Genomic Genetic Evaluations Have Arrived
Posted: February 16, 2009
Genomic sire evaluations were released for the first time in January. Many people believe this will have the largest impact on genetic improvement programs since the advent of frozen semen. Cows and bulls can now be genomically tested for approximately $250 with the new Illumina BovineSNP50TM BeadChip.
This DNA “chip” was developed in collaboration with scientists from Illumina, USDA, University of Missouri, and the University of Alberta. The cattle genome is a code that consists of four letters (or nucleotides): A, C, G and T. Each parent contributes roughly 3 billion nucleotide pairs which are packaged on 30 chromosomes to their offspring. Variation in this code is responsible for much of the difference in performance among cows. For example, a “G” at a specific location in a sire’s DGAT1 gene (located on the 14th chromosome) instead of an “A” is associated with an increase of approximately 0.15% fat and a decrease 300 pounds of milk in his daughters.
Scientists identified 54,000 of these single nucleotide polymorphisms (called SNPs or snips) spread evenly throughout the cattle genome. In contrast to the previous example, a majority of the identified SNPs have a very small effect by themselves, but cumulatively they can predict performance accurately. After building the DNA chip, USDA scientists went to work to determine how each SNP was associated with 29 different production, type and health traits. They did so by associating SNP information from over 7,000 Holstein sires with their traditional progeny test genetic evaluation.
For bulls with both a genomic evaluation and traditional progeny test evaluation, the two sources of information are combined into a single sire proof. For bulls with many daughters, genomic results contribute very little to their proof. The biggest advantage for genomic evaluations is obtained for cows and for those bulls with no daughters. We have relied solely on pedigree information in the past to predict a bull’s genetic merit when he had no daughters milking. Now we can combine parent information with genomic results to significantly increase reliability. A Holstein bull with 0 daughters is expected to have reliability of 60% to 70% for all traits. In the past, such a bull would have reliabilities of 35% or less.
These changes will have a very large impact on genetic selection programs. Few bulls were actively marketed until they were nearly 5 years old in the past. Now, we have a reasonable idea of a bull’s genetic potential before he is sexually mature. Studs will genomically test all bulls before they enter their progeny test program. Some bull studs will reduce the number of progeny test bulls in their program, and may forgo progeny testing entirely at some point in the future. Genetic progress will increase more rapidly in the future because the interval between sire generations will decrease. Previously, bull studs waited until a bull had a progeny test (5 years of age) before they used a bull for contract matings. They are now using these bulls as soon as they are sexually mature. The change in generation interval is going to make genetic evaluations for genomically proven bulls look too good to be true over the next year or two. We made a big leap forward because we are now looking at bulls that, traditionally, would still be 3 years from having their proof.
Genomic testing will also have a large impact on marketing of cows. Our confidence in a cow’s genetic evaluation will be greatly enhanced because of genomic testing. The elite genetic cows will increase in value because of this increase in confidence. Unfortunately, that also means that some current high index cows will lose substantial value because of an inferior genomic profile. If I am buying the pick of a flush, I would surely have the calves genomically tested before making my selection.
While genomic evaluations are a big step forward, they are not perfect. Reliability estimates for breeds other than Holstein are not as favorable. Gains in reliability for Jersey bulls are less than ½ those observed for Holstein, and Brown Swiss bulls have experienced very little gain due to genomic testing. It requires a large number of progeny tested bulls to accurately estimate the effect of a SNP. While over 7,000 animals were used to derive genomic estimates for Holsteins, less than 2,000 Jersey and less than 500 Brown Swiss were available. Efforts to combine breeds into a single genomic evaluation haven’t been successful at this point. Genomic evaluations will improve for the color breeds, but it will require time.
While 60% reliability without a single daughter for a trait like daughter pregnancy rate is good, most producers are used to higher reliabilities. It takes over 1,000 daughters to reach 90% reliability for daughter pregnancy rate, and the gain realized with genomic evaluations is equivalent to perhaps 100 daughters. New Zealand stopped marketing one of their genomically tested “bull teams” because daughters were not performing up to expectations based on their sire genomic test once they began milking. They had multiple “bull teams” and the others were performing as expected, but it does highlight that the reliabilities are lower than traditional evaluations. In other words, genetic evaluations from progeny tests will remain the gold standard for the foreseeable future. Because these bulls will have lower reliability than traditional progeny tested bulls, it is advisable to spread risk by using more bulls. Some companies are marketing bulls that have a genomic evaluation and no daughters in “6-packs” to help spread risk effectively.
Genomic evaluations are here and they will help accelerate our quest to breed for more profitable cows. We can be confident in these evaluations when we look at a group of bulls, but need to recognize that the reliabilities are lower than progeny testing. Don’t spend much time worrying whether a bull’s proof is from a genomic test or a progeny test. Continue selecting bulls that meet our goals and manage the risk of lower reliability by using a larger selection of bulls than in the past.
Chad Dechow, Assistant Professor of Dairy Cattle Genetics, Department of Dairy and Animal Science