBL score 1,474 514.4 120.2 160.0 1030.0 23.4 C-EMA cm 1,460 90.2 9.6 66.0 124.0 10.6 C-RIB mm 1,450 18.7 5.5 6.0 40.0 29.4 C-RUMP mm 1,462 23.2 6.3 10.0 50.0 27.2 C-WT kg 1,462 460.2 37.4 334.9 568.6 8.1 Steers ultrasound scanned at feedlot intake at an average age of 511 d (SD 72.4). Steers ultrasound scanned after an average of 103 d on feed, at an average age of 614 d (SD 78.4). Steers harvested and graded at an average age of 796 d (SD 70.0) following an average feedlot period of 285 d. is an identity matrix for the total number of obser-     ñ ô y Xb vations; and ⊗ is the Kronecker product.     u 0 E = ; V Heritability estimates from the univariate mod- e 0 els, as well as phenotypic and genetic correlations from the bivariate models, were calculated from the The bivariate animal models are expressed as resulting variance components. Y = Xb + Zu + e RESULTS AND DISCUSSION where Y is the vector of the trait phenotypes; X is the matrix which relates to the fixed effects; b is the Summary Statistics vector of the fixed effects of the traits analysed; Z is the matrix which relates to the animal effect; u is Summary statistics for the ultrasound scan the vector of the random animal effects; and e is measurements and carcass traits are shown in Table the vector of residual effects for the traits analyzed. 1. Comparing scanning systems, CUP consistently The expectations and variance matrices for random produced more variation as indicated by higher vectors are described as standard deviations and higher coefficients of vari-     ation. This was most noticeable for the IMF ultra- ñ ô ñ ô ñ ô y Xb u G A ⊗ G 0 sound scan trait, for example, 100dC-IMF had a     u 0 E = ; V = = e R 0 I ⊗ R 0.8 higher standard deviation and 16.5% higher co- e 0 efficient of variation compared to 100dP-IMF. Across all ultrasound scan trait traits, there were Where G and R denote the 2  × 2 matrices con- consistently fewer animals measured using the CUP taining additive genetic and residual variance com- system mainly due to the image capture technology ponents; A is the numerator relationship matrix; I not being available for some scanning events. There Translate basic science to industry innovation Comparison of ultrasound to predict carcass was also a decrease in the number of steers between For subcutaneous fat traits of ultrasound RIB scanning and harvest due to the normal attrition and RUMP, the heritabilities were similar across during the lot feeding and pre-harvest phase. This systems, with the exception being the higher herit- was also a function of the time between scanning ability observed for InC-RUMP at feedlot intake at events and harvest with the steers being an average 0.52 compared to 0.38 for InP-RUMP. of 511 d of age at feedlot intake, 614 d of age for The heritability estimates for InC-IMF and the second ultrasound scanning event, and 796 d of 100dC-IMF were noticeably higher than found in age at harvest. previous studies. For example, Walkom et al. (2015) obtained heritability estimates for ultrasound scan IMF of 0.28 in heifers and 0.20 for bulls, based on Variance Component and Heritability Estimates phenotypes collected mainly on the PIE system in Variance components and heritability estimates primarily Angus breeding animals. Similarly, Kelly for the intramuscular fat and marbling traits, eye et  al. (2019) estimated the heritability of ultra- muscle area, rib fat and rump fat are shown in sound scan IMF as 0.25 from a combined dataset Tables 2, 3, 4, and 5, respectively. of bulls, steers and heifers measured using the PIE All heritabilities were moderate to high, con- system. The heritability estimates of InC-IMF and firming that ultrasound scan and direct carcass 100dC-IMF from the current study were similar to traits provide valuable information for genetic the estimates from Kemp et al. (2002) of 0.51. This evaluation of beef cattle. This finding is consistent is a more comparable study as it was undertaken on with previous studies that have estimated vari- Angus steers, rather than bulls or heifers, and with ance components and heritabilities of ultrasound the ultrasound scan images interpreted in a labora- scan traits and direct carcass traits on Angus and tory setting rather than crush-side in real-time. Angus influenced beef cattle populations (Reverter The heritability estimates for InP-IMF and et al., 2000; Kemp et al., 2002; Boerner et al., 2013; 100dP-IMF in the current study were closer to Walkom et al., 2015; Kelly et al., 2019). most previous studies from Australian cattle popu- Comparing the two ultrasound scanning sys- lations, particularly heifers, which was expected tems, CUP resulted in higher phenotypic and gen- as most phenotypic data analyzed in those stud- etic variances compared to PIE for all ultrasound ies were based on the PIE ultrasound technology. scan traits of IMF, EMA, RIB, and RUMP at both Estimates of IMF from bull phenotypes from the steer feedlot intake and after 100 d of feeding. previous studies found lower heritability which is Heritability estimates of InC-IMF and likely to be the result of lower mean intramuscular 100dC-IMF were consistently higher than InP-IMF fat, and therefore genetic differences expressed to a and 100dP-IMF, with a 0.14 increase at steer feedlot lesser degree in bulls compared to heifers and steers intake and a 0.09 increase after 100 d of feeding. (Reverter et al., 2000; Boerner et al., 2013; Walkom Ultrasound EMA, at both feedlot intake and 100 et al., 2015). d of feeding, were the only group of traits in which For ultrasound scan EMA, this study showed the heritabilities were higher for PIE (0.52 for InP- higher heritability compared to previous stud- EMA compared to 0.40 for InC-EMA and 0.46 ies (Reverter et  al., 2000; Kemp et  al., 2002; for 100dP-EMA compared to 0.43 100dC-EMA). Boerner et  al., 2013; Kelly et  al., 2019). For the Table 2.  Heritabilities, additive genetic variances, phenotypic variances, genetic, and phenotypic correl- ations for IMF and carcass marbling traits (standard errors in parenthesis) Variance/trait InP-IMF InC-IMF 100dP-IMF 100dC-IMF C-IMF C-AMBL C-MMBL h 0.37 (0.08) 0.51 (0.09) 0.45 (0.09) 0.54 (0.09) 0.62 (0.09) 0.42 (0.09) 0.46 (0.09) σ a 0.25 0.73 0.13 1.16 5.91 0.57 5,872 σ p 0.68 1.45 0.29 2.13 9.46 1.35 12,794 InP-IMF – 0.79 (0.09) 0.73 (0.10) 0.71 (0.11) 0.64 (0.11) 0.45 (0.14) 0.46 (0.14) InC-IMF 0.34 (0.03) – 0.78 (0.10) 0.98 (0.06) 0.75 (0.09) 0.59 (0.12) 0.64 (0.12) 100dP-IMF 0.39 (0.02) 0.30 (0.02) – 0.76 (0.09) 0.59 (0.10) 0.62 (0.12) 0.63 (0.12) 100dC-IMF 0.30 (0.02) 0.49 (0.02) 0.43 (0.02) – 0.66 (0.09) 0.68 (0.10) 0.74 (0.09) C-IMF 0.27 (0.03) 0.36 (0.02) 0.28 (0.03) 0.43 (0.02) – 0.97 (0.04) 0.96 (0.03) C-AMBL 0.19 (0.03) 0.27 (0.03) 0.24 (0.03) 0.34 (0.03) 0.56 (0.02) – 0.99 (0.01) C-MMBL 0.21 (0.03) 0.30 (0.03) 0.25 (0.03) 0.38 (0.03) 0.62 (0.02) 0.94 (0.01) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Translate basic science to industry innovation Duff et al. Table 3.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for EMA traits (standard errors in parenthesis) Variance/trait InP-EMA InC-EMA 100dP-EMA 100dC-EMA C-EMA h 0.52 (0.09) 0.40 (0.09) 0.46 (0.09) 0.43 (0.08) 0.60 (0.10) σ a 10.01 12.27 13.48 24.55 37.91 σ p 19.43 30.82 29.41 56.85 62.92 InP-EMA – 0.94 (0.06) 0.80 (0.07) 0.92 (0.06) 0.83 (0.07) InC-EMA 0.76 (0.01) – 0.84 (0.08) 0.94 (0.07) 0.90 (0.07) 100dP-EMA 0.52 (0.02) 0.49 (0.02) – 0.94 (0.04) 0.86 (0.07) 100dC-EMA 0.51 (0.02) 0.50 (0.02) 0.71 (0.01) – 0.78 (0.08) C-EMA 0.35 (0.03) 0.36 (0.02) 0.38 (0.03) 0.38 (0.03) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Table 4.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for Rib Fat traits (standard errors in parenthesis) Variance/trait InP-RIB InC-RIB 100dP-RIB 100dC-RIB C-RIB h 0.42 (0.08) 0.44 (0.09) 0.50 (0.09) 0.59 (0.10) 0.40 (0.09) σ a 0.56 1.24 1.70 4.47 10.35 σ p 1.35 2.83 3.39 7.56 26.00 InP-RIB – 0.98 (0.03) 0.75 (0.08) 0.58 (0.10) 0.42 (0.08) InC-RIB 0.75 (0.01) – 0.75 (0.08) 0.58 (0.11) 0.33 (0.15) 100dP-RIB 0.48 (0.02) 0.49 (0.02) – 0.83 (0.04) 0.58 (0.11) 100dC-RIB 0.40 (0.02) 0.42 (0.03) 0.75 (0.01) – 0.60 (0.11) C-RIB 0.33 (0.03) 0.26 (0.03) 0.41 (0.02) 0.41 (0.03) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. Table 5.  Heritabilities, additive genetic variances, phenotypic variances, genetic and phenotypic correl- ations for Rump Fat traits (standard errors in parenthesis) Variance/trait InP-RUMP InC-RUMP 100dP-RUMP 100dC-RUMP C-RUMP h 0.38 (0.08) 0.52 (0.10) 0.61 (0.10) 0.61 (0.10) 0.50 (0.09) σ a 0.99 1.59 4.97 5.78 14.71 σ p 2.58 3.04 8.12 9.52 29.10 InP-RUMP – 0.80 (0.09) 0.85 (0.06) 0.84 (0.06) 0.55 (0.11) InC-RUMP 0.85 (0.01) – 0.86 (0.06) 0.87 (0.07) 0.55 (0.11) 100dP-RUMP 0.62 (0.02) 0.61 (0.02) – 0.99 (0.05) 0.75 (0.07) 100dC-RUMP 0.59 (0.02) 0.59 (0.02) 0.94 (0.01) – 0.71 (0.07) C-RUMP 0.41 (0.02) 0.41 (0.02) 0.55 (0.02) 0.55 (0.02) – For traits genetic correlations above diagonal, phenotypic correlation below diagonal. subcutaneous fat ultrasound scan traits of RIB of C-RIB and C-RUMP the heritability esti- and RUMP, the heritability results were like pre- mates where similar to the associated ultrasound vious studies. scan traits. The heritability estimates for the carcass marb- The heritability estimates for the carcass traits ling and EMA traits were higher than the scan were generally higher in the current study com- traits with estimates of 0.62, 0.42, 0.46, and 0.60 pared to some previous reports. For example, for C-IMF, C-AMBL, C-MMBL, and C-EMA, re- Borner et al. (2013) estimated heritabilities for car- spectively. For the marbling traits, the higher her- cass IMF, carcass rump fat, carcass rib fat and car- itability for C-IMF is expected given the objective cass eye muscle area of 0.33, 0.36, 0.23, and 0.39, NIR assay used to precisely measure this trait, respectively. In the current study, steers were killed compared to the subjective scoring by a human at an older age and higher carcass weight resulting grader and categorical nature of both C-AMBL in higher means and variances for all carcass traits. and C-MMBL. For the subcutaneous fat traits However, several other reports, based on similar Translate basic science to industry innovation Comparison of ultrasound to predict carcass cattle and production systems, showed comparable heritabilities to this study. For example, the Torres- Vázquez et al. (2018), Jeyaruban et al. (2017), and Kemp et  al. (2002) obtained heritability estimates for carcass IMF at 0.61, MSA marbling score at 0.48 and USDA marbling score at 0.40, respectively. In most genetic evaluation programs, it is more common for bulls and heifers to be ultrasound scanned for the correlated carcass traits, rather than steers as in the current study. For example, from the 593,376 ultrasound scan IMF records on the Angus Australia database, 49.8% are from bulls, 45.6% from heifers, and 4.6% from steers (A Byrne, Angus Australia 2020, pers. comm., January 28, 2020). While this should be a consideration in the interpretation and application of the results from this study, a similar study (Duff et  al., 2018) of combined steer and heifer data showed comparable results to this study, particularly the higher herit- ability for CUP IMF compared to PIE IMF. It is common practice to combine heifer and steer ultrasound scan data for parameter estima- tion and genetic evaluation. For example, Walkom et  al. (2015) observed substantially higher genetic Figure 1. Distribution of ultrasound scan intramuscular fat (IMF) steer phenotypes for the PIE (top) and CUP (bottom) systems at 100 variance and heritabilties for ultrasound scan IMF d feeding. from the combined heifer and steer phenotypes, compared to bull phenotypes. There are no known previous reports where Angus bulls have been ultra- significant. For example, the genetic correlation sound scanned for IMF, EMA, RIB, and RUMP with C-IMF was 0.11 higher for InC-IMF and with both the PIE and CUP systems. 0.07 higher for 100dC-IMF compared to the PIE estimates at the same event. A  possible explan- Genetic Correlations ation for this may be the narrower range of IMF prediction for PIE compared to CUP. The PIE Genetic correlations for the intramuscular system, and its in-built algorithm used to predict fat and marbling traits, eye muscle area, rib fat IMF, is known to be most effective between 2.0% and rump fat are shown in Tables 2, 3, 4, and 5, and 8.0% IMF range (R. Evans, Bovine Scanning respectively. Services Pty Ltd, pers. comm., January 6, 2021). The genetic correlations between ultrasound For this reason, we observed few records in this scan traits and the direct breeding objective car- study that are less than 2.0% or greater than 8.0% cass traits presented were positive and moderate from PIE, compared to CUP, particularly in the to strong. This is consistent with previous studies 100-d scan (Figure 1) where we expect to observe (Reverter et  al., 2000; Kemp et  al., 2002; Borner a higher proportion of IMF values greater than et  al., 2013; Walkom et  al., 2015), showing that 8%. The CUP system can predict IMF to a wider ultrasound scanning is a valuable indirect carcass range (Table 1) and can more precisely determine measurement for informing genetic evaluation pro- genetic merit by explaining greater genetic vari- grams of beef cattle. ance (Table 2). This is evident and consistent at Comparing the two ultrasound scanning sys- both steer intake and 100 d on feed. This finding tems, the genetic correlations of ultrasound scan also highlights the difference observed in the gen- IMF and the breeding objective carcass IMF etic correlations between the same ultrasound (C-IMF) and marbling traits (C-AMBL and scan system at steer intake and 100 d feeding, C-MMBL) tended to be higher for CUP compared being 0.98 for CUP and 0.73 for PIE. In con- to PIE, at both ultrasound scanning events, how- trast, the genetic correlations with the breeding ever, also acknowledging that the standard errors objective carcass traits tended to be similar when of these estimates indicate the differences are not Translate basic science to industry innovation Duff et al. comparing CUP to PIE for the ultrasound scan fat and rump fat are listed in Tables 2, 3, 4, and 5, traits of EMA, RIB, and RUMP. respectively. The genetic correlations for the ultrasound IMF Like the genetic parameters, the phenotypic traits with the carcass IMF and marbling traits were correlations for the CUP system for IMF to generally stronger than reported in previous studies. C-IMF, C-AMBL, and C-MMBL were higher For example, Reverter et al. (2000), obtained genetic than for the PIE system. While for ultrasound correlation estimates for carcass IMF to bull IMF scan EMA, RIB, and RUMP the phenotypic ultrasound of 0.47 and heifer IMF ultrasound of correlations to the associated breeding objective 0.46. Kemp et al. (2002), reported a much stronger traits tended to be similar between systems. An genetic correlation of 0.90 in steers, but the time exception being the 0.33 correlation for InP- interval in that study was much shorter between RIB with C-RIB, compared to 0.26 for InC-RIB ultrasound scanning and harvest of i.e. 52 d, com- with C-RIB. pared to a 285 and 182-d interval in the current study. Herring et al. (1998) reported stronger pheno- For EMA, the genetic correlations of PIE and typic correlations for the ultrasound predicted CUP ultrasound to carcass EMA were stronger in IMF traits with carcass IMF and marbling score this study than those reported in previous studies across four different ultrasound systems, including (Kemp et  al., 2002; Reverter et  al. 2000; Borner CUP (described as CVIS) and PIE in crossbred et al., 2013). The high correlations observed in the beef steers. They reported phenotypic correlations current study may have been a function of the use for CUP IMF with carcass IMF and marbling of highly experienced ultrasound scanning techni- score of 0.61 and 0.74, respectively. While for PIE cians on the live steers, experienced carcass grad- IMF to carcass IMF and marbling score, the re- ers in the abattoir and controlled data collection, ported estimates were 0.31 and 0.39, respectively. whereas field data from large scale bull breeding The contrasting results between studies is likely to herds were mostly utilized in other studies. be due to the different time intervals between ultra- In contrast, for the fat traits, genetic correlations sound scanning steers and their harvest followed of ultrasound rib and rump fat to the respective by carcass data collection. The interval was much carcass measures were weaker than those observed shorter in the Herring et  al. (1998) study ranging in previous studies. The possible reason being unin- from 8 to 14 d. tended abattoir effects, such as hide puller damage The results from the current study are more on the fat distribution on the long-fed steer car- likely to reflect industry practice, as phenotypic se- casses. With more subcutaneous fat observed on lection (e.g., drafting pre-harvest) with short time the long-fed steer carcass compared to shorter fed intervals between ultrasound scanning and harvest steers, there is higher probability of damage to the is unlikely due to the associated stressors having subcutaneous fat which may lead to reduced pre- negative impacts on meat quality through dark cut- cision of measurement in the chiller. For example, ting (Ponnampalam et al., 2017) or welfare implica- in this study steers averaged 18.7  mm for carcass tions of increased injury risk. It is more practical, rib fat at an average of 460.2 kg carcass weight. In and therefore more likely, to ultrasound scan ani- contrast, in the Reverter et al. (2000), Borner et al. mals on-farm or at feedlot induction, well before (2013) and Kemp et  al. (2002) studies the carcass harvest. rib fat measurements were leaner at 6.2, 9.0, and For the carcass traits, the correlations between 14.1 mm, respectively, and with lighter carcasses. C-IMF with C-AMBL and C-MMBL were mod- The genetic correlations between the three car- erate at 0.56 and 0.62, respectively. These esti- cass traits of C-IMF, C-AMBL, and C-MMBL mates were lower than those reported by Lee et al. were very strong and positive ranging from 0.96 to (2019) of 0.87 in a different breed and produc- 0.99. A study with temperate beef cattle by Johnston tion system having higher mean carcass IMF and (2001) showed similar genetic correlations between marbling scores with greater variability. Konarska carcass IMF to MSA Marbling and AUS-meat et al. (2017) reported a closer correlation between marbling score of 1.00 and 0.96, respectively, sup- MSA marbling score and carcass IMF by NIR in porting the findings in this study. M. longissimus thoracis, the same muscle as meas- ured in the current study, of 0.75. The phenotypic correlation of C-AMBL with C-MMBL was high Phenotypic Correlations at 0.94. This was expected as both scores were Phenotypic correlations for the intramus- assessed by the same grader, albeit on different cular fat and marbling traits, eye muscle area, rib scales. Translate basic science to industry innovation Comparison of ultrasound to predict carcass Breeding Program Design CUP, compared to the crush-side and real-time process of PIE. Comparing ultrasound scan methods to predict To benefit from the results of this study, beef carcass traits is an important step in understanding cattle genetic evaluation programs could consider strategies, particularly breeding program design, transitioning all live animal ultrasound phenotype aimed at increasing accuracy of selection and gen- recording to the CUP system or similar systems etic gain for carcass traits in breeding objectives. In using a centralized processing approach and predic- a companion study (Duff et al., 2019), we modeled tion algorithms. An alternative, but more complex, several phenotyping and genotyping scenarios fo- approach is to receive ultrasound phenotypes from cused on the breeding objective traits of C-IMF, a range of systems (e.g., both PIE and CUP) and C-AMBL, and C-MMBL. The study investigated model each specifically to recognize the differences how breeding programs may be enhanced by using in trait variances, heritabilities and genetic correl- genomic-based information as derived from a refer- ations to the breeding objective traits. ence population with direct carcass IMF and marb- Consistent with Kemp et al. (1998), this study ling score phenotypes coupled with genotypes, as confirmed that ultrasound scanning can be used described by Goddard et  al. (2010). This study to effectively predict carcass phenotypes in Angus found the highest rates of selection accuracy and steers, including those from progeny test programs, response would be achieved through a combination to inform genetic evaluation programs, particularly of CUP ultrasound scan phenotyping for IMF where collection of effective carcass data from the and genotyping with a reference population of re- abattoir is not possible or difficult. lated animals with carcass IMF and marbling score A unique feature of this study was the inclusion phenotypes. However, the value of ultrasound scan of three measurements of marbling traits on each phenotyping diminishes as the GBV prediction ac- carcass, being C-IMF, C-AMBL, and C-MMBL. curacy increases, which is mainly a function of the The results indicate that these measures are all reference population size. strongly and positively correlated, both phenotyp- ically and genetically. As a result, for beef cattle CONCLUSIONS genetic evaluation, the collection of just one of the marbling traits is likely to be sufficient. Additional This study compared the phenotypic and gen- benefit is attained from measuring C-IMF due to etic parameters for the ultrasound scan measured the higher heritability of this trait and stronger gen- traits (IMF, EMA, RIB, and RUMP) for two etic correlations with live animal ultrasound scan live-animal ultrasound systems (PIE and CUP) IMF, but there are added cost and sample collec- and estimated their relationship with the direct car- tion considerations associated with the C-IMF cass traits. The results showed substantial genetic phenotype. variation in carcass performance can be measured It is also recognized that the ultrasound scan- using either ultrasound scan system, even when ning hardware used in the study, which is still com- there is a considerable interval (e.g., 285 d) between monly used to phenotype live animals for carcass the ultrasound scanning event and harvest. This traits for genetic evaluation, was developed in the is based on the moderate to high heritabilities ob- 20th century. More sophisticated ultrasound scan served, coupled with moderate to strong relation- systems are available today which can capture ships with the related breeding objective carcass higher quality images and potentially predict more traits. A noticeable difference was the CUP system precise phenotypes, if coupled with appropriate explaining more variation, particularly for ultra- prediction algorithms. A  study to understand po- sound scan IMF, resulting in a higher heritability tential benefits of modern ultrasound scan sys- and stronger correlations with the carcass IMF and tems for beef cattle genetic evaluation programs is marbling traits in the breeding objective. This indi- recommended. cates that the CUP system, compared to PIE, pro- Further research is also warranted to under- vides an advantage for genetic evaluation of carcass stand genotyping and phenotyping strategies traits in Angus cattle. This advantage should be for beef herds with carcass traits included in the considered with knowledge of possible additional breeding objective. As it is more common for costs involved with interpreting ultrasound images breeding candidates to scanned for the correlated through a centralized laboratory. Furthermore, carcass traits, rather than steers as in this study, the there is also a turn-around time of 24–48  h in re- future research also needs to better understand the ceiving the phenotype measurement results from Translate basic science to industry innovation Duff et al. https://www.ausmeat.com.au/WebDocuments/Producer_ genetic and phenotypic relationship between bull, HAP_Beef_Small.pdf [accessed January 6, 2021]. heifer, and steer measurements. Börner,  V., D.  J.  Johnston, and H.  U.  Graser. 2013. Genetic relationships between live animal scan traits and carcass ABBREVIATIONS traits of Australian Angus bulls and heifers. Anim. Prod. Sci. 53: 1075–1082. doi:10.1071/AN12435 100dC-EMA, 100 days on feed eye muscle area Duff,  C., J.  H.  J.  van  der  Werf, and S.  A.  Clark. 2018. using CUP; 100dC-IMF, 100  days on feed intra- Comparison of two live-animal ultrasound systems muscular fat using CUP; 100dC-RIB, 100  days to predict carcass intramuscular fat and marbling in Australian Angus cattle. In: Proceedings of the 11th World on feed rib fat depth using CUP; 100dC-RUMP, Congress on Genetics Applied to Livestock Production 100  days on feed rump fat depth using CUP; No. Electronic Poster Session. Auckland, New Zealand. 100dP-EMA, 100  days on feed eye muscle area p. 262. http://www.wcgalp.org/system/files/proceed- using PIE; 100dP-IMF, 100 days on feed intramus- ings/2018/comparison-two-live-animal-ultrasound-sys- cular fat using PIE; 100dP-RIB, 100  days on feed tems-predict-carcass-intramuscular-fat-and-marbling.pdf rib fat depth using PIE; 100dP-RUMP, 100 days on [accessed February 29, 2020]. Duff, C., J. H J. van der Werf, and S. A. Clark. 2019. Should feed rump fat depth using PIE; ASBP, Angus Sire Angus breeders live-animal ultrasound scan for intramus- Benchmarking Program; BIN, Beef Information cular fat in the genomics era? Proc. Assoc. Adv. Anim. Nucleus; C-AMBL, carcass measured AUS-MEAT Breed. Genet. 23: 496–499. http://www.aaabg.org/aaab- marbling score; C-EMA, carcass measured eye ghome/AAABG23papers/122Duff23496.pdf [accessed muscle area; C-IMF, carcass measured intramus- March 1, 2020]. Gilmour, A. R., B. J. Gogel, B. R. Cullis, and R. Thompson. cular fat by near-infrared spectrophotometry; 2009. ASReml user guide release 3.0. Hemel Hempstead, C-MMBL, carcass measured Meat Standards UK: VSN Int. Ltd. Australia marbling score; C-RIB, carcass measured Goddard,  M.  E., B.  J.  Hayes, and T.  H.  Meuwissen. 2010. cold rib fat; C-RUMP, carcass measured hot P8 fat; Genomic selection in livestock populations. Genet. Res. CUP, Central Ultrasound Processing ultrasound (Camb). 92:413–421. doi:10.1017/S0016672310000613 scan system; C-WT, carcass measured hot carcass Herring,  W.  O., L.  A.  Kriese, J.  K.  Bertrand, and J.  Crouch. 1998. Comparison of four real-time ultrasound systems weight; EMA, eye muscle area; GBV, Genomic that predict intramuscular fat in beef cattle. J. Anim. Sci. Breeding Value; IMF, intramuscular fat; InC- 76:364–370. doi:10.2527/1998.762364x EMA, feedlot intake eye muscle area using CUP; Jeyaruban,  M.  G., D.  J.  Johnston, and B.  J.  Walmsley. 2017. InC-IMF, feedlot intake intramuscular fat using Genetic and phenotypic characterization of MSA index CUP; InC-RIB, feedlot intake rib fat depth using and its association with carcass and meat quality traits in Angus and Brahman cattle. Proc. Assoc. Adv. Anim. CUP; InC-RUMP, feedlot intake rump fat depth Breed. Genet. 22:313–316. http://www.aaabg.org/aaab- using CUP; InP-EMA, feedlot intake eye muscle ghome/AAABG22papers/71Jeyaruban22313.pdf [ac- area using PIE; InP-IMF, feedlot intake intramus- cessed August 10, 2019]. cular fat using PIE; InP-RIB, feedlot intake rib Johnston,  D.  J. 2001. Selecting for marbling and its relation- fat depth using PIE; InP-RUMP, feedlot intake ship with other important economic traits. What impact rump fat depth using PIE; MSA, Meat Standards does it have? In: Proceedings of Marbling Symposium, Coffs Harbour, Australia. p. 88–93. https://pdfs.seman- Australia; PIE, Pie Medical Esaote Aquila ultra- ticscholar.org/e8f9/70fa903a9ba929b83409db8acecb- sound Scan system; RIB, rib fat; RUMP, rump fat 2d2ac196.pdf [accessed March 11, 2020]. Kelly,  D.  N., M.  Murphy, R.  D.  Sleator, M.  M.  Judge, ACKNOWLEDGMENTS S. B. Conroy, and D. P. Berry. 2019. Feed efficiency and car - cass metrics in growing cattle. J. Anim. Sci. 97(11):4405– The research was supported by Meat and 4417. doi:10.1093/jas/skz316 Livestock Australia through MDC matching funds Kemp, D. J., W. O. Herring, and C. J. Kaiser. 2002. Genetic and environmental parameters for steer ultrasound and car- in project PSH.0528. Generous support has been cass traits. J. Anim. Sci. 80(6):1489–1496. Available from provided by numerous Angus bull breeders, co-op- http://search.proquest.com.ezproxy.une.edu.au/scholar- erator herds, supply chain partners, technicians, ly-journals/genetic-environmental-parameters-steer-ul- and research organizations. trasound/docview/218122729/se-2?accountid=17227 Conflicts of interest statement. The authors de- [accessed March 11, 2020]. Konarska,  M., K.  Kuchida, G.  Tarr, and R.  J.  Polkinghorne. clare that there are no conflicts of interest to disclose 2017. Relationships between marbling measures across that relate to the research described in this paper. principal muscles. Meat Sci. 123:67–78. doi:10.1016/j. meatsci.2016.09.005 LITERATURE CITED Lee, B., and Y. M. Choi. 2019. Correlation of marbling char- acteristics with meat quality and histochemical character- AUS-MEAT. 2020. Handbook of Australian Beef Processing, istics in longissimus thoracis muscle from hanwoo steers. Version 7. Murarrie, QLD: AUS-MEAT Ltd. Available from Translate basic science to industry innovation Comparison of ultrasound to predict carcass Food Sci. Anim. Resour. 39(1):151–161. doi:10.5851/ Reverter,  A., D.  J.  Johnston, H.  U.  Graser, M.  L.  Wolcott, kosfa.2019.e12 and W.  H.  Upton. 2000. Genetic analyses of live-ani- Parnell, P. F., C. J. Duff, A. I. Byrne, and N. M. Butcher. 2019. The mal ultrasound and abattoir carcass traits in Australian Angus sire benchmarking program—a major contributor to fu- Angus and Hereford cattle. J. Anim. Sci. 78:1786–1795. ture genetic improvement in the Australian beef industry. Proc. doi:10.2527/2000.7871786x Assoc. Adv. Anim. Breed. Genet. 23:492–495. http://www. Torres-Vázquez, J. A., J. H. J. van der Werf, and S. A. Clark. aaabg.org/aaabghome/AAABG23papers/121Parnell23492. 2018. Genetic and phenotypic associations of feed ef- pdf [accessed November 1, 2020]. ficiency with growth and carcass traits in Australian Perry,  D., W.  R.  Shorthose, D.  M.  Ferguson, and Angus cattle. J. Anim. Sci. 96:4521–4531. doi:10.1093/ J. M. Thompson. 2001. Methods used in the CRC program jas/sky325. for the determination of carcass yield and beef quality. Upton,  W.  H., K.  A.  Donoghue, H.  U.  Graser, and Aust. J. Exp. Agric. 41:953–957. doi:10.1071/EA00092 D.  J.  Johnston. 1999. Ultrasound proficiency testing. Polkinghorne,  R., J.  M.  Thompson, R.  Watson, A.  Gee, Proc. Assoc. Adv. Anim. Breed. Genet. 13:341–344. http:// and M.  Porter. 2008. Evolution of the Meat Standards www.aaabg.org/proceedings/1999/AB99079.pdf accessed Australia (MSA) beef grading system. Aust. J. Exp. Agric. August 10, 2019]. 48:1351–1359. doi:10.1071/EA07177 Walkom,  S.  F., M.  G.  Jeyaruban, and D.  J.  Johnston. 2015. Ponnampalam, E. N., D. L. Hopkins, H. Bruce, D. Li, G. Baldi, Impact of scanning lean cattle on the genetic correlation and A. E. Bekhit. 2017. Causes and contributing factors between scan and carcass intramuscular fat in Angus and to “Dark Cutting” meat: current trends and future direc- Hereford cattle. Proc. Assoc. Adv. Anim. Breed. Genet. tions: a review. Compr. Rev. Food Sci. Food Saf. 16:400– 21:253–256. http://www.aaabg.org/proceedings/2015/ 430. doi:10.1111/1541-4337.12258 Walkom21253.pdf [accessed August 10, 2019]. Translate basic science to industry innovation

Journal

Translational Animal ScienceOxford University Press

Published: Jan 25, 2021

Keywords: Angus; beef cattle; carcass; genetic parameters; phenotypic parameters ultrasound

References