To read about the program scroll down below 'Latest Updates'.
There have already been more than 1,200 orders for the 12k chip (from breeders). Delivery of new genomic breeding values is as planned, with imputation now automated in the pipeline. We have prepared a research paper on 12k SNP-chip design and imputation accuracy.
Information Nucleus samples for 700k HD genotyping are being analysed with the aim of improving prediction accuracy and providing information for future imputation based on full sequencing. We have about 400 sample results back, and these look good from a genotype perspective, but we have not started imputation and association analysis. Plans to genotype 200 Merinos using full sequencing depends on the extension outcome.
Getting the Information Nucleus data for reproduction in good shape has proven to be a challenge. It is expected to be finalised by the end of the year. There is a Technical Committee meeting scheduled for February. The sheep Genetics Technical Committee has also decided that we can release RBVs as ASBVs.
Stephen Lee has worked on further case studies using MOET programs. Brian Horton's models have proven very useful, modelling all three layers of the breeding structure (with close to 100,000 commercial ewes).
Lu Hogan has further developed a plan for an industry training program in March for genetics service providers.
The delivery of genomic breeding values started the new year with a new product: the 12k SNP chip. This is a lower cost genomic test that can deliver a similar genomic prediction accuracy as the 50k SNP chip. We have used imputations to recover 50k information from 12 genotype data and the imputation accuracy is usually over 90%. The prediction accuracy is over 95% of that of the full 50k chip. The cost of the 12k chip is ~$50 which constitutes no real subsidy on the chip cost and does not fully cover the people involved in delivering the service.
We are going to explore the accuracy of predicting animals from some minor breeds, as this may be higher than predicted from theory, due to strong relatedness within the breed. This will also involve genotyping more progeny from these minor breeds.
The first 700k genotypes have been returned from Geneseek and we are working on getting the samples to Geneseek (or using existing ‘left over’ samples at Geneseek) for the remainder of the set of 950 tests that have been planned. Data analysis is expected to be well underway before the end of 2013.
We are currently reviewing the parentage test service and the parent assignment algorithm, and will document carefully under what circumstances the parent assignment rates are impaired, e.g. with small sample sizes or closely related candidate parents. Overall, the assignment rate is very good and is close to 100%.
The CRC recently held a very successful workshop with breeders involved in the pilot projects. The total number of participants was 56, including researchers and attendees from the Sheep CRC.
Lu Hogan has developed a plan with Sheep Genetics on industry support and training, with components including development of modeling and description of case studies, an intensive training of consultants, and development of extension material.
We will still have two years of followers left - in total about 1600 animals over six sites. This requires ongoing coordination and checking of data entry.
Genomics Pilot Project
Test results for Pilot Project III and the Large Scale Genotyping Project are currently being finalised. The Large Scale Genotyping Workshop with breeders in May was very successful, with a range of questions/issues identified for further work in order to support breeders implementing genotyping into breeding programs. A subsequent workshop has been scheduled for the end of August to report back on modelling, case studies and scenarios that are being developed in response to the May workshop. Guidelines and information for breeders and service providers is being developed to support use of the 12K test, which is likely to be launched at the August workshop. Arrangements for transition of Parentage, Poll and Genotyping Tests to Sheep Genetics are being developed.
The genomic Pilot Project results are almost complete and as they become available they will be reported on a routine basis. GBVs for NLW could soon be available but released only after a report on a trial run(s) has been prepared by AGBU.
DNA extraction is now complete for the final drop of Information Nucleus progeny (2011 drop) and has been sent for genotyping. The genotyping of INF progeny is therefore now almost complete. We will have genotyped ~13,000 progeny out of 18,800 progeny weaned. Animals not genotyped are mainly wethers and progeny from minor breed sires.
The production of an ovine low density (LD) chip (12k) has been ordered from Geneseek and this is expected to be available during May. We ordered 14,000 chips over 2 years to be used for resource flocks and commercial breeders. The high density (HD) chip (700,000 SNP) is now expected in May-June and we are sorting out 950 samples to be genotyped (500 of those will be INF sires).
Pilot Projects - Pilot Project I and II participants have received updated RBVs and accuracies that have been re-scaled so that they can be compared with ASBVs within flock. Across flock comparison of RBVs is not yet possible. Updates and information on research outcomes have been made available to breeders through media stories, Sheep Genetics communications and the recent Leading Breeder Forum in Bendigo. The presentations from this forum are available on the Sheep Genetics website by clicking here.
The DNA tests have all been allocated in Pilot Project III and the Large Scale Genotyping Project. Most of the blood cards have been received and reporting of RBVs is progressing as planned. A list of questions and issues on "implementation of genomic testing and incorporation into breeding programs" has been collated. This list will form the basis for a workshop with breeders, service providers and geneticists to be held during May
The genetics group held a review workshop during October to develop plans for the final stages of genetic analysis for the Information Nucleus data. Groups responsible for each of the key discipline areas such as meat science, wool, reproduction and parasites now have a consistent approach for the analysis of all traits. There was a further meeting with the genomics team in Armidale during December that focused on a review of accuracies in estimating Research Breeding Values. Most of the accuracies are now very “usable” and are improving with increasing amounts of data. There are plans to genotype a number of animals from the Information Nucleus flock using the higher density (700k) SNP chip. It is hoped that the higher density genotyping will improve the accuracy of prediction and provide a lead as to whether full sequencing of key animals will be a valuable option.
The main activity on the Information Nucleus sites is the continued management and monitoring of the ‘Followers’. Reproduction data from these animals is contributing to the improved accuracy of prediction of number of lambs weaned from genomic analysis. Lambing data for 2012 will only be finalised once the DNA test has been applied to sort out parentage.
Pilot Project 3 is now fully subscribed with breeders having ordered the 3000 genotyping tests based on the 50k product. To date there have been 17,000 tests allocated for parentage and horn-poll testing.
A workshop held in October with a number of breeders was designed to share information on the way that the genomic predictions of breeding values can be best used in commercial breeding operations. There are 11 large-scale genotyping projects that will be closely monitored to determine the value of additional genomic information to their respective breeding programs.
In the genomic analysis, work is underway to finalise analysis of reproductive data, combining information from Information Nucleus (INF) progeny and industry sires.
We will genotype the last 1800 INF progeny in the coming months. We have also planned genotyping using a high density chip (700k), for 950 animals, and will then use this information to see if we can increase accuracies of smaller breeds.
We have had a very high level of interest from breeders to take part in the 50k testing in Pilot Project 3 (where 50k tests are subsidized to a $50 cost per breeder). We have sold 1500 tests in the ‘normal category’ where breeders can do a maximum of 50 tests.
We will also allocate 1500 tests to breeders that are prepared to do a few hundred tests. We will work closely with them to advise on optimal strategies to genotype and use the genomic information. For more information about the tests and to order online click here.
This program continues to run smoothly with data collection and reporting on schedule. The last quarter has involved preparation of the Operational Plan for FY13. Planning for the Pilot Project genotyping and for the parentage/horn-poll test has involved a number of discussions with Sheep Genetics and other stakeholders. It has been agreed that the parentage and genotyping pilot projects will be managed through the new Project 4.4 with Lu Hogan as the Project Leader.
Design and Analysis
Fay Haynes has been working with Steve Potts to get a reconciliation of all genotyping conducted by the CRC and account for all 50k chips used over the last few years. This has been a useful exercise and has identified a number of loose ends that have now been resolved. We currently have no unused SNP-chips available through Geneseek. All future testing is now linked to specific activities. An order has just been placed for 5,000 50k tests. Of these, 2,000 will be allocated to progeny of the 2011 joining, 2,000 for the genotyping pilot project in FY13 and 1,000 for genotyping industry sires that have accurate ASBVs for ‘number of lambs weaned’ and ‘WEC’ (parasite resistance).
Pilot Project III is close to completion with over 860 rams entered in the project and all data returned to breeders within the specified 12 week period.
There have been a number of meetings to plan the parentage/horn-poll test in terms of logistics, reporting and the allocation of available tests to provide most useful information for research and industry feedback. There has also been an analysis of costs and pricing options.
Information Nucleus Operations
All sites are up to date with measurement programs and management of the flocks. Animals caught up in the footrot eradication program at the Hamilton site have been finished to weight specifications and destocked. The staff at Hamilton and the VIC DPI management team have done a great job in finishing the slaughter group of animals to specification. The female progeny have been lost to the project and this represents a setback as reproduction data is of particular importance to the overall Information Nucleus Program.
The Site Advisory Groups held their final meetings during June. These groups have made a very important contribution to the success of the Information Nucleus program. It is anticipated that advisory groups will be set up at Katanning and Kirby to assist with management of the 2012 joining and measurement program. Establishment of these new groups will depend on the MLA-led consortium responsible for the new Resource Flock project.
Information Nucleus Data Management
All data files are up to date. During the last quarter there has been progress in development of management systems for the parentage SNP data. Further work is required to automate parentage analysis and reporting and this will be a priority.
In March we completed the second batch of 50k chips with Geneseek, consisting of remainder samples from the 2008-2010 drop and around 400 samples from the 2011 drop (total 2900 from INF progeny buffy coats), pilot project 2 samples (~750) and industry sires (~250). The number of Information Nucleus (INF) progeny remaining to be genotyped is ~2000 and we need to do a few hundred more industry sires. For these we need a new service contract with Geneseek (~250k).
The additional genotypes will need to be analysed to obtain increased accuracies of genomic prediction. In this quarter we obtained the first estimates for NLW, with results looking pretty good (genomic prediction accuracy ~.25-.30 for merino and Border Leicester. The delivery of results in the pilot project 2 is being achieved in a timely manner (for Sheep Genetics registered samples the average is 38 days, longer delays occur for non-SG registered samples). We have also started to do some validation of research breeding values from pilot project 1, as those sires now have some progeny. Correlations of GEBV and current ASBV are substantial and as expected. A potential issue to be resolved in the procedure is in the accounting for genetic groups: this needs to be consistent between GEBVs and ASBVs, i.e. both predictions need to consist of a between group component (representing bloodline, breed etc) and a within breed deviation.
We have obtained results from phase 2 of the validation of parentage testing. The first phase was ‘in silico’ based on the INF database. This validation was based on samples from real flocks, and genotypes were obtained on a sequonom platform (which will be the platform for cheap parentage testing). Genotyping results look good and we concluded that we need in first instance ~150 SNPS (3 panels for a parentage test. This will also include 3 SNPs to predict poll/horn. The next step is to negotiate price and condition for a larger pilot project (20,000 samples) that will include samples from commercial flocks, breeders and INF flocks.
We have started to investigate options for cheaper genotyping, such that breeders can more easily afford genotyping for the purpose of genomic selection. Samples genotyped from a 10k test could often be imputed to 50k data, allowing enough SNP density for genomic prediction. Options are to run a 10k SNP chip (cost ~$70 via Geneseek) or a test for 10k SNPs obtained via Genotyping by Sequencing (DPIVic). The latter is a new technology that seems very promising and we also explore this with a Canberra based company (possibly delivering 100k SNPs for $50). It was decided that we take out time to properly develop and test these options in the next financial year, not to confuse the market place too much and for 2012 keep offering a 50k test (possibly subsidised in a ‘pilot project 3 framework). Together with a parentage/horn-poll test there will be enough excitement and activity at this front later this year.
Work on phenotypic analysis is progressing, but slowly. The team is attempting to at least complete some of the across trait correlations before the end of the financial year.
During the last quarter, there has been additional genotyping data added to the database. This significant quantity of new information and the commencement of the second Genomics Pilot Project provided the justification to re-analyse all of the genomic association data. The results have sparked a lot of interest. The accuracies with which a wide range of important traits can now be predicted, based on genomic information, has proven beyond doubt that the genomics technology will be of lasting value for the Australian sheep industry. The results of the improved predictions are already flowing through to industry.
The first recipients of the upgraded information were breeders with rams entered in the Pilot Project. The next groups to receive new information were breeders with rams that had been used in the Information Nucleus and participants in the first Genomics Pilot Project. A valuable aspect of the genomic approach is ability to update and expand predictions for all animals for which genotypes are available. The prediction of horn/poll was also released as part of this new analysis.
Information Nucleus Operations
A review of reproductive performance in the ewe progeny identified a number of problems leading to low fertility. There appear to be a number of factors contributing to low lamibing percentages including the complication of measuring age at puberty and some ewes below target weight at first joining. A series of measures have been put in place to overcome these problems.
All arrangements are now in place for the 2012 AI program. Approximately 6,000 ewes have been allocated across two sites (UNE/Kirby and the Katanning/DAFWA sites). Moving to two sites and increasing the number of ewes per site provides a more cost-effective model for data collection and reflects the transition to a post-CRC model for the Information Nucleus program.
Design and Analysis
Further progress has been made with extracting DNA samples of INF progeny with most of the animals from the 2007-2009 drops now extracted and a start was made with animals from the 2010 drop. Extractions have been done at UNE (2000 samples) and at the animal genetics lab of UQ in Gatton (1500 samples). Data on genotypes has been returned and there is now a more efficient pipeline to QA and store genotype data with AGBU. ~4200 INF progeny have been genotyped since October 2010, with ~800 to follow this FY. Some animals that failed in first instance in the Genomics Pilot Project have also been re-genotyped. Further results have been sent out in the Pilot Project, e.g. carcass traits for 155 terminals as well as further results for 82 merinos (staple strength, birth weight, and wrinkle) and 16 maternal sires. Freezers have purchased and good progress is being made with a solid documentation system for storing DNA samples.
In the last quarter we will start a whole new series of association analysis with hopefully higher accuracies for genomic breeding values across the board. In first week of April we have a genomics workshop in Armidale to discuss the further ways for data analysis, the design of INF2 as a reference population for future genomic selection, and further collaboration with the International Sheep Genomics Consortium to ensure that developments of high density SNPs and sequencing projects are carried out with sufficient Australian stake holding (e.g. using Australian genetic resources).
Various papers have been written by the group for AAABG and the Livestock Genomics Conference in Melbourne 2-5 May 2011.
The Dorper site near Narrogin on the property of Adrian Veitch continues with assistance being provided by DAFWA staff from the Katanning INF site. Sires for the second mating have been identified (10 Dorper and 3 link) and semen assembled for the 2011 AI mating.
Genotyping data are in the new genotyping database and a letter with the latest Research Breeding Values were sent out in mid March to all Genomic Pilot Project participants.
A commercial Laboratory Information Management system (FreezerPro) has been ordered and will be available in Q4 for tracking of DNA samples from the INF and Genomics Pilot Project Stage 2.
Ongoing genotyping has provided a few thousand extra genotypes and these are being used to increase accuracy of genomic breeding values. Most of the time, accuracies are increasing with more data, which is encouraging. Some traits require further work, e.g. shear force tenderness and a couple of new carcass traits seem to have low accuracies, which could be related to quality of phenotypes, especially in the first year. Data on reproductive traits is still limited and will be enhanced with genotyping of industry sires with high ASBV accuracies.
All INF DNA and blood samples are now stored in duplicate in different freezers in Armidale and a lab inventory software/database is linked up to the CRC INF database.
The Genomics Pilot Project 2 project is now running and first samples have been received and genotyped. Close to 800 samples are expected before November. Genomic breeding values will be delivered quite promptly via AGBU. Another workshop with ram breeders has analysed the (costs and) benefits of genomic testing and the breeders are generally very positive about the prospects.
Information Nucleus Version 2
Following the CRC’s Third Year Review in February 2011 which strongly recommended industry continuation of the INF, the CRC has examined various designs for an Information Nucleus Version 2 (INF2). The resultant design is similar to the current INF, but with the number of sites reduced from 8 down to 2 sites – Kirby in northern NSW and Katanning in WA. The focus will initially be on the major breeds (Merino, Border Leicester, White Suffolk and Poll Dorset) due to the large numbers needed to be measured from each breed to deliver meaningful genomic predictions. Also more progeny will be needed for Merino due to their diversity. Strategies are being explored to determine the feasibility of genomic predictions for smaller breeds.
Information Nucleus data analysis
Work on phenotypic analysis is progressing, with more attention now directed to analysis of parasites and reproduction. The parasite data is challenging, with various ages measured, different worms types etc, and correlations of sire ranking between sites seem to be lower than expected (e.g. based on industry data).
Design and Analysis
Analysis of the accuracy of genomic predictions across breeds and of individuals unrelated to the reference population have been undertaken, required for input into the design of the next version of the Information Nucleus Flock (INF2) to follow the work of the Sheep CRC. Results show that genomic predictions across Australian sheep breeds have low accuracy. Prediction within breed can be high for animals related to the reference population, but even unrelated animals can have an acceptable accuracy provide there are many records of the breed in the reference population. Consequently a design of INF2 has been proposed. Two scientific papers have been drafted and a discussion paper has been prepared.
Further progress has been made with extracting DNA samples of INF progeny with most of the animals from the 2007-2009 drops now extracted and a start was made with animals from the 2010 drop. 2000 genotypes have been added to the database, and another 1000 are planned this month which will complete the number planned for the financial year.
The Genomics Pilot Project 2 is underway with DNA samples to be collected via blood cards. 1000 young sires are planned to be sampled in July from Terminals, Maternals and Merinos.
The Dorper site near Narrogin in WA continues with sires for the second mating identified (10 Dorper and 3 link) and semen assembled for the 2011 AI mating.
A letter and the latest RBVs were sent out in mid March to all Genomics Pilot Project 1 participants.
A commercial Laboratory Information Management system (FreezerPro) has been implemented for tracking of DNA samples from the INF and Genomics Pilot Project 2. All DNA samples have now been relocated to the new freezers in Armidale.
The Information Nucleus Program integrates sophisticated genetic design and analysis with comprehensive measurement of biological and production parameters. It will allow breeders and commercial producers to quickly exploit information on breeding values for new traits potentially supported by information on animals’ genomes to achieve more rapid genetic improvement in their flocks and across the whole sheep industry. The Information Nucleus will also provide the base for core research activities in other CRC programs and enhance the application of results by industry.
The Information Nucleus will test progeny of key young industry sires for an extensive range of traits in differing environments. This information will be added to the Sheep Genetics (SG) database to enhance the accuracy of Australian Sheep Breeding Values (ASBVs) used by industry. The Program will also analyse and use genotypic information on IN progeny and industry sires to estimate and test SNP panels as predictors of breeding value. Genetic information will be generated about new and novel traits and traits that are difficult or expensive to measure on-farm that may be related to wool and meat quality, disease resistance and reproductive fitness. The Information Nucleus will also be the focus of management, wool and meat research being undertaken in the other CRC research programs.
The Information Nucleus will provide progeny test data to Sheep Genetics (SG) that will flow immediately into routine SG genetic evaluations to enhance the accuracy of the ASBVs for immediate delivery to industry. SG has the breeder networks and proven ability to source all of the semen required for the Information Nucleus. Molecular ASBVs will be developed and tested based on data from the Information Nucleus, the previous Sheep Genomics project and industry sires and jointly with phenotypic IN data. This should lead to increased accuracy of selection of young males combined with existing quantitative ASBVs for various traits to improve the rate of response to selection in industry. The result is a potential increase in value of genetic gain of 20-30%, with considerable more gain for some hard-to-measure traits that currently are difficult to improve.
SG’s program for adoption of ASBV involves extensive marketing and training in the seedstock sector and with commercial breeders. This includes information workshops, one-on-one instruction and promotion via industry events and publications. Additionally, a pilot project has been designed to engage breeders in genotyping young breeding animals and use resulting information for breeding decisions. Extension and training activities will be coordinated with Project 5.2 (Industry Training). Strategies and supportive software will be developed to assist them in making optimal genotyping and selection decisions. These programs build pull-through demand for the provision of ASBV by the stud breeders and ensure demand for the products of the CRC Information Nucleus that add value to commercial production.