Many beef producers will recognise the expression that when looking at an animal, “70 percent of what you see is a result of what’s gone down his neck.”
It can be easy to take this as just another old saying that is used in conversation. And like many old sayings, the truth that lies behind it is often overlooked or not really considered.
In looking at the physical appearance of a beef animal, whether it be a new sire on sale day, or selecting replacement breeders, it is essential to remember that the average heritability across all the economically important traits falls in a range from 25pc to 40pc.
That means that the environment (which includes not only nutrition, but management, health or other factors) explains the 60-75pc of the physical (or phenotypic) variation.
This matters when producers start to place much greater emphasis on raw data when making their selection decisions. It’s not uncommon for producers to spend more time comparing bulls based on a supplementary sheet of raw numbers, than on the genetic information that is provided within the sale catalogue.
This emphasis often extends to commentary at industry events and frequently on social media, where claims that raw data, or raw data and pedigree, have more value than performance information.
In making this argument, producers are saying that they are making selection decisions on the 60-75pc variation that has nothing to do with genetics. The things that contributed to that variation, feed, health, management cannot be bred from, and cannot be passed to the next generation of calves.
While recognising this, those producers will often quickly point out that they know the animal’s pedigree and that combining this knowledge with raw data, they can reach a better selection result. However, there are also some serious flaws in this approach.
Pedigree a starting point only
It may surprise many producers, but pedigree is actually the weakest form of genetic information available to a buyer. All it really does is to identify who the parents were.
After that, there is no more information since a pedigree never updates. It contains no performance data, no adjustment for environment, and no measure of how much confidence to place in the estimate. So, in simplest terms, a pedigree serves as a starting point.
This is very different to genetic information provided by EBVs. All EBVs include pedigree and then expand the information with the animal’s own performance records, data from siblings and progeny, contemporary group comparisons, and connections across herds.
Arguing for pedigree over EBVs is therefore an argument for less information, not more. The producers making that argument are choosing the thinner dataset.
Beyond over-reliance on pedigree, there are specific risks that come with placing heavy emphasis on raw data. These risks can push selection decisions in a consistently wrong direction.
Regression to the mean
One of the most significant risks for a herd from this approach can be described as regression to the mean. This can happen when a producer chooses an exceptional animal, based on phenotype, pedigree and some raw measurements.
However, that exceptional performance is the result of components. The first, genuine genetic merit, and the second is the contribution from environment, management, and biological luck.
When that animal breeds, only the genetic portion is passed on. The seasonal conditions that produced the performance don’t repeat. The careful preparation doesn’t transfer. And the biological luck may not occur again.
The result is that progeny will on average perform considerably below the phenotypic level of the parent. That lower performance is not because something went wrong, but because the non-genetic portion of that exceptional performance simply evaporates.
The more exceptional the raw phenotype, the more of that exceptionalism is likely to be environmental rather than genetic.
This often leads to greater disappointment in the next generation. Producers who repeatedly select on impressive raw data and are repeatedly surprised when progeny underperform are not experiencing bad luck. They are experiencing regression to the mean.
Quietly going backwards
The second risk is that raw data selection makes it impossible to know whether a herd is genuinely improving, standing still, or quietly going backwards relative to the broader industry.
Without an external reference point, improving management can feel identical to making genetic progress. Producers can invest more in feeding, preparation, and selection by eye, while their genetic merit stalls across generations. As a consequence, the gap between their herd and the genetic trend widens invisibly.
Supplementary data sheets
This is precisely what a supplementary data sheet cannot correct for. It records what happened to that animal, in that environment, under that management, in that season. It tells a buyer the bull performed well under the conditions the vendor provided.
It does not tell how much of that performance was genetic. Nor can that data really be used to determine how his progeny will perform in a different paddock, under different conditions, in a different season. It cannot be meaningfully compared to a bull prepared differently on a different property, because the environments are not the same.
This is not to say that raw data has no role. Raw data does provide useful context about how a bull has been managed, and it feeds into the dataset that generates EBVs.
But as a standalone selection tool it is incomplete by design. It simply cannot tell a buyer what genetics the bull will transmit to his calves.
EBVs change because of new data
A common criticism of EBVs is that they change over time. Producers often use this as evidence of unreliability.
In fact, the opposite is true. EBVs change because new data comes in and the system updates its estimate accordingly. This is what a well-functioning information system should do.
By contrast, a pedigree does not change. It is frozen at the moment of birth and receives no new information.
The other key factor to consider is that EBVs also come with an accuracy value, unlike raw data.
Accuracy tells the buyer how much weight to place on the estimate, and how much it is likely to change as more data arrives. A low accuracy EBV on a young bull will move as progeny data comes in. That movement is expected, and those movements show the system working as designed.
However, the large degree of uncertainty that exists in raw data because of environment effects is equally real. The difference is that it is hidden, which makes it considerably more dangerous.
Advocating for raw data and pedigree over EBVs is not a more cautious approach. It is a more restricted one.
For many producers, familiar bloodlines and past performance feel like the safe choice. But it means selecting predominantly on environmental effects that will not repeat, from a reference point that cannot be verified, and with no way of knowing whether the herd is improving or falling behind.
However, producers making these choices risk compounding genetic losses across generations. More crucially, this slowing rate of gain may go undetected because those tools which would show a reduction in genetic gain have been dismissed as unreliable.
You can’t pass a feedlot or a good season down to the next generation. You can only pass genes.
SOURCE — BEEF CENTRAL WEEKLY GENETICS REVIEW WRITTEN BY GENETICS EDITOR ALISTAIR RAYNER