Last week, the Harvard Medical School-Partners Health Care Center for Genetics and Genomics announced that it will embark on a large-scale genotyping study to find genetic risk factors for atherosclerosis. The study will analyze DNA from 1,200 women using Affymetrix’s GeneChip 500K Human Mapping arrays under an early-access program. The samples come from the Women’s Health Study, an ongoing 10-year NIH-funded study of cardiovascular events with 38,000 women participating.
BioArray News recently spoke with David Kwiatkowski, one of the project’s study leaders, to discuss the new undertaking. Kwiatkowski is the director of genotyping at HPCGG, which is based in Cambridge, Mass., as well as an associate professor of medicine at Harvard Medical School and a physician at Brigham and Women’s Hospital.
How are you going to conduct this study? How many patients and controls do you have?
[In] this initial study of 1,200, these are all controls. In these individuals, we are going to be looking at a wide range of serum markers that correlate, to some extent, with atherosclerotic risk as well as other conditions. Those serum markers include C-reactive protein, but also other inflammatory markers, [for example] interleukin 6, and several different lipid measures, [such as] cholesterol. [In] this control population, we will be looking at sites of [genetic] variation and their influence on all these measures of the biochemical milieu. When that is completed, we plan to do another 400 atherosclerosis cases that will be compared with these controls to look at those same aspects but also things that correlate with risk of atherosclerotic events.
How do you plan to transform your results into a genetic test?
It’s a long path; it doesn’t come quickly. There will need to be application of refined statistical and analytical methods for looking at this data to pick out the SNPs and genes that seem to correlate with these various things we are examining, as well as [disease] events. And in fact, we will need to replicate these in another set, another population, to be confident of the findings. But the final outcome would be to identify genes and variants within them that do influence both these various serum markers and also atherosclerotic risk, and to put that together into what could be a genetic risk profile.
Do you expect to find differences between men and women, since this study only includes women?
It’s an important point, and there is evidence that atherosclerosis in various parts of the body does differ according to sex. This population in certain respects is ideal for this study. They have been followed up [on] for ten years, very carefully followed clinically, so that the epidemiologic basis for this is very strong. We also think that starting with a single sex is probably better to help avoid differences between the sexes in terms of the analysis.
Why did you choose Affymetrix’s platform for this project?
We wanted to be able to look through the genome in an unbiased way at many different sites of variation and genes to try to identify those genes and sites of variation that have a role in influencing both inflammation and risk of atherosclerotic events. The reason why the Affymetrix technology is attractive for this purpose is that it’s much cheaper than any other technology currently available.
How much cheaper?
I think a reasonable estimate would be that it’s at least five-fold lower in terms of costs than other technologies. [For example], it’s more than 10-fold cheaper than Illumina[‘s platform]. Another one is ParAllele [BioScience’s platform], which comes closer in terms of cost, but still, the Affy approach is cheaper.
So cost was the decisive factor?
Cost is extremely important as well as the fact that we are getting very high coverage in terms of the number of SNPs in the genome that are being checked. [That number] will in aggregate be somewhere in the range of 500,000.
Is that number actually high enough?
Well, that is debatable. It’s still not entirely certain. There is work going on to figure that out. This will not give absolutely complete saturating coverage of the genome, that is clear. But we believe, based upon our calculations, that this will get coverage somewhere in the range of 50 percent of the variation in the genome. Therefore, this is an attractive initial experiment to do, to try to identify those regions that are important.
How did you become an early-access user for the 500K arrays, and when will you start using them?
I think the initial contact came from Affy. Basically, [we will start using the 500K arrays] as soon as they are available, which is expected in May or June. In the first phase of the study, we will use the 100K chips.
You mentioned you are also using other genotyping platforms. What do you have at the center?
We have really a broad range [of genotyping technologies.] [ABI] TaqMan genotyping is performed, Sequenom genotyping is done, and we have an Illumina instrument as part of the center, we are also experimenting with SNPlex, which is another ABI approach.
How do you decide when to use which platform?
The choice of method to use is very much dependent on the number of SNPs one wishes to analyze. You would scale among those different [platforms] depending upon what size your project was. They also all have different costs. In general, the costs go down as you go to more and more SNPs being analyzed at once. The TaqMan, for example, is the most expensive, but that is designed for a singleton assay, a single SNP that one would be analyzing.
Do the platforms differ in the quality of the data they produce?
Well, that certainly is possible, and some of these seem to be better than others in that respect. We are still learning about that. Some of them are also somewhat older technologies. I would say that the Illumina is one that is relatively new and looks very attractive in terms of being robust and high quality. The SNPlex we are really just learning about, so it’s too early to comment.
How much funding does the Reynolds Foundation provide for the study?
The total award of the Reynolds Foundation to the Donald W. Reynolds Center at Brigham and Women’s Hospital and Harvard Medical School is $26 million over four years. It’s just a fraction of that overall budget that is devoted to this particular study. Many other things are being looked at with respect to atherosclerosis. This is one project out of four main projects.
When are you hoping to finish the study?
It depends somewhat on the availability of the chips, but the ideal and probably unreachable [date] is fall. More realistic, I’d say, would be by the end of the year.