2017 Analysis Upgrades
The LAMBPLAN and MERINOSELECT analyses are continually being developed to improve the genetic evaluation and supply the best tools possible to producers, aiming to help improve genetic gain across the sheep industry. Major updates to the Sheep Genetics analyses are implemented once a year to minimise any disruption that may result from the changes. This year there are significant updates to both LAMBPLAN and MERINOSELECT. It is important to understand the enhancements made to the analyses and why you may see changes to the ASBVs of your animals. All updates detailed below will be applied to LAMBPLAN on the 15th of April and MERINOSELECT on the 21st of April.
2016- SINGLE STEP CARCASE ANALYSIS
Until 2016 all genomic information was incorporated into ASBVs using a “blending” approach. Blending combined pedigree based estimated breeding values with genomic breeding values from separate analyses to create a blended ASBV.
In May 2016 Sheep Genetics moved to the “Single Step” approach to calculate ASBVs for carcase traits to improve the application of genomics within this sub-analysis. The Single Step analysis combines genomic and pedigree information with performance records in a single model to calculate ASBVs for all traits in the model.
Updating to the Single Step sub-analysis for carcase traits allowed genetic correlations with related traits to be used more effectively. This resulted in a larger number of animals having eating quality breeding values estimated.
To compliment the Single Step carcase analysis, two new indexes were introduced with a focus on joint improvement of eating quality and lean meat yield for terminal sire breeders.
2017- SINGLE STEP IS HERE
Since the implementation of the Single Step carcase sub-analysis, we have been working on the implementation of Single Step for the main analyses in the TERMINAL, MATERNAL and MERINO runs.
In 2017 we make the transition to using the single step approach over all main analyses, as well as the WEC analysis for all three breed groups and visual trait analyses for Merinos. The reproduction analysis will remain unchanged at this stage due to the limited records available. However, there is ongoing work to develop genomically enhanced Research Breeding Values (RBVs) for reproduction traits later this year.
The benefits of moving to a full single step analysis include:
Simultaneous use of all information – no need for blending.
More accurate genetic relationship between animals identified and utilised, meaning better use of information from similar animals, with the possibility of correcting pedigree errors.
Genomics used for a wider range of traits and the genetic correlations between all relevant traits are now considered.
The accuracy of ASBVs better reflects how effectively genomic information contributes due to genotyped animals relationship to the reference population.
Genomic information contributes to the calculation of relatives ASBVs, i.e. the siblings or progeny of a genotyped animal will benefit.
Industry data now contributes to the reference population.
While we are confident that the changes to the analyses will give better estimates of animals genetic merit you will see some changes to ASBVs. The correlation between the old and new analyses is very strong, although some animals will change for some traits. Changes you will see in ASBV’s are due to the following:
Influence of genomic information from a larger number of animals and traits.
Impact of correlated information from a larger number of traits.
The carcase traits being moved back into the main analysis, so that all animals have breeding values estimated for carcass and eating quality traits. The addition of the carcass and eating quality data into the main analysis also has a correlated impact on the other production traits.
Genetic groups will be fitted using a different methodology
There are still some limitations to be able to use genomic information in the analyses. We can only use genotypes from four major breeds, Border Leicester, Poll Dorset, White Suffolk and Merino. We are continuing to collect genomic data from other breeds, but at this stage we cannot include these in the routine genetic evaluation.
The value of genomic testing to the genetic evaluation is still determined by linkage (genomic relationship) to reference animals and other records available on those animals. Some animals from relatively closed flocks that are not well represented in the reference population may end up with low accuracies for breeding values that may not pass reportable thresholds.
Animals will still need to have some pedigree and performance included in the analysis in addition to a genotype. Having only a genotype on an animal is currently not enough information for ASBVs to be calculated.
Single Step across the main analyses would not be possible without a significant upgrade to the computer software used. Run times with the existing software were too slow to complete the analyses every two weeks. A new program which “solves” the analysis has been developed and we are now able to run these more demanding Single Step models in acceptable run times which will be sustainable for some time to come.
Single step is a major achievement
The full implementation of single step is a major achievement and is the result of a long term collaboration between Sheep Genetics, AGBU and the Sheep CRC. We have large multi trait across breed analyses, where there are many genetic groups fitted and significant numbers of animals genotyped. Despite this, Single Step has been implemented without compromising run times. The advancements made to the analysis are world leading and result in more accurate, reliable breeding values on more animals.
Through across flock genetic evaluations, the sheep industry has made significant genetic progress in growth and carcase traits in young animals, as well as wool production. The other area needing increased effort is maternal efficiency, including reproduction.
The reproduction analysis is currently run as a separate sub-analysis. While the reproduction analysis is not yet being rolled over to use Single Step with the April changes, improvements will be made to the validation of data in the analysis. The main change is that if a flock has 100% survival in their twins within a year (i.e. all twins born are then raised) that year drop of lamb survival records will be removed from the reproduction analysis. This will improve the information being used to identify lamb survival. For these groups, although records on number of lambs weaned will be removed, records on number of lambs born will be retained.
To assist breeders to understand how much of their data are being used in the reproduction analysis an additional report has been developed. This report summarises how much data was used from each drop in the analysis, also highlighting areas in which the data recording for reproduction might be improved.
Sheep Genetics and AGBU are aiming to release RBVs for component traits of reproduction based on the Single Step model later in 2017. These traits include Fertility (whether or not a ewe became pregnant), Fecundity /Litter size (how many lambs she had) and Survival (how many of those lambs survived to weaning). Information on these three traits and joining details can be entered through the Mating Module. We are encouraging breeders to adopt the Mating Module system to improve the general recording of maternal efficiency traits.
- Single Step for Main Analyses.
- Performance records are still needed to calculate ASBVs.
- You will see ASBVs change from upgrades.
- More accurate and reliable ASBVs will be estimated.
- These upgrades are world Leading in genetic evaluation.
- Reproduction analysis will not be single step
- More data is needed to develop individual repro ASBVs
For further information, we recommend that you:
- Watch Daniel Browns presentation from Leading Breeder “Development and Future Plans from AGBU” which can be found on the link below.
- Attend a Regional Forum in a town near you.
Leading Breeder Presentations
Regional Forums click here to register.