You would expect the genomic score to be more accurate, but you would be wrong. My original thoughts were…people can manipulate data, but you can't manipulate DNA, so the genomics should be more accurate. Right? It turns out that I was wrong. The DNA may not change, but our understanding of what the DNA represents is constantly evolving. The EPD's include the most recent genomic information and are updated every Friday, but genomic scores are only updated for you to see once every year.
Kelly from American Angus explained it to me. When you run a Genomic test, the SNP's, of that animal are compared to those from hundreds of thousands of others. Each SNP represents a difference in a single DNA building block, called a nucleotide. Scientists somehow determine that certain SNP's are responsible for certain traits, so it is assumed if the SNP's of a certain bull or cow are similar to those of others with a certain trait, then they will share that trait. Assumptions are not always accurate. At one time this was done manually, but now computers are able to analyze and calculate similarities by analyzing thousands of pieces of data at the same time. EPD's are recalculated every Friday using the actual data turned in by producers and looking at the information provided through the genomic analysis. The EPD's are always being refreshed, however the numbers you see in a genomic score are not updated until the computers make an annual run adding all of the information collected over the prior year. The genomic scores you now see on your animals when you log in to AAA were just updated May 28th, so are pretty current. Those numbers will be different than what you saw over the last year and some may change quite a bit. When you find a genomic score for an AI bull on your computer, it could be years old if the stud does not update their information. The EPDs on the other hand will have changed very little, if at all, when those genomic numbers were recalculated and reposted on the 28th.
The fact is, even though DNA doesn't change, we are still learning what that DNA does, and original assumptions are often found to be inaccurate. If you see a genomic score on a bull that says he is in the bottom 20% for Heifer pregnancy, but his EPD puts him in the top 20%, it is because the data since that genomic score was run has changed expectations. With something like heifer pregnancy, there is less data to analyze, so these numbers are still likely to keep changing. Birth weight has more actual data and is less likely to move. Weaning weight is one that can be off in your environment. Environment plays a huge part in weaning weights, and I expect most data on this comes from purebred breeders and cows that are treated way better than the cows in the average large commercial herd. For that reason, I expect weaning weight EPDs and genomic scores should often be suspect. I, for example, have an eleven-year-old cow with a weaning weight average of 105% in my herd on 9 calves. She has a genomic score that puts her in the bottom 16% for weaning weights, but in my tougher environment she easily out weans cows with better scores. In past years her genomic score has put her as low as the bottom 1% for weaning weight.
I have found that most of the genomic scores I first received in 2015 have changed dramatically, and today's scores are definitely closer to my actual data. As an example, a bull I ran a genomic test on in 2015, had a score that put him in the bottom 15% for Birth weight and his birth weight EPD was + 3.2. I expected this based on his own performance and he was chosen as a growth bull to be used on cows. His genomic score for weaning weight put him in the top 22%. He sired about 100 calves, and we found his to be some of our smallest calves, but weaning weights were unimpressive. His genomic score now shows him in the top 9% for low birth weight and his BW EPD is at .5. His WW genomic score is now 89 and his WW EPD has dropped to 39 pounds.