ICAR Conference 2017 Programme Abstracts

Transcription

1 ICAR Conference 2017 Programme s Sheep, Goat and Camelid Session Measuring out of season breeding ability in dairy goats S Desire 1, S Mucha 1,2, M Coffey 1, R Mrode, 1,3, J Broadbent 4, J Conington 1 1 Animal and Veterinary Sciences, Scotland s Rural College, Easter Bush, Midlothian EH25 9RG, Scotland, United Kingdom 2 Poznan University of Life Sciences, 33 Wolynska, Poznan, Poland 3 Animal Biosciences, International Livestock Institute, Naivasha Rd, Nairobi, Kenya 4 Yorkshire Dairy Goats, St Helen s Farm, Seaton Ross, Yorkshire, YO42 4NP Producing a consistent, year-round supply of fresh milk can be challenging in goats, which have highly seasonal breeding cycles. There is within-breed variation in the length and timing of oestrus, suggesting there is potential to select for shifted, or extended breeding cycles. Recording out of season (OOS) breeding capability is difficult, as it is impractical to log the onset and duration of oestrus in individual animals in a large, naturally bred milking herd. The aim of this study was to develop a method of recording out of season kidding ability in dairy goats using routinely recorded data, and to estimate genetic parameters for this trait. Does were group housed indoors with mature bucks for between 60 and 120 days, then tested for pregnancy via trans-abdominal ultrasound. Goats that failed to become pregnant during this period were regrouped and given a further mating opportunity. First kidding (parturition) dates from 9,546 goats, bred across two farms between the years of 1992 and 2013, were used to define peak kidding season (PKS), which was a 4-week period from the end of February. The out of season phenotype was the number of weeks a doe first kidded, relative to PKS. A 14 generation pedigree containing 12,617 animals (231 sires, 7,201 dams) was used for genetic parameter estimations. Breeding values were estimated using an animal model that included the fixed effects described below, as well as a random animal effect. The average age at first kidding was 15.5 months (SD 3.0), which means the average age at first conception was approximately 10 months. As the time of year a goat is born will affect whether she is given the opportunity to breed out of season, herd-year-season of birth was fitted as a fixed effect in the model. Likewise, a goat that kids at an older age may have had several unsuccessful attempts to breed, therefore age at first kidding was also fitted as a fixed effect to adjust for this. A total of 1,818 (19%) of goats kidded for the first time during PKS. A total of 654 goats kidded one week either side of PKS, which represented a 12% decrease in kidding rate. Only 610 goats (6%) kidded between 20 and 24 weeks outside of PKS. Out of season kidding was found to have a low but significant heritability of 0.11 (s.e. 0.02), suggesting selection for this trait may be possible. Out of season kidding as defined in this study offers a proxy measure of aseasonal kidding ability, using routinely recorded information. Keywords: Dairy goats, aseasonal breeding, genetics

2 Potential for a routine health and robustness monitoring in dairy goats P. Herold 1, M Wolber 1, H. Kettnacker 1 & K. Droessler 2 1 Landesamt BW, Stuttgarter Str. 161, Kornwestheim, Germany 2 LKV Baden-Wuerttemberg, Heinrich-Baumann-Str. 1-3, Stuttgart, Germany Dairy goat farming in Southern Germany is dominated by organic goat breeders and keepers. Breeding value estimation is established since 2013 for dairy traits. In frame of developing an organic breeding program, improving functional traits is within the focus. In dairy cattle breeding, health monitoring systems have developed over several years. Combining the experiences of existing health monitoring in dairy cattle with the result of stakeholder interviews, a health and robustness monitoring for goats is currently set up in Baden- Wuerttemberg and Bavaria, Germany. The monitoring system will be based on an ADED standard code which will be developed in the frame of the project. Setting up the standard codes is a joint project of animal scientists, veterinarians as well as members of the breeding and the performance recording association. The standard code will contain diagnoses, observations, diagnostic findings and single animal or herd (preventive) treatments. Besides animal diseases it will also be possible to record postpartum observations on kids and does, horn-status and horn-related injuries. Additionally, behavior traits like targeted meanness or nervousness will be recorded. The codes will be implemented in the existing database Ziegendatenverbund (ZDV = goat data network). Within the farmer application ZDV4M the farmers themselves will be able to enter health, behavior and robustness related observations. Also, breed wardens will be able to enter diagnoses by the veterinarians. The collected data will be directly analyzed for management feedback to the farmers. It is expected to be a powerful tool for farmers to improve herd health. In future, cumulated data will be objective of further development in breeding value estimation. Parallel, breeding value estimation for conformation traits is developed and the combination of information from both phenotyping systems seems promising for a reliable genetic evaluation of the goat population. Keywords: dairy goat, health monitoring, ADED standard code

3 Phenotyping tools for genomic selection to reduce footrot in Texel sheep A.McLaren 1, W. Sawday 2, K. Kaseja1, J. Yates 2 & J. Conington 1 1 SRUC, Animal & Veterinary Sciences, Easter Bush Campus, EH25 9RG, Edinburgh, UK 2 UK Texel Sheep Society, Stoneleigh Park, Kenilworth, Warwickshire, CV8 2LG, UK Footrot, an endemic disease caused by Dichelobacter nodosus, is one of the most common causes of lameness in sheep and has been a major challenge to farmers for generations. In addition to being of significant concern in terms of animal health and welfare, the presence of footrot in a flock can have a negative impact on productivity resulting in serious economic consequences for both the individual flocks themselves as well as the sheep industry as a whole. As a trait that is hard to measure and has been previously shown to be low to moderately heritable in a number of different populations, it would be an ideal candidate for genomic selection. The use of genomics means that the animals under selection do not need to be exposed to the disease to determine if they are susceptible or not. The objective of this study was therefore to estimate genetic parameters for footrot in the commercial, purebred, Texel population, as an initial step to allow future genomic breeding values to be estimated. Ewes based on 29 different Texel flocks throughout the UK, were scored twice a year (midlactation and late-lactation) during 2015 and Each hoof of the ewe was awarded a score ranging from 0 (healthy foot) to 4 (demonstrating signs of severe footrot). A total of n=5,449 foot score records were collected from 2,767 individual ewes (732 ewes were scored in both years). The scores given to each of the 4 hooves of the animal were then summed and logtransformed (LnFSum). The data were analysed using univariate analysis by means of a model with fixed effects of ewe age, scorer, use of foot vaccine and the combination of flock x year x month of scoring as well as the direct genetic effect of the animal. The pedigree file used contained sire and dam information for 31,775 animals. In total, 867 records had a score of 1 or above (15.9% of the total number of records in the dataset). The univariate heritability estimated for LnFSum was 0.18 (0.02), with a direct genetic effect of 0.04, variance of 0.18 and a phenotypic variance of 0.22 (0.005). The results therefore indicate that genetic improvement can be achieved in this population, providing a long-term and sustainable solution to help reduce the incidence of footrot and the risk of antibiotic resistance. The use of future genomic breeding values, based upon these results, will also allow the identification of animals suitable for further breeding at an early age, which combined with the benefits already mentioned, will also further improve flock productivity and health and welfare. Keywords: footrot, sheep, genetic selection

4 Genetic Parameters for Longevity Traits in UK Dairy Goats L.J.Geddes 1, S. Desire 1, S. Mucha 1, M. Coffey 1, R.A. Mrode 12 & J. Conington 1 1 SRUC, Edinburgh, UK 2 International Livestock Institute, Nairobi, Kenya Goats with long productive lives are desirable as they reduce the number of replacement animals required, allow greater selection intensity, and reflect the health of the herd. The objectives were to estimate the heritability (h2) of seven longevity traits and investigate their genetic and phenotypic correlations with production traits. Data was provided on 23,984 female dairy goats from two genetically linked farm sites. The goats were a composite of 3 breeds: Alpine, Saanen, and Toggenburg. The pedigree file contained records of 40,422 individuals where 447 were sires and 17,166 were dams. The dataset consisted of milk yield records, birth year( ), kidding year( ), farm(2), maximum lactation number(1-11) and age at first kidding( days). Four lifetime longevity traits were considered: age at death (AD, as the number of years from birth to death); predicted age (PA, as the number of years from birth to death or if the animal was alive predicted based on current age plus the survival probability); lactations completed (LC, as the number of completed lactations from birth to death); and predicted lactations (PL, as the number of lactations completed from birth to death or if still alive predicted from the completed number of lactations plus the survival probability). In addition, three production longevity traits which included the number of days in milk over lifetime (DIM), the average lifetime daily milk yield (ADY), and the total lifetime milk yield (LY) were examined. AD, PA, DIM, ADY and LY were analysed using a univariate mixed animal model with the fixed effects of birth year, farm, maximum lactation number, age at first kidding, birth year by farm and maximum lactation by farm interactions. All fixed effects included in the model were significant (P<0.05). LC and PL were fitted without maximum lactation number. A bivariate model was used to calculate the correlations between the lifetime and production longevity traits, with fixed and random effects fitted as described above. Heritabilities (s.e) of the four lifetime longevity traits ranged from 0.09 to 0.11(0.01). For DIM, ADY and LY, h2 were 0.12(0.01), 0.44(0.02) and 0.17(0.02). Phenotypic and genetic correlations between the lifetime and production longevity traits were all positive with moderate to high correlations. Positive correlations indicate that increasing longevity in dairy goats results in a greater number of days in milk during the animals lifetime, a larger lifetime milk yield and an improved average lifetime daily milk yield. These results will allow genetic improvement in dairy goats by incorporating longevity into a selection index. This study has shown that longevity traits in UK dairy goats are under low genetic control and are suitable for use in selection programmes. Keywords: goats, dairy, longevity, heritability

5 Plenary 1 Legal implications of data provision services Data protection aspects by merging cattle data of various origins R. Knyrim 1, E. Dolamic 1, M. Mayerhofer 2, M. Koblmüller 3, J. Perner 4, R. Janacek 5, G. Schoder 6, F. Gstöttinger 7, R. Weissensteiner 2, B. Fürst-Waltl 8, M. Schagerl 9, H. Eder 10, C. Egger-Danner 2 1 Knyrim Trieb Rechtsanwälte, Mariahilferstraße 89A,, A-1060 Vienna, Austria 2 ZuchtData EDV-Dienstleistungen GmbH, Dresdner Str. 89, 1200 Vienna, Austria 3 LfL Upper Austria, Auf der Gugl 3, 4012 Linz, Austria 4 Chamber of Veterinarians, Hietzingerkai 83, 1130 Vienna, Austria 5 Animal Health Organisation Lower Austria, Landhausplatz 1, 3109 St. Polten, Austria 6 Animal Health Organisation Upper Austria, Bahnhofplatz 1, 4021 Linz, Austria 7 RZV, Sportplatz 7, 4840 Vocklabruck, Austria 8 University of Natural Resources and Life Sciences Vienna (BOKU), Gregor Mendel-Str.33, 1180 Vienna, Austria 9 AGES, Spargelfeldgasse 191, 1220 Vienna, Austria 10 Chamber of Agriculture, Schauflergasse 6, 1010 Vienna, Austria To facilitate improved herd management, easier access and compatibility of various data sources on farm and from external databases are of high priority for Austrian farmers. Recent research projects have focused on extended services for farmers to generate added value by linking a variety of external data sources. This includes extended health and treatment data, findings from laboratories, milk quality information from dairies and feeding information. These new online services will be provided by the cattle database (RDV) jointly operated by the Austrian and German performance recording organizations. The precondition for generating added value by merging data from various origins are beside standardization, data exchange and data communication legal implications on data protection regulations. Within the project ADDA (Advancement of Dairying in Austria) the legal implications and requirements for merging data from different data sources have been elaborated. Due to the fact that there is no data ownership, the different roles person affected, contracting authority and the service provider have to be defined and assigned to the data processed. The new European data protection regulation and its impact on the implementation related to provision of services based on cattle data of different origins and different circumstances and legal aspects for documentation and collection have been taken into account. The challenge is to set up a transparent system that guarantees the compliance of data protection regulations and minimized the administrative work for all parties involved when data from different data sources are integrated for routine applications as well as for research and development of advanced services. The presentation covers an outline of the basic legal data protection aspects and the example of implementation based on merging data from farmers, veterinarians, performance and breeding organizations, labs and dairies in Austria. Keywords: data protection, cattle data, legal implications, data integration

6 A computerized consent management tool for breeders: why, how? B. Balvay 1 1 French Livestock Institute, Agrapole - 23 rue Jean Baldassini LYON Cedex 7, France Access to livestock data has become a major issue in recent years, both for the breeders and for the organizations providing them with services. Breeder's control over these accesses requires his prior consent towards the organization wishing to use them. Professional organization "France Génétique Elevage" (FGE) manages a database gathering zootechnical data collected for the purposes of genetic improvement. In order to perpetuate recording of these data and allow their better valorization, this organization must consolidate breeders confidence in particular on the respect of their consent. Regulatory texts concerning access to livestock data are numerous and fall into different legal fields, which makes concrete rules complex to be defined and implemented: whether or not the consent of the farmer is mandatory depends on data type, use made of data and person who wants to use it. They also evolve over time and a new 2015 text on genetic information systems brings new obligations whose impact is yet to be measured. At the same time, relationships between organizations involved in genetic improvement are evolving in an increasingly competitive context. This is why data are becoming a matter of differentiation and their access an increasingly sensitive issue between organizations but also with breeders. Since 2009, FGE has been providing a data exchange service between its database and breeders or, more recently, a body designated by them. This service has recently been enriched with a consents management tool with 2 features: Registration procedures adapted to various organizations in the field, Consultation of all consents granted (in order to be able to terminate them if necessary). This tool has several innovative features to address consents management needs: Choice of web service with a standardized interface that allows a smooth use by all information systems (breeder, company, mutualized...), Functional wealth with a detailed description that not only contains the "basic" consent data (holder's exploitation, consent beneficiary) but also clarifies its scope by indicating: the species concerned, Access to data is granted and, where appropriate, the breed and the family of data (eg dairy control, animal insemination...). Presentation outlines legal context, challenges and features of this new consent management tool and refers to its possible positioning to address wider use than in the field of genetic data alone. Keywords: data management, access, consent, livestock

7 A Central Database for the Australian Dairy Industry S. Jenkins 1, T. Francis 1 1 DataGene Ltd., AgriBio, 5 Ring Road, 3083 Bundoora, Victoria, Australia In 2012 the Australian Dairy Herd Improvement Scheme began work on developing a replacement for their outdated genetic evaluation system. This project was entitled GESII (Genetics Evaluation System II). In late 2015 the industry looked to create a single industry body in Australian to manage genetic evaluation, data and milk recording standards and training. This along with other synergies enabled the GESII development to merge with existing milk recording software by the way of creating a Central Data Repository (CDR). This new single project is entitled GESNP (Genetics Evaluation System New Platform) and is delivered by DataGene, the newly established industry body. The goal is to provide a central industry data store where previously disparate and unconnected data can be shared and analysed. DataGene established, with the broader industry a Data Governance Group to manage the rules, standards and data management guidelines for data entering CDR. The GESNP program currently has a technical delivery team of 4 business analysts and 25 software engineers. There are 6 subject matter experts from across DataGene and Department of Economic Development, Jobs, Transport and Resources (DEDJTR) The first stage of the GESNP platform delivery will include the replacement of the legacy Genetic evaluation system and phase I of CDR. This will be delivered for beta and internal release in June 2017, with further phases staged for release into CDR will feature an open architecture, and the data will be a combined dataset of all industry data without duplication and with validation, verification and standards agreed and applied. The Data Governance Group (DGG) will be instrumental in setting these agreed parameters. There will however, be conditions of entry to CDR which will include committing to the terms of a Service Level Agreement (SLA). Software development partner will be encouraged to innovate and deliver new services on this combined data for the greater industry good. The CDR component of the GESNP program is broken into a phased approach with Phase I Acquire the Data featuring data collection, governance and support for operational systems. Phase II known as Report and Inform includes data analytics, increased functionality in reporting and analysis to better drive business decision making. Additional data providers will also come on stream during this phase. The final program phase, Phase III, Enhance and Extend sees increased analysis using Big Data, use of IoT, AI and machine learning, identifying new areas of complimentary data such as Social media and international data that can improve our understanding of customers, behaviours and trends. This phase will help provide predictive analytics and what if scenario planning. It is expected that CDR will then enter a product lifecycle development phase of continuous improvement and enhancement to always stay relevant and valuable to the user community and our industry partners. Keywords: central data repository, genetic evaluation, software, Australia

8 Smart Dairy Farming 3.0: multiplying innovations on the farmyard B. van t Land 1, G. Smeenk 1, H. Lucas 1, A. Lamers 1 1 CRV, Box 454, 6800 AL, Arnhem, The Netherlands SmartDairyFarming 3.0 is an initiative of three farmer-owned Dutch cooperatives: Agrifirm (feed), CRV (genetics/herdmanagement) and FrieslandCampina (dairy products). In March 2016 this organizations signed a cooperation agreement: Smart Dairy Farming in a Cooperative Perspective. One of the main reasons to launch the program was the lack of accessibility of data from individual animals. Stock management at herd level is today not accurate enough to enable optimal attention for individual animals. Sensors, index figures and decision-making models can help farmers establish the precise needs of individual cows, and make the right choices. At the same time new digital technology is being implemented on farms with high speed. Automatic milking robots, feeding systems, sensors to track movement and behaviour of the cow are examples and all are producing lots of data. In SmartDairyFarming 3.0 a digital highway for farm-generated data is being developed: The Smart Dairy Farming Datahub. To share farm data in a way where the farmer stays in control of this own data, the Datahub functions as an safe and secure electronic highway for an optimal data exchange. Data from farmers, dairy technology suppliers, research institutes and the participating companies are all combined to improve dairy farming as a whole. To be successful, it is essential to gain the trust of all parties involved but especially the farmers trust. This is why there is a strong emphasis on privacy and security of data. To harvest from all this data we need to make smart use of these data: data collection, combining data from different farm and agri-business data sources, analysis and create valuable information to steer on. This offers dairy farmers the opportunity to work more sustainable, more efficient and make better choices in the area of health, fertility and nutrition, ultimately improving revenue and our planet. The technology used in the Smart Dairy Farming Datahub follows a best-of-breed strategy: the datahub is using industry-standards with existing components. An important consideration is the position of the dairy farmer as the owner of his personal data and an architecture that safeguards data flows. The data hub is expected to be operational at the end of Another aim set for end 2017 is inviting other stakeholders in the dairy sector to participate in the program. In this lecture this will be further explained and a some actual use cases will be presented. Keywords: Needs of individual cows, IoT, trusted share of farm data, industry-standards

9 Session 1 Robots, Sensors and ICAR Automating the dairy farmer? Understanding the barriers to uptake and use of precision technology in dairy systems. D. A. McConnell 1 1 Agri-Food and Biosciences Institute, Large Park, Hillsborough BT26 6DR, Belfast, Northern Ireland Rapid development in precision technology presents significant opportunity to revolutionise livestock management practices on farm. In addition, the wide biological, climatic and temporal variation which can occur on farm, the dairy industry remains a perfect application for this innovation. Although the development of precision technologies suitable for use in the dairy sector continues at pace, the true value of these technologies is unlikely to be realised without significant social and economic change on farm. To assist farm businesses in achieving maximum return on investment when implementing new technologies, it is important to understand how farmers are currently using data provided by established precision technologies, understand the skill base required by these techniques, and consider how these findings can be integrated into the development process. This paper will present the findings of a 2016 Nuffield scholarship which, through a range of interviews with industry, farmers and scientists will provide a global view of precision technology uptake in the dairy industry. This paper will discuss the socio-economic factors influencing uptake, and importantly, the true use of new technologies on dairy farms across the world. It will also evaluate the ability of the current skills base both on-farm and within industry to fully exploit the new developments in technology now available to dairy farmers. Finally, it will consider how farmers and industry currently evaluate the success of precision technology purchases. Recommendations for scientists, industry and farmers will be made to ensure the potential of precision technology is maximised within the dairy sector. Keywords: precision technology, dairy, grassland, uptake

10 Use of daily robotic progesterone data for improving fertility traits in Finnish Ayrshire J. Häggman 1, J.M. Christensen 2, J. Juga 3 1 Department of Agricultural Sciences, University of Helsinki, FI-00014, Helsinki, Finland 2 Lattec I/S, Slangerupgade 69, 3400 Hillerod, Denmark 3 Department of Agricultural Sciences, University of Helsinki, FI-00014, Helsinki, Finland Currently, cow s ability to return to cyclicity after calving is mostly evaluated using the first insemination measurements, which are highly influenced by management decisions. However, if consecutive progesterone measurements are used the first heat can be estimated accurately even if the cow does not show clear signs of heat. The data from 14 Finnish dairy herds using DeLaval Herd Navigator (HN) system were used to study cow s ability of returning to cyclicity after calving Ayrshire cows from parities 1-3 were included in the analysis. HN system takes milk progesterone samples automatically during milking and apply biological models to calculate the time of estrus and probability for pregnancy if cow is inseminated. HN use the change in progesterone level from high to low to determine that a heat has occurred and will look for a new heat around 21 days later. In this study the data of first estimated heat based on progesterone data between 1-100d after calving were included. The mean number of days from calving to first heat (CFH) detected by HN varied from 49.9d to 65.8d depending on parity. Most of the cows in the data had been inseminated for the second or third heat identified by the HN system. The mean number of days from calving to first insemination (CFI) varied between d and interval from first to last insemination (IFL) between d depending on the parity. The mean number of inseminations in HN herds was around 2, which is similar with the estimates from previous studies. When phenotypic estimates were compared to those from previous studies CFI was 4.5d shorter and IFL 10.3d shorter in HN herds for the first parity Ayrshire cows. CFI was 1d and 5.7d longer and IFL was 6.6d and 4.9d shorter in HN herds for parities 2 and 3, respectively. According to previous studies the average cost of CFI is 0.51 /d and IFL is 2.56 /d for the red dairy breed. Heritability estimates were calculated using DMU software and linear animal model for first parity cows being 0.27±0.13, 0.07±0.07 and 0.03±0.06 for CFH, CFI and IFL, respectively. Because of the small number of animals in the data most of the estimates had high standard errors. However, the magnitude of the estimates are in line with previous studies where CFH have found to be more heritable compared to other fertility traits. Results showed that using milk progesterone information to detect heats shortened CFI in first parity cows and IFL in parity 1-3 cows. Keywords: progesterone, cyclicity, dairy, ayrshire

11 Collecting milking speed data as part of official milk recording. R.H. Fourdraine 1, H.A. Adams 2 & A.D. Coburn 3 1 CRI International Center for Biotechnology, 1706 County Road E, Mount Horeb, WI 2 CRI International Center for Biotechnology, 1706 County Road E, Mount Horeb, WI 3 AgSource Cooperative, 135 Enterprise Drive, Verona, WI Due to the growing use of robotic milking systems, the interest in optimizing the milk output of the robotic milking unit has added a new dimension to breeding and managing dairy cows. Milking speed, milking unit attachment speed and time required for cows to enter the robotic milking unit are three major factors in determining which cows are more suitable for robotic milking systems and maximize returns on investment. Milking speed also has application in conventional parlors. Milking speed can have a direct factor on operational expenses associated with milking the herd. High producing cows with consistent milking speed will optimize parlor throughput and increase the amount of milk collected on a daily basis. Dairy producers have had the opportunity to purchase in-parlor milk meters and collect data that would help in the optimization of parlor performance, however costs and maintenance concerns have limited the adoption in the United States. The data that is produced from existing systems varies in format and archive history and is rarely transmitted as part of milk recording services, thus no national genetic evaluation of milking speed currently exists in the United States. AgSource has been a long-time user of Tru-Test (Tru-Test Inc, Mineral Wells, TX) Electronic Milk Meters (EMMs). EMMs are used to collect monthly DHI milk weights, milking durations and milk samples. Due to the growing interest in parlor efficiency, in 2015, AgSource started collecting the milking duration times using its EMMs. Milking speed records in the form of milk weight and milking duration data are collected on approximately 100 large farms totaling over 100,000 cows. Beneficial to analysis and utilization, data is measured as a continuous variable (kg/minute) versus a standard categorical measurement. AgSource milking speed data proved to be a consistent measure based on stage of lactation and parity. Milking speed values averaged 2.6 kg/minute and ranged from 1.4 to 4.5 kg/minute. Further analysis showed that milking speed data was positively correlated to DHI Mature Equivalent (ME) 305-day milk production and negatively correlated to somatic cell score at low and high milking speeds. Using genotypes supplied by the USA A.I. cooperative, Genex, resulting breeding values were calculated on over 60,000 cows and bulls. Employing new technologies in regular DHI recording result in new reliable and consistent phenotypic measures that can be combined with genotype data to identify new markers, new genetic traits of economic importance and be incorporated in DHI value-added management reports. Keywords: milking, speed, parlor, efficiency

12 Objective Carcass Measurement to Improve Lean Meat Yield and Eating Quality in Australian Beef, Sheep and Pork D.J. Brown 1,5, D.W. Pethick 2,5, P. McGilchrist 2,5, C.K. Ruberg 3,5,W.S Pitchford 4,5, R. Apps 3,5 & G.E. Gardner 2,5 1 Animal Genetics and Breeding Unit, University of New England, 2351, Armidale, Australia 2 School of Veterinary and Life Sciences, Murdoch University, 6150, Murdoch, Australia 3 Meat & Livestock Australia, 40 Mount Street, 2060, North Sydney, Australia 4 School of Animal and Veterinary Sciences, University of Adelaide, Roseworthy, Australia Advanced Livestock Measurement Technologies project, Meat & Livestock Australia, 40 Mount Street, 2060, North Sydney, Australia Australia's meat supply-chains are a significant contributor to GDP and thus continued improvement in productively across the whole value-chain is vital. To this end a new project has recently commenced aiming to transform industry competitiveness by creating feedback and decision support systems linked to accurate carcass measurements. The project is implementing the following in major beef, lamb and pork supply-chains; accurate measurement systems for live animals, carcases, and cuts; actionable producer feedback to effect decision making on compliance and profitability; and, value-chain information systems to extract maximum economic return from products. Current pricing systems reflect the lack of accurate technologies that can measure eating quality (EQ) attributes and proportion of saleable meat (lean meat yield; LMY) in a carcass. Thus for lamb and pork there are no price signals for EQ, with carcases mainly traded upon weight, with some value-chains also using a single manual assessment or measurement of fat depth to indicate LMY. Although similar in beef, the Meat Standards Australia grading system enables price signals for EQ. To address these deficiencies, the project is developing measurement technologies to assess both LMY and EQ, and carcass inspection support tools. The technologies being investigated include: 3D-imaging, high throughput dual energy x-ray absorptiometry (DEXA), near infra-red spectroscopy, RGB digital and hyperspectral image analysis. Data capture systems are planned to enable data-to-decisions forward and backward between retail, abattoir and on-farm sectors. These technologies will assist on-farm for both seed-stock and commercial producers by providing more accurate prediction of carcass composition for improving genetic gain for difficult to measure carcass traits and improved compliance to carcass weight, fatness, and EQ specifications. The data will also allow processors to precisely value carcases, optimising market-based cutting decisions, market allocation of products, and improving labour efficiency. For retailers this project will result in greater consistency of product size and EQ, and an improved ability to predict supply. The project will also pilot assisted or automated offal assessment to aid in disease detection and feedback to producers. The DEXA technology being developed by this project is delivering very promising results for measurement of carcass composition in both sheep and beef at commercial abattoir line speed. Work is continuing on measurements related to eating quality, with hyperspectral imaging for tenderness, intramuscular fat, meat colour and ph. The resulting data from all of these measurements will facilitate the development of novel pricing mechanisms that reflect true value of the carcass to the whole value-chain.

13 It is known that antagonisms exist between LMY and EQ, and these measurement technologies will assist producers to manage these, both at the phenotypic and genetic levels. The industry is also working towards using these new measurement technologies, with progeny testing and genomic testing to provide data to underpin the genetic evaluation for these traits and reduce the current reliance on expensive resource populations. This project has key advantages for the whole value-chain by providing a foundation for objective value-based price signals. This will also enable enhanced product description and more sophisticated value-based marketing. Keywords: DEXA, CT, lamb, hyperspectral imaging

14 Towards a robust protocol for enteric methane measurements using a hand held Laser Methane Detector in Ruminants ThiphaineBruder 1, Benoit Rouille 1, Tianhai Yan 2 & Mizeck G.G. Chagunda 3 1 Service Productions Laitières, Monvoisin BP , LE RHEU Cedex, France 2 The Agri-Food & Biosciences Institute, Hillsborough, Northern Ireland, UK 3 Future Farming Systems Group, Scotland s Rural College (SRUC), Edinburgh, EH9 3JG, Scotland, UK Direct measuring of enteric methane in breath of ruminants is becoming popular. Since the first peer-reviewed publication (Chagunda et al., 2009) showed the potential application of the proprietary Laser Methane Detector (LMD) in ruminants, it has been shown to have strong relationship with traditional techniques such as respiration calorimetric chambers. For example, Chagunda et al, (2013) reported sensitivity and specificity for cows of 95.4% and 96.5%, and for sheep, sensitivity was 93.8% and specificity was 78.7%. However, there is no joined-up protocol covering all aspects, including, data collection, data extraction, data handling, and estimating methane volume from the measured concentration. Using data from two studies this paper presents results from tests and analysis to develop a method for data extraction, determine optimal recording duration, differentiating breath from eructation; and conversion of methane concentration to volume. The first study used a group of 71 dairy cows with repeated measurements over a 5 week period. Methane was measured by pointing the LMD at the nostril of the cow from a distance estimated to be 1m in the feedface after midday milking. Measurements lasted 4 to 5 minutes. For each individual timeseries measurement, time of recording and cow s tag number were recorded. In the second study measurements were taken from 18 Holstein Friesian heifers simultaneously by the LMD and the metabolic chamber. In differentiating eructation from breath, one standard deviation for the individual measurement-window, was used as a threshold. This proved to be a biologically meaningful and statistically effective way of distinguishing methane coming from the rumen through eructation and that from the normal breath. An example is the mean of (with a standard deviation of 182.7) ppm. To determine the optimum recording duration, five levels of 60s, 120s, 180s, 240s and 300s were created. Gross average of methane emissions was calculated for each recording window. Significant difference was tested using analysis of variance (ANOVA). In this test the only group that resulted in significantly low measurements (p<0.001) was the 60s. Given that eructation episodes in cow breath cycles are estimated to be one to three per minute, measurement windows of less than 3 minutes would risk missing out on some eructation episodes. When methane was measured when animals were standing the relationship between LMD methane and Chamber methane was highest (r = 0.65) while daily averages had the weakest relationship (r = 0.48). This strong and positive correlation allowed us to build regression equations for estimating methane volume (g/day) from methane concentration (ppm) measured by the LMD. Key words: Enteric methane; measuring protocol, breath cycles

15 Sharing data through an API platform - API AGRO. Erik Rehben 1, Béatrice Balvay 2, Theo Paul Haezebrouck 3 1 IDELE - Institut de l Elevage, 149 rue de Bercy, 75012, Paris, France 2 IDELE - Institut de l Elevage, 23 rue Jean Baldassini, 69007, Lyon, France 3 ACTA-149 rue de Bercy, 75012, Paris, France In order to facilitate data sharing among stakeholders in agriculture, French leading organizations of Research and Development launched an API (Application programming interface) platform one year ago. The goal of that project, so called API AGRO, is to establish a business environment to facilitate the development of new services. A detailed presentation of the main functionalities of API AGRO is given and the different licences which may be used by the data providers are also addressed. The key IT techniques which have been implemented are presented briefly. The current data sets available through the platform are detailed as well as the future development. As a conclusion, the interest, the cost and the limit of that solution are discussed. Keywords: data sharing; API; livestock

16 Manufacturer s showcase and carry-over effects Milk sample carry over in the field identifying and resolving the challenges. Justin Frankfort 1 1 National Milk Records, Fox Talbot House, Greenways Business Park, Chippenham, SN15 1BN, UK Obtaining an accurate milk sample from individual animals at a milk recording session is the point at which unique challenges are presented, on both the equipment and at the human levels. It is where the milk recording technician or milk sample collector becomes directly involved in the herd test. It is also the single most important role of a milk recording organization (MRO). While other event data may be captured by computers and milk yields estimated by milk meters, collection of a representative milk sample is critical for accurate results delivery by the MRO and delivery of effective management data for decisions by the herd owner. Dairy producers demand more from MROs, as milk margins continue to tighten. In response to this demand, MROs have investigated and delivered other useful information for management that is contained in the milk sample. The net result in an increase the value of the milk sample, where investment in collection has been made, and consequently, increase the value of the services offered by the MRO. Additional tests such as ELISA on preserved milk samples collected by MROs may have different tolerance levels to carry-over when compared to traditional infrared analysis. Further, these ELISA analyses have a larger impact on herd performance, where management decisions are immediately made on milk analysis results that include culling, breeding, and or administration of prescribed pharmaceuticals. Milk pregnancy testing, through the ELISA pregnancy-associated glycoprotein (PAG) test, has a low tolerance to carry-over which may provide inaccurate results, including both false negatives and false positives, to a dairy producer. These inaccuracies in PAG testing have a high visibility in the marketplace, a significant effect on the bottom line of the producer, and subsequently damage the reputation of the MRO. Compared to traditional milk component testing, which includes multiple test events over the course of a lactation, where the net effect of carry-over is minimal and management decisions are not as finite (culling of cows), the impact of carry-over may in value-added milk testing is problematic. Knowledge of and addressing potential carry-over in MRO-collected milk samples is key to delivering both qualified and accurate results to the dairy producer, and increasing the value of MRO services. MROs have a responsibility to address the challenge of carry-over through education, training, and implementation of best practices in their respective field operations. Keywords: carry-over, milk recording organizations, best practices

17 Practical Considerations to Reduce Carry-Over in Design of Recording & Sampling Devices John Baines 1 1 Fullwood Ltd, Grange Road, Ellesmere, SY12 9DF United Kingdom Traditionally, the main objective of collection of milk samples from individual animals was to provide an objective estimate of the proportion of fat, protein and lactose present in that animal s milk. Noting that fat percentage particularly rises towards the end of the milking, carryover of significant residues may have a negative effect on the accuracy of the results. Milk recording organisations continue to seek opportunities to add value to the routine sample through availability of additional analyses such as ELISA based pregnancy tests and PCR based mastitis pathogen identification. Culling and/or treatment decisions can be influenced or based on these types of tests. As a result, the importance of accuracy in this type of testing cannot be over stressed. The sensitivity of such tests emphasises the need for limitation and/or avoidance of carryover. The physical properties of milk result in a liquid which tends to stick to the surfaces of the milking machine and to drain only slowly. Milk tubes, by necessity, need to be flexible enough to permit attachment to teats and so that milking clusters can be stored out of the way of cows feet and operators. Those contrasting needs result, inevitably, in the potential for milk to be trapped at the end of an animal s milking. Attachments and/or components included in the milk tube may also create the potential for trapping of small quantities of milk. Some milk yield measurement devices work on a fill and dump principle. Inevitably, the final flow from an animal does not fill that measuring chamber resulting in the need for an arbitrary final emptying of the valve. Sampling devices themselves may add to the potential for residue carryover. In contrast, the speed of operation in many milking facilities means that, practically, there is little or no time for any clean of equipment between animals to remove residues. This paper considers the potential for carryover of milk, and cleaning solutions, the difficulty in eliminating residues and practical considerations in design and operation of milking machines and sample collection devices to minimise the effects. Keywords: carry-over, design, best practices

18 The new CombiFoss 7 DC Differential Somatic Cell Count and other Advancements in Milk Testing D. Schwarz 1 1 FOSS Analytical A/S, Foss Alle 1, Hilleroed, Denmark FOSS has launched the 7 th generation of CombiFoss milk analysers in October The new CombiFoss 7 DC seamlessly integrates MilkoScan 7 RM and Fossomatic 7 DC. The instrument allows to test raw milk for up to 19 parameters, including the brand new Differential Somatic Cell Count (DSCC) parameter, simultaneously in just 6 seconds. The objective of this work is to provide an overview on the key advances of the instrument and an update on the latest developments in terms of working with new parameters for milk testing, particularly DSCC, from around the world. The MilkoScan 7 RM can be used to test for up to 17 different milk component parameters. The latest generation technology includes improvements of the optics and flow systems that result in better statistics, in particular for minor components such as urea and BHB (beta hydroxybutyrate). Apart from that standardisation of spectra is still possible using FTIR equalizer (FTIR Fourier transform infrared spectroscopy), which is particularly important nowadays where full spectra information is utilised for various purposes. The Fossomatic 7 DC allows to measure 2 parameters, SCC and DSCC, simultaneously at a speed of 600 samples per hour. The key elements of the new milk analyser are a new chemistry, a new incubation unit, and a new measuring module. Besides, the design of the instrument allows easy accessibility of the different modules inside the instrument. DSCC is a new biomarker for mastitis management. Mastitis remains to be a significant challenge on dairy farms and still causes tremendous economic losses to the dairy industry. DSCC provides more information on the actual inflammatory status of the cow s udder by revealing the percentage of immune cells (i.e. DSCC represents the combined proportion of neutrophils and lymphocytes). Several research projects on the practical application of DSCC in the frame of dairy herd improvement (DHI) testing are currently running around the world. A first research study, where the DSCC parameter was investigated before, during, and after artificially induced mastitis under controlled conditions was recently completed. The results showed that DSCC values changed significantly during the course of the experiement (i.e. <60%, >90%, and <70% before, during, and after infection, each). Hence, first indications on where to set a threshold for DSCC to distinguish between normal and active (e.g. mastitis) inflammatory response are available. In conclusion, the new CombiFoss 7 DC allows highly accurate, fast, reliable, repeatable, and robust determination of up to 19 parameters from raw milk samples at low cost. DSCC is a new parameter providing more detailed information on the actual inflammatory response of the mammary gland and thus opens up the possibility to develop new tools for mastitis management that can be offered through DHI testing programmes.

19 Employing high-resolution big data for predictive modeling in precision dairy farming. G. Katz 1 1 afimilk, Kibbutz Afikim, 15148, Israel Modern farms exercise decision making employing high level of automated computerized data acquisition by sensors installed in the dairy parlor or on the individual cows. These data are utilized in the management system to support decision making in high precision farming. Analytical reports are extracted from the raw data estimating farm and cows events (what is happening). More advanced systems can apply descriptive moddeling suggesting the causes for te event (why did it happend). New megga trends incorperated in IOT (inernet of things), such as big data, cloud and mobile enables moving on to predictive modeling (what is going to happen) in high precison farming. Such predictive moddeling is demonstrated in early prediction of total lactation production. Estimating total lactation yields from truncated data early in lactation allows for better selection decisions and future production planning. Selection decisions are made as early in lactation as possible, before breeding (50 DIM). Continuous production planning aimed at getting the best financial returns are essentials in quota systems and seasonal (summer/winter) differential payments. The predicted total lactation yields must also be accurate as possible to avoid selection bias. Many formulas have been adapted for trunkated prediction of total production. However, these fromulas were constructed retrospectively, based on huge amount of inividuals, thus neglecting to atrribute locality herd phenomena. Current data, employing monthly milk tests is also inefficient for making prediction at the most critical descision making time (50 days) for having only two data sets per individual at most. The AfiFarm Herd Management system, including the AfiLab TM milk analyzer (AfiMilk, Afikim, Israel) and milk meter, which provides on-line data on gross milk composition (fat, protein, lactose and coagulation properties) is employed with additional data collected on the farm and it s eco system for Prediction of total lactation production (305-days yield or complete lactation yield, fat protein and ECM). These predictive models calcualted updated continuously to be local and herd specific. Such an approach can also be used to predict emerging of health associated problems(post partum deseases, mastitis). The flaw of this new approach is the failure to predict black swans events. Keywords: on line milk analysis, descriptive analysis, predictive modelling, Big Data.