HOW fast we make genetic improvement in our own breeding programs and as a breed is heavily influenced by our accuracy of selection.
Higher accuracy of selection simply means we can make the right decision more often when we are choosing the best animals to breed from. The role of Hereford BREEDPLAN is to give us the highest possible accuracy of selection for the traits that we can measure objectively.
We are in a fortunate position because Hereford BREEDPLAN is backed by an extensive research and development program of work. This R&D is conducted by ABRI and AGBU, but also by a range of other organisations supported by MLA’s National Livestock Genetics Consortium. The alignment between R&D and the delivery of BREEDPLAN means that as soon as new R&D is completed and the outcomes are implemented into the genetic evaluation, Hereford breeders get the benefit with higher rates of genetic improvement based on better accuracy of selection.
As a result of the most recent developments, we will be implementing several updates to the Hereford BREEDPLAN evaluation later this year. As with all updates to the evaluation these developments have been extensively tested and we are confident the updates will give Hereford breeders better outcomes and enable better rates of progress.
Of course, any changes to the evaluation means some breeding values will change and this means there will be some reranking of animals. To minimise any disruption that this may cause the changes are due to be implemented as of the October BREEDPLAN run.
The key changes are outlined below. If you would like any additional information about these updates, please contact the office.
Genomic Pipeline update
The Hereford BREEDPLAN analysis makes use of genomic tests of various sizes, typically from 30,000 SNPs to 100,000 SNPs.
These genotypes are then imputed up to a much larger SNP list for use in the BREEDPLAN evaluation. Imputation is best described as filling in blank spaces in a sentence or story, if we have read a similar story before it is much easier to fill in the blank spaces.
Similarly, imputation of genotypes relies on having a library of genotypes for similar animals. Updating our library means that we can impute genotypes with greater accuracy.
As more genotypes are now available for the Hereford breed, we will be updating the library used for imputation. This will mean some improvement in accuracy and means a small number of animals that were previously having their genotypes excluded from the analysis will now have that information brought in.
Alignment between the current library and updated library is high, therefore changes to EBVs because of this update are expected to be relatively minor.
Lambda update
Single step BREEDPLAN analyses use pedigree, performance, and genomic information simultaneously to calculate the highest possible accuracy EBVs using all available information. In this type of genetic evaluation genomics contributes by telling us the number of genes individual animals share, or in other words, it is giving us a much better understanding of the relationship between animals. The challenge with this approach is the relationship you get from pedigree records and the relationship from genomic information are not the same.
The diagram (Figure 1) shows the relationship between an animal and its grand-progeny. Based on pedigree alone we have to assume that all grand-progeny share the expected 25 per cent of their genes with their grand-parent. Once we have genomic information, we can see that there is variation in how closely related each of those animals is to the grandparent, in this case a range from 23 to 27 per cent. Understanding this difference allows more appropriate weightings to be applied when using relative’s performance to calculate higher accuracy EBVs.
Lambda is the name given to the weighting used to combine the relationship information from these two sources. Since Hereford BREEDPLAN was first introduced in 2017 a lambda value of 50 per cent has been used. More recent work conducted by AGBU has shown significant advantages in moving to a lambda value of 95 per cent.
Doing this means the relationship value between animals used in the analysis aligns much more closely with the genomic records. More importantly it has been demonstrated to give us higher accuracy EBVs as well as improvements in other quality metrics for stability and bias.
The scatter plots on Figure 2 show the impact of changing lambda on the accuracy of EBVs. The blue dots represent EBV accuracy for animals out of an analysis using pedigree only and not genomic information. The orange dots represent EBV accuracy for animals using the current model with lambda at 50 per cent and the green dots represent the new analysis moving to lambda at 95 per cent.
As shown in the plots the impact differs between different traits depending on the heritability of the trait and the size of the reference population for that trait. As an example, we see greater improvement in accuracy for Weaning Weight in comparison to Days to Calving as Weaning Weight has a significantly higher heritability and we also have a bigger reference population.
The other point to note is we see a much bigger improvement for the animals that have low accuracy EBVs in the pedigree only model. For example, we see a much bigger improvement in accuracy for young animals that don’t have any phenotypes recorded in comparison to sires with progeny information already on file.
This is worth considering when planning the genotyping strategy for your herd. Genotyping calves earlier on in life will give you higher accuracy EBVs early on in life that will also be more stable and less likely to change as more information becomes available.
Calving Ease Analysis update
Calving Ease is a challenging trait to include in a genetic evaluation because of the categorical scoring system and the low frequency of calving problems. Calving Ease has to date been analysed in a separate sub-analysis using calving difficulty scores, birth weight and gestation length.
It has not been possible to use genomics as part of the Calving Ease analysis as appropriate methodology has not been available. A new model has now been developed that will allow Hereford BREEDPLAN to move to Single- Step Calving Ease meaning that for the first time genomic information will contribute to calculating Calving Ease EBVs.
The changes to the Calving Ease analysis are based on a new solver (software) being used but will also mean the inclusion of genetic groups and changes to contemporary groupings.
The benefit of these changes is there will be a large increase in the number of animals included in the Calving Ease analysis, there will be a significant increase in the accuracy of Calving Ease EBVs and these EBVs will be more stable with less bias.
As shown in Figure 3, these changes will mean that we will see a much larger range in Calving Ease EBVs, and we will see significant changes for some animals.
The scatter plot shows results from the existing analysis along the horizontal axis against results from the new analysis on the vertical axis. The yellow shading on the chart indicates that most animals sit in a range where there are not particularly large changes between current and new analyses. But the darker blue dots show that there are animals that will change significantly.
It will be important to keep in mind the new analysis will give us more accurate Calving Ease EBVs that are more reliable and for more animals. This means as a breed we will be well placed to make better progress for this economically important trait. It will also be important to remember the new Calving Ease EBVs should only be compared to the breed average and percentile bands table from the same analysis. The previous benchmarks will no longer be relevant.
Updated indexes
At the 2023 Herefords Australia AGM, the board announced a new strategic direction for the breed focussing on sustainability. In developing the approach for addressing this from a breeding program perspective we have considered what can be done now, what comes next as a focus for the medium term, and what needs to be done in the longer term.
What we can do in the immediate term to address sustainability priorities is update Hereford indexes so responses to selection decisions in our breeding programs have better impacts on sustainability. This can be done using traits that we already have access to.
In the medium term our focus shifts to capturing better information to improve the accuracy of selection for existing traits important to improving sustainability, for example mature cow weight and net feed intake. Improving the accuracy of EBVs for these traits means greater responses to selecting based on the index. The longer-term strategy is to then develop breeding values for new traits that address sustainability such as a direct measure of methane production.
As part of the immediate term strategy, we have been working with AGBU to develop two updated indexes for the Hereford breed. Methane production is the biggest sustainability indicator for us to use, it represents lost or wasted energy. More efficient animals and more efficient production systems produce less methane.
However, we need to be conscious of how we define methane production. We can talk about it in two ways; total methane produced, and methane intensity or the amount of methane produced per unit of product. To improve both sustainability outcomes and continue improving productivity we need to include both metrics in our breeding objective.
The development work that has been done using BreedObject to update indexes has shown this can be achieved. By increasing the feed cost used in the model it puts more pressure on traits that influence total methane production, such as mature cow weight and feed intake. At the same time, it retains as much as possible of our improvement in productivity traits that influence methane intensity such as growth and reproduction.
The bar chart (Figure 4) shows the changes in selection differentials for different traits for the Southern Self Replacing index as feed cost is increased by 10, 20 and 30 per cent. The most acceptable option is increasing feed cost by 20 per cent. The outcome is that we end up with an index that is quite balanced across the major trait groups impacting on profitability as well as sustainability.
The index puts more downward pressure on cost of production through mature cow weight but has similar pressure on improving cow/calf survival, growth, reproduction, carcase weight and eating quality.
Changing the Southern Self Replacing index to a Sustainability index with an additional 20 per cent feed cost gives an index that is strongly correlated with current values with a correlation of around 99 per cent.
This would lead to some reranking of animals. However, the other changes to the analysis, particularly changes to the Calving Ease analysis, will have a flow on impact on index values. Once changes to the analysis are accounted for in addition to changes to the index, we will be expecting a correlation of 92 per cent.