What happened when nobody was paying attention and other things to make your head explode.
Understanding the ARI – EPD program can appear daunting. It is unfortunate that much of the “discussion” to date has been argumentative and has taken the focus off the reality that EPD’s do work. Some who disagreed with the initial decisions made to get the program up and running felt a need to voice their negative opinions. It also didn’t help that “moneyed interests” may feel a bit threatened with transparency provided by EPDs. Looking back at some of the “chat site” archives is like a watching a discussion taking place around the wood stove at the old general store exclaiming what brand of tractor is best…a great deal of emotion with very little substance. What is even more surprising is that I can’t find anything being discussed about how this science has played a big role in improving other livestock industries. Nothing is really even discussed about the potential uses or goals for the alpaca industry. It actually feels like a vacuum when it comes to finding anything alpaca related on how EPDs are actually used. So, it isn’t surprising that most of the questions I get about our breeding program and the use of EPDs has to do with exactly that question. The questions I hear the most are related to how EPDs will benefit a small breeding program and what to do with the information if they had it. Hopefully, I can illustrate what the process is. I am not a scientist nor claim to be an expert by any means… far from it. I became an advocate of how science could help our small breeding program after attending a few seminars put on by NW Alpacas and Snowmass Alpacas about 5 and 10 years ago. Since then I’ve asked questions, read and researched on the ol’ Inter-Web. Here’s my take on EPDs. Hopefully it will help.
It may be of help to start with a definition of what EPDs are. This is Wikipedia’s definition of Expected Progeny Difference. “EPD are an evaluation of an animal’s genetic worth as a parent. They are based on animal models which combine all information known about an individual and its relatives to create a genetic profile of the animal’s merits. These profiles are then compared only to other individuals of the same breed.”1 For me, this definition is perfect because in this article I am trying to tackle the “animal’s merits” and comparing them to others.
I frequently hear that EPD trait vales are confusing when Fluffy’s first histogram comes in at an AFD of 18 microns but her EPD trait value for AFD is a “plus one” which predicts that Fluffy’s genetics for AFD is stronger or coarser than the average of the whole EPD group. Keep in mind that histograms describe phenotype. What that animal looks like at a point in time. EPDs predict the genetics that will most likely pass on to its offspring. I say “most likely” because an EPD prediction is only relevant 68% of the time. EPD predictions are “wrong” 32% of the time. So, when the prediction is “wrong”, the actual results could be better or worse than the original prediction. So, I look at the EPD prediction as that I have an 84% chance that the outcome will be as expected or better. I like those odds. Anybody could win in Vegas with those odds.
How EPDs are calculated is complicated at best. It hurts my head just thinking about all the math that is needed to put this all together. That’s what computers are for. Besides, it would probably make me cry if I tried to write it down. So, here is my simple summary. EPDs are calculated using a statistical science called BLUP. It is the same science that we’ve seen on TV where the investigators can clear up a noisy recording and end up with a clear voice or sound. BLUP tries to resolve what is related and what is not relevant. I don’t need to know how they do that. I’ve seen it catch criminals on TV, that’s enough for me. If you need more information Wikipedia and Google have articles and links to click on till your heart is content.
But that is only part of the story. What really makes EPDs interesting is the ability to compare animals being raised in different locations of the country, under different environments that are of different age and sex. So how is that done? BLUP also uses a science called connectedness. The program looks for a path to connect all the animals in the program and make a prediction based upon what it has learned. I have included a simple visualization of what connectedness is. (Connectedness Graphic) East Farm has animals 
from A & B bloodlines. Central Hacienda has animals from B & C bloodlines and West Ranch has animals out of the C & D lines. It isn’t difficult to grasp that science could evaluate environment differences in the farms that share common ancestry like East Farm and Central Hacienda share B bloodline genetics. But connectedness can also determine connections from perceived disconnected parts like A and D through their environment connections with B & C. Obviously this is a very big simplification of very complicated process. If you reported your clip data to ARI this year, you may have noticed more attention being directed to imputing data for animals that are cared for at different locations. There is a link below if you would like to learn more about Connectedness 2.
Has this helped so far? I hope so, because my head is about to explode. I get the same feeling when I look at trait values by themselves
without other trait values to compare them to. A trait value is just a number. Alone, by itself, it really doesn’t tell me much. It helps to have other trait values to compare so I have an idea of how that animal fits in the overall picture for that specific trait. Also, in the ARI EPD database, not all data is available to the public. Owners can choose if they want an animal’s data made public. So, what I find most useful in my research is the use of Trait Rank. An animal’s ranking for a particular trait can be determined by using Table 7 in the Producers Report. In this table, every EPD Trait Value has an associated Rank. 
For me, seeing an alpacas AFD Trait Value is in the top 1% is easier to categorize than a “-2.257” Trait Value for AFD. (Trait Value Graphic) Ranking provides a quick visual on where an animal fits in according to the whole population and will help simplify your research for breeding and purchase decisions. Utilizing ranking also very helpful in avoiding the issues involved in single trait breeding. If you like show results as a method to rationalize some decisions you will like ranking your traits. Ranking also facilitates more complex sorting of large groups of animals I have a program that will calculate ranking per Table 7 of the latest Producers Report. I am available if you have a lot of animals that you would like to get converted to ranking.
Earlier I stated that “EPDs predict the genetics that an animal will most likely pass on to its offspring”. Then what? I have EPD Trait Data on my herdsires and breeding females. What’s next? How does this help predict what the offspring will be? This is an area that I find many loose the connection of how EPD’s can help with their breeding decisions. The genetic EPD values of the parents don’t predict the Phenotype of the offspring. They predict the offspring’s EPD’s. With EPD trait values of the offspring, in advance, it is easy to resolve if the mating will help you reach your goal. If you don’t have a goal EPDs should help conclude the appeal that animal will have in the marketplace. Will a male offspring have herdsire appeal? For me, getting a general idea of the results is good place to begin. I 
just add the trait values of the Sire and Dam and divide by two. For the most part, the offspring’s EPD Values will be an average of the parents. In this method we are not taking into account heritability rates, differences that sex will determine and more, but then again, I never went into this looking for an exact number. I just want to know that I am maintaining or achieving progress toward my goal. I do this calculation on every breeding that I am planning and it helps me narrow the field of prospective herdsires. I have an example for you to view in the Marilyn Monroe graphic. As you can see from the ranking of the offspring, this alpaca could be one generation awayfrom elite offspring production. EPDs are a great tool for a bred up program.
I don’t just rely on the numbers. Other factors also help with making our decisions. The point is, in the 21st century, the art of successful breeding, should include knowing what science to use. Tomorrow, many more advances will be available such as: EPDs being considered in show ring placements and DNA Enhanced EPDs for higher accuracies. I believe EPDs are the best thing to happen to alpacas. Don’t they finally deserve something good to happen for them? I believe EPDs will provide the most predictable method to help bring alpacas back to the prominence they should have held in the world of natural fibers over the last 500 years.