Ultrasound proficiency testing.

Abstract

Proc. Assoc. Advmt. Anim. Breed. Genet. Vol13 ULTRASOUND PROFICIENCY TESTING W. H. Upton, K. A. Donoghue, H. -U. Graser, and D. J. Johnston NSW 2350 Animal Genetics and Breeding Unit **,University of New England, Armidale, SUMMARY Scanning of live beef cattle for the estimation of breeding values for carcase traits is an important facet of the selection of beef cattle through BREEDPLAN EBVs. Scans are collected by real time ultrasound contractors who must undergo a proficiency test prior to their data being accepted by the breed societies for analysis. May `97 and, June `98 proficiency tests included scanning for intramuscular fat as well as subcutaneous fat and eye muscle area. Proficiency testing of operators has led to confidence in the measurement technique ahd rapid adoption of the technology for genetic evaluation. Keywords: Beef cattle, carcase, ultrasound scanning INTRODUCTION Since 1989, real time ultrasound measurements of subcutaneous fat and eye muscle area (EMA) have been successfully used for Australian genetic evaluations of beef cattle. Scanning for percent intramuscular fat (PIMF) has been researched at AGBU for the last five years via a Meat and Livestock Australia (MLA) supported research project. Heritabilities of scanned PIMF are estimated to be 0.297 (I 0.029) for heifers and 0.12 1 (I 0.027) for bulls (Reverter 1999). To maintain the standard of data records for use in BREEDPLAN the contract scanners who have taken the measurements have been required to undergo proficiency tests on a regular basis. Accreditation is awarded, based on satisfactory proficiency, by the Performance Beef Breeders Association (PBBA) who represent the major Australian breeds undertaking BREEDPLAN genetic evaluations. This accreditation allows those scamem who have received accreditation to submit data into the individual breed society data bases for analysis by BREEDPLAN. This paper reports the conduct and results from the two most recent proficiency tests. These were heavily subsidised by MLA Validation projects that were designed to research and develop new traits for BREEDPLAN. In the last five years considerable emphasis has been on the development of techniques to assess intra-muscular fat and these proficiency tests were essential to introduce the technology to the seedstock industry. CONDUCT OF THE PROFICIENCY TESTS Thirty cattle of different age and sex were accumulated for each of the two tests. Accreditation was attempted for ukrasound recording of rump and rib fat depth, eye muscle area (EMA), eye muscle depth (EMD) and intramuscular fat. All animals were scanned twice by each scanner in two separate * AGBU is a joint institute of NSW Agriculture and the University of New England 341 Proc. Assoc. Advmt. Anim. Breed. Genet. Vol13 runs with identities of the cattle changed between runs to conceal the real identity from the scanners. The cattle were slaughtered within three days of scanning. Thirteen scanners attempted accreditation in May `97 (3 using PIE and 10 using Aloka ultrasound The majority of scanners who sat for machines) and 16 in June `98 (6 PIE and 10 Aloka). accreditation in `97 re-sat the test in `98 because of changes to the PIMF scanning technology. Two previous Aloka operators switched to using PIE equipment. CARCASS RESULTS Carcass data were recprded in the chiller the day after slaughter, by three different AGBU staff in the chiller. All measurements were taken for both the left and right sides of the carcass. This effectively meant that six records were taken for each of the traits evaluated. Eye muscle areas were traced on acetate sheets by th' three independent operators. EMA was calculated using a computerised digitizer by an AGB e staff member. A cross sections slice of the'longissimus dorsi (eye muscle) approximately 10 mm thick was taken from the forequarter and analysed for percent intramuscular fat (PIMF) by the Meat Science Laboratory of UNE using replicated Soxlet extraction procedure. All carcass data were averaged over the six records for comparison with the scanner data. Table 1 presents a summary of the carcass data. , Table 1. Summary qf carcass data May `97 and Juae `98 proficiency tests I Trait ~ P8 (mm) Rib fat (IIUQ) EMA (d)i PIMF May 1997 Range 2.8-14.7 2.2-12.5 53.2-75.0 0.28-6.50 Mean 8.5 6.1 63.0 1.95 SD 3.25 2.58 6.2Q 1.34 June 1998 Range 4.7-18.0 1.7-12.2 58.4-88.8 2.0-7.75 Mean 9.8 5.6 71.0 3.95 SD 2.99 2.99 9.09 1.31 The May `97 test us d only grass, fed cattle but the &me `98 test had cattle specifically prepared for the course. It shoul be noted that eye muscle area and intra-muscular fat were higher in the June `98 $ test but the subcutaneous fats were similar. The standard deviation of EMA was also larger for `98 than for the `97 test. RESULTS FOR SQWNERS Scanners are assess d for their ability to be repeatable and for their relationship to the cat-case values. Repeatability is test d using the standard deviation of the difference between repeated scans on the i same animals. The relationship to the carcase is tested using both as the standard deviation of the difference between live scan measurements and the carcase values as well as the correlation between live and carcase results. Repeatability. The atamlard deviations between first and second scans on the same animal for PIMF in the June `98 testi were considerably smaller than for the May `97 test. Three of the more likely reasons being that the scanners were more experienced, scanners were advised to use the average of 5 scans rather than the 3, as in `97 and the cattle had higher mean PIMF values. Both the Aloka (ISU software) and the PIE systems appear to work best between about 2 and 8 % PIMF. 342 Proc. Assoc. Advmt, Anim. Breed. Genet. Voll3 Table 2. Standard deviations of the difference between first and seconds scans - May `97 and June `98 proficiency tests Profit Level' Mav 1997 M& June 1998 Mean 1.07 0.93 3.71 Trait P8 mm Rib fat mm EMA cm* PIMF Min aProficiency 0.60 5 1.5 1.25 5 1.0 0.60 0.96 1.50 5 6.0 3.61 0.59 < 1.1 0.95 Level = acceptance level for proficient scanners Max 2.30 1.80 6.40 Min 0.58 0.55 1.59 Max 1.82 1.34 6.42 1.19 0.68 0.33 1.65 Accuracy with Carcase Measures. Correlations of scans with mean carcass measurements ( avei-age of six measures for each carcase: 3 abattoir measurers; left and right) are presented in Table 3 and the standard deviation of the difference between live and scan in Table 4. Table 3. Correlations of scans with mean carcass measurements - May `97 and June `98 Profit Level a > 0.90 > 0.90 > 0.80b May 1997 Mean 0.87 June 1998 Mean 0.87 Trait P8 mm Rib fat mm EMAcm' Min 0.80 Max 0.92 Min 0.75 Max a90 0.84 0.81 0.92 0.42 0.65 0.82 h 0.7ob PIMF' 0.71 0.63 0.83 aProficiencyLevel = acceptancelevel for proficient scanners b Proficiency 0.88 0.85 0.77 0.73 0.77 0.72 0.96 0.90 0.83 levels were adjusted in 1997 to 0.65 for EMA and 0.50 for PlMF due to lower mean and standard deviations Table 4. Standard deviation of difference of (Scan-Carcass-Bias) - May `97 and June `98 tests Protic May 1997 Min Level a Mean 5 1.5 2.01 1.6 5 1.0 1.37 1.00 5 5.5 5.44 3.60 < 1.0 1.07 0.77 = acceptance level for proficient scanners June 1998 Mean 1.51 1.48 4.91 0.93 ' Trait PfXmm Rib fat mm EMA cm2 PlMF Proficiency Level Max 2.4 2.20 7.30 1.32 Min 1.17 0.82 3.96 0.82 Max 2.09 2.07 5.89 1.41 In the June `98 test the correlation between scans and carcase for EMA and PIMF are considerably higher than results for May `97. The cattle used in May `97 had lower mean EMA and a smaller standard deviation. The required level to pass the `97 proficiency test was adjusted to account for the lower mean and standard deviation. The mean value of PIMF in May `97 was also low and this almost certainly contributed to the lower correlations between scan and carcase. Taking the `98 results into account, and provided that a similar standard deviation of carcase PIMF can be obtained in future test animals, it has been recommended that the threshold for the correlation between scan and carcase PIMF be increased to 0.75. Proficiency was awarded for all three traits (Fat, EMA and PIMF) to 7 scanners in the May `97 test and 10 scanners in the June `98 test. Other scanners were accredited for individual traits. 343 hoc. Assoc. Advmt. Anim. Breed. Genet. Vol I3 DISCUSSION Proficiency testing of scanners for submitting data to BREEDPLAN has resulted in high credibility of the data and the resulting EBVs. Scanned data are the highest cost of any data collection currently used to produce EBVs in BREEDPLAN yet the growth in scanned data has increased dramatically as demonstrated by the Angus statistics with 10,000 scans recorded in 1993 jumping to 53,000 analysed in the 1999 group BREEDPLAN. Proficiency standards for EMA and subcutaneous fats have remained the same from the first proficiency test in 1989 and field research has shown that heritabilities of carcass traits based on Standards for PIMF have been set only in the scans taken by the accredited scanners is moderate. last two years but the first estimates of heritability for scanned PIMF are also moderate. The mean and the range of trait values for the cattle used in the test are critical if repeated tests are to be equitable as the statistics used can be influenced by the value. This is especially evident in the PIMF trait where the current algorithms used in the computation are calibrated for a relatively small range of values. Knowledge of the background of the cattle and pre-course preparation in a feedlot is almost certainly necessary to ensure suitability. There are a number of issues that need to be addressed for future testing. Currently the PBBA stipulates that currency of accreditation requires a test every three years. To date the tests (with the exception of one) have been conducted at Armidale but have been subsidised by research funds as As the system becomes strictly user pay, they have been essential to the research objectives. relatively high costs will be incurred by the scanners to attend the proficiency tests. The estimate from the June `98 test was an overhead cost to run the course of $1000 per participant which did not include salaries for organisers nor did it include travel and accommodation for scanners. Currently no formal training is available for new and aspiring scanners. `on the job' by the contract scanners who are accredited. This has been entry of new scanners into the industry as contractors are disinterested have trained only new members of their team. If the requirement for the current rate the issue of training and proficiency testing may need to REFERENCES Reverter, A. (1999) Proc. Assoc. Advmt. Anim. Breed. Genet. 13:377 Training has been conducted successful but is limiting the in training competition and scanning continues to grow at be addressed. 344