USING RELATIVE RISK ANALYSIS AND OPEN HEALTH REGISTRIES TO PLAN MATINGS
This article is important for breeders who are planning a breeding and would like to compute the odds of the proposed mating producing CA carriers. Until we have a genetic test for carriers of CA, relative risk analysis is our only tool for assessing the likelihood of producing dogs affected with CA. Dr. Bell can also assist you in the relative risk pedigree analysis. (Please see "Requesting a Risk Analysis".)
Important points covered:
1. With a simple autosomal recessive genetic disorder, pedigree analysis can be used to compute relative risk factors for carrier and affected status.
2. The objective risk factor for a parent of an affected dog is 100% chance of being a carrier. A full sibling of a carrier has a 50% chance of being a carrier. A non-affected offspring of a carrier with no affected littermates has a 50% chance of being a carrier. A non-affected full sibling to an affected dog has a 67% chance of being a carrier.
Notice that these percentages refer to carriers, not affected dogs. The chance of producing an affected dog would be even less.
3. It is erroneous to add risk factors for each parent of proven carriers. Only one parent of a carrier needs to pass on the defective gene and we cannot determine which parent is a carrier. Assigning carrier risk to both parents of a carrier also incorrectly places selective pressure against dogs based on their matings, and not their proven genetic background.
AUTOSOMAL RECESSIVE INHERITANCE
This article will be helpful in understanding how CA is inherited. Breeders especially are urged to study it carefully.
Autosomal = a pair of like chromosomes.
Recessive = two copies of a gene must be present before a dog is affected by the disease or trait, thus a carrier would have one copy of the gene to pass on to offspring but would not actually have the disease or trait.
Important points covered:
1. Simple autosomal recessive genetic conditions are inherited only from parents that are carriers for the recessive gene or are themselves affected by the condition.
2. Both affected and carrier parents can pass the recessive gene on to their offspring.
3. Both parents of an affected animal must be carriers of the recessive gene.
4. Breeding carriers will not always produce an affected offspring.
5. Clear bred to Clear will only produce Clear offspring
CA – DR. BELL & OESCA
These articles are helpful to get an overview of the problems we face, how we began, and how far we have come in working together.
Some of the important points:
1. CA is not a single kennel problem
2. CA is an autosomal recessive gene
3. The closest common ancestor in pedigrees is not necessarily a carrier
4. To keep our gene pool diverse, we should not discard dogs from breeding programs simply because they are related to a dog affected with CA.
CEREBELLAR ABIOTROPHY (CA) SYMPTOMS
This short article gives a description of the typical gait of a dog affected with CA.
CEREBELLAR ABIOTROPHY IN THE OLD ENGLISH SHEEPDOG: Its Cause And Diagnosis
This article is a great resource for understanding CA in our breed. A must read before you view the CA registry.
Important points covered are:
1. CA is caused by a degeneration of cells in the cerebellum of the brain.
2. CA causes uncoordinated movements of the limbs.
3. CA is not painful and need not shorten a dog's life.
4. CA can be diagnosed through clinical observation, examination of the cerebellum after death, and an MRI in advanced cases.
5. CA is caused by an autosomal recessive gene, which means both parents must carry the gene to produce an affected dog.
Cerebellar Ataxia Update
Progress report for OES Club of America, January 2009
from Dr. Natasha Olby
The current grant we are working on was funded in April of 2008, and since that time we have done a lot of work both focusing on the region of linkage we had established in our prior work, and screening the rest of the genome more completely.
In our previous work we had identified a region on one chromosome that had significant LOD scores. In other words, this region was linked to the disease based on our linkage analysis. However, the region was large and contained literally hundreds of genes. Our work since April 2008 took up the search from this point. To see if we could narrow the region, we identified new markers within the region and genotyped them in our families of dogs. Unfortunately, this did not make the region any smaller, probably because of the relatively low number of dogs we are working with. For our next step we looked for candidate genes in the region that was linked to the disease. These are genes that are either known to cause similar diseases in other species, or to have a function that might affect neuronal survival. We sequenced four of these genes in a small group of dogs and very disappointingly, we did not find any consistent differences between the normal and affected dogs.
Our next step was to try to look at the whole genome again using SNP chips. This very new technology allows us to look at markers that are relatively closely spaced over the whole genome (just over 22,000 markers in total). We were very lucky to collaborate with a laboratory at NIH for this – their expertise, equipment and financial support made it possible. Using this data, we then performed a different type of analysis, an association study, to see if we could find a region that was associated with the disease. We did this analysis in 22 dogs and unfortunately, the association study did not provide any significant values for us either in the region we had already identified as linked, or in the rest of the genome.
There are numerous possible reasons for this: first and foremost is the number of dogs we are working with. We have DNA from only 16 affected dogs and most association studies are run with at least 30 affected dogs, most use more than 100 affected individuals. There is not much we can do about this other than wait for more affected dogs to appear, and the flip side of this problem is that it does indicate that the prevalence of the disease is not extremely high within the breed, which is a good thing. Other possible reasons for failing to identify a region are that we have the incorrect phenotype (are the affected dogs really affected and are the normal dogs really normal?), incorrect dog ID or a more complex mode of inheritance than anticipated. We continue to be as careful as possible in deciding the phenotype of each dog and considering different modes of inheritance in our analyses.
What are we doing now? Rather than be disheartened, we are moving forward in a couple of ways. Firstly, we have statisticians evaluating the data in different ways to see if they can find any statistical link to the disease. This takes time – with over 22,000 data points for each animal, the data files are enormous. If they show up anything interesting we will be looking to sequence candidate genes in those regions. Secondly, we are planning to genotype another 48 dogs on the SNP chips provided we have sufficient funds. Our hope will be that including more dogs will give us the statistical power we need to identify a linked region. This all takes time, so please be patient and be assured we are working hard on the project. We are limited by the low numbers of samples we have so if you know anyone with an affected dog, please ask them to contact us (919 513 7235). If we have any good news, we will of course pass it on to the OES Club of American Health Committee. In the mean time, we will continue our work and hope that an answer will be forthcoming soon. Thank you and the AKC Canine Health Foundation for your continued support of our work.
Natasha Olby