In the last blog post, I defined Expected Progeny
Differences (EPDs), the accuracy of said EPD, and the percentile rank. An EPD
is an estimate of the genetic merit an animal will pass on to its progeny, on
average. They can be used to sort animals according to their potential to make
genetic change within a herd, or when making bull buying decisions. A
percentile rank simply reports where the specific EPD for that animal ranks
across the entire breed from 1 (top) to 100 (bottom). While I touched on the
value of genomics and how it increases accuracy, to do that justice really
takes a separate post – so here goes.
Remember, EPD stands for ‘Expected’ Progeny Difference. In
other words, it is the evaluation’s best assortation of an animal’s genetic
merit based on the information available at the time. Traditionally, that would
be the animal’s (assumed) pedigree, phenotype data, and progeny information. In
the case of young sires, sometimes that can be limited simply due to their age.
Accuracy, reported on a scale of 0 to 1, is an estimate of
the confidence that the EPD provided is the ‘true’ EPD of the animal. As more
progeny records and phenotypes on an animal are reported, evaluations
inherently know more about that animal. This results in the EPD fluctuating up
and down over time, as the accuracy increases.
Now enter in genomic testing. Simply taking a DNA sample on
an animal takes an EPD and makes it ‘Genomically Enhanced’. But what does that
mean? Well now, instead of assuming that a sire is 25% related to each of his
grandparents, now the evaluation has an exact understanding of which pieces of
DNA he inherited from whom, and how that DNA impacts the phenotypes that animal
will produce. This additional information then increases the accuracy of the
EPD.
But still – what does an increase in accuracy mean? Let me
give you an example:
Figure 1. Comparing accuracies of bulls with and without a
genomic test
|
|
Birthweight
(BW)
|
|
|
Bull
A
|
Bull
B
|
|
EPD
|
-1.7
|
-1.7
|
|
ACC
|
0.39
|
0.15
|
|
Percentile Rank
|
3%
|
3%
|
Figure 2. The impact of accuracy on the 95% confidence
interval and possible range of “true” EPDs
Figure 1 is meant to help picture what is meant by
‘increased confidence’ in an EPD with increased accuracy. The image includes
birth weight (BW) EPD information for two different yearling bulls. Notice that
while they have the same EPD for BW (-1.7 lbs), thus the same percentile rank
in the breed (3%), the accuracy is higher for Bull A than Bull B. This is
because Bull A has been genotyped while Bull B has not. Just looking at their
respective EPDs, both animals would be considered ‘calving ease’ bulls and
would be purchased as such. However, remember that EPDs are a prediction, and
can change over time. Therefore the second image (Figure 2) represents the
possible range (95% confidence interval) of ‘true’ EPDs for both Bull A and
Bull B, given their respective accuracies. Notice how much smaller the range is
for Bull A compared to Bull B. Genomic testing, which results in a higher
accuracy, helps reduce the risk to the buyer and allows them to confidently
purchase a heifer bull, knowing the seedstock producer has done everything
possible to ensure he is indeed a heifer bull.
In my opinion, accuracies are the second-most important
number to consider when selecting a bull for purchase. As more information is
reported, be it in the form of phenotype records or DNA, that estimate can move
up or down as the accuracy of the prediction increases. However, the take home
message here is that the amount of potential movement decreases as the
accuracy increases.
Risk vs. Reward – How
much value to you place on accuracy compared to genetic potential?
Unfortunately,
increased accuracy of an EPD doesn’t equate to “better” genetic potential. Just
as often as an EPD for a trait will increase as a result of genetic testing or
additional phenotype information, it is just as likely to decrease. Hence why
it is most beneficial to purchase yearling bulls who have already had a genomic
profile added to the evaluation. Being the equivalent of anywhere from 3 to 30+
progeny or a 7% to 17% increase in accuracy, depending on the trait, a genomic
test can add almost an entire calf crop worth of information to the prediction,
reducing the likelihood of an EPD shifting over time.
The
question for the commercial bull buyer, however, is what is the value of that
increased confidence to you? Yes, increased accuracy reduces the risk of the
EPD changing over time, but that must always be balanced with the EPD on the
animal and how much “risk” you’re willing to take. For example, if there are
two bulls being considered, one with a WW EPD of +68 lbs and another with a WW
EPD of +58 lbs but with a genomic profile on file, which would you choose? A
more risk averse person would likely prefer to purchase the second bull with
the lower WW EPD because they are more confident in the certainty of the
prediction, while a producer who is willing to accept a bit more risk would see
the value in purchasing the first bull understanding that 30 calves at +10 lbs
of WW is over 300 lbs of added value, assuming the EPD prediction doesn’t
change over time. This is the push and pull of understanding the value of the
information available to you and being able to use it as a tool to make more
informed decisions.
In my
opinion a bull with a genomic test, no matter the difference in EPD, is more
advantageous than one without. But mainly because of the large leap in
confidence that we see in the accuracy as a result. The true take home message
here is to not split hairs over small differences in that value, especially as
a result of a few progeny records being added to the evaluation on older,
proven bulls. As with anything, the values placed on these animals are a tool
to help provide insight and expectation to an otherwise unknown outcome. It is
up to you to decide what value you place on specific pieces of the puzzle and
how much ‘change’ or risk you’re willing to accept.
Comments