Marker Validation

Marker validation means different things to different people. For example, the National Beef Cattle Evaluation Consortium (NBCEC) http://www.nbcec.org/nbcec/ucdavis/validation.htm considers marker validation to be the process by which the claims made by the commercializing company are validated. To accomplish this, a marker would be genotyped in an appropriate set of populations and the produced genotypes would be statistically tested for an association with a particular phenotype or set of phenotypes.

We view marker validation somewhat differently. Ideally we want to identify the mutation or mutations within a gene that directly cause an effect on phenotype. If we are able to do this, the test should have the greatest value to industry since it is expected to work in every population in which variation in the gene is found. If it is discovered not to work in any one population we can conclude that it is due to an interaction with other, unidentified genes that form part of a gene network that regulate the expression of the trait. Disappointing perhaps, but we know that the genome is a complicated place. On the other hand, when a test is based on linked markers, this means that the test is examining a mutation that does not directly cause the phenotypic effect but is predicting the true effect due to the fact that it lies somewhat close to the causal mutation on the chromosome. When this is the case, there are several other reasons why the test may not work in all populations.

Consequently, we view the process of validation as being more than simply asking whether we find an effect due to a DNA marker on a trait. If an effect is discovered, we further ask the question "Does it look like the test is detecting a causal mutation, or is it simply a linked marker?" To accomplish this, we need to score several additional markers on a chromosome that flank the DNA marker undergoing validation. This way, we can obtain information concerning how far the tested marker lies from the detected true effect. If the answer is "zero distance" we can at least feel comfortable that the marker may be the causal mutation. But if we find distance > 0 we know that the marker is a linked marker and that this means it may not work in all populations. Further we know that we should continue to look for the true causal marker because this will be of greater value to breeders than linked markers, once discovered.

Here are a list of genes for which commercial tests are either available or may become available that we have examined in our test populations:

• Thryoglobulin
• Growth Hormone Receptor
• DGAT1
• Leptin