I run a registered Angus commercial herd and I DNA test all of my bulls. This year I DNA'd a group of cows that had 10+ calves with data collected in contemporary groups of 30 plus. I was really impressed with how close the DNA evaluation came to predicting the actual data. For instance, the DNA was able to predict the cows with heavier birth weights and cows who milked better giving me heavier calves. With the big drop in price, I am seriously considering testing my heifers before I make replacement decisions in the future. Unfortunately this year I sold all but 5 heifers for Fall delivery shortly after their birth. I only AI a few cows and my heifers and I made the decision to keep those AI sired heifer calves. Looking at them now, I kind of wish I had selected differently. Testing would probably make me feel worse, so I will save that for next year.
I have a 4 year old bull that I selected to breed cows. He had an exceptional weaning weight himself, however he was a 99 pound calf and out of a line of cows known for heavy birth weights and heavy weaning weights. His mother was an exception to the birth weight prediction, with most of her calves weighing less than the average. He was sort of a throw-back in BW to the weights I saw with most of that line of cows. He was DNA'd as a yearling. His initial EPD put him at a -3 for calving ease and his genomic profile put him in the bottom 10% of the breed for BW, but the top 22% for weaning. At that time I did not see accuracy in many of the genetic profiles I looked at, but those numbers kind of matched my own expectations. When he received his initial genomic profile there were far less animals in the data base and animals that were not from the most popular and tested cattle lines could have less accurate profiles. He now has two years of calves and to my great surprise his calves were born lighter than the calves of AI bulls with low birth weight EPD's, but weaning weights were pretty average. Since his genomic profile was first run four years ago, there has been a huge increase in the number of animals tested and accuracy has greatly improved. His genomic profile which originally made him appear to be unsafe for heifers, but capable of producing heavy weaners, now shows him as a heifer bull who will likely have average weaning weights. In other words, the genomic profile matches the data collected on two calf crops. Had his initial profile been the one I see now, I would have used him on heifers from the beginning.
I myself will no longer use a bull that has not been tested. It is too easy to manipulate EPD's by reporting inaccurate or selective data, but it is not possible to manipulate a genomic evaluation. It is very important that data continue to be collected and turned in if we want those genomic profiles to predict accurately. If everyone discontinues actual data and only does a genomic profile in order to get EPD's, the genomic profiles won't have the information they need to match phenotype to genotype and genomic accuracy will deteriorate. In my opinion, genomics predict some things well, but some others I still don't trust.