|
|
|
Getting Tight with Genes
BNH speaks to J. Claiborne Stephens, director of population genomics at Genaissance Pharmaceuticals in New Haven.
|
Business New Haven
11/15/1999
By: Tammy Rachau
|
Give us an overview of your current research at Genaissance regarding variation in drug targets.
Well, it's ongoing, first of all, and it's not done by any stretch of the imagination. We have a pipeline of genes that we are routinely surveying for genetic variation. This pipeline consist of genes that are pharmaceutically relevant, either in being drug targets or pathways that are relevant to drug metabolism, genes that are implicated in some of the epidemiology and some of the diseases that we're going after. These are the genes that we think are most important to the pharmaceutical industry. [We hope to have done] 2,000 genes by the end of next year.
What have you found to date?
One of the really important findings is a substantial variation in these genes. It's not like everybody has the same version of the gene. We've seen an average of about eight variable sites per gene, which are arranged as what we call haplotypes. Genaissance has a trademark actually called a HAPMarker that we're using.
What is the significance of the variation?
The reason we're accumulating the information about the variations is twofold. First is what you could think of as sort of a direct approach. A molecular biologist who is working on a specific gene would like to know what are the variations that cause different versions of the protein or different expression levels - different amounts of the protein. So that's sort of a basic, direct molecular biological approach. The second one, though, may be even more important. By establishing what the major pattern of variation is for each of these genes, we have what we like to call a molecular sieve if there are rare variants that fall outside of the region that we routinely assay. If we have a really high resolution picture of the variation, by stringing these variable positions together into these HAPMarkers, even if we don't directly look at a site that may cause some sort of an effect, we are able to detect that that particular HAPMarker is associated with the effect. This is what we call a molecular sieve, which is a way of actually doing these rather complex statistical analysis, which are aimed at connecting critical variation, such as drug response or non-response, to the underlying genetic variation. Which is what we're specialists at - looking at that genetic variation.
What specific methods are you using for this research?
We have developed some proprietary software algorithms for making the connection, for doing the statistics, the population genetics - all the stuff that goes into using the data. In terms of the actual wet-lab work, where we establish the variation, we have a very high-tech facility that we call the HAP factory, where we're doing DNA sequencing at about four mega-bases of DNA sequencing a day.
How does your work at Genaissance differ from that of other genomics companies?
We are the only company that is making that connection between gene variation and drug response. There are a lot of genomics companies out there that address gene discovery, but we're not so much gene discovery as we're discovering gene variation. On top of that we know what to do with that variation. We're really focused on connecting that genetic variation to very practical important issues, such as will a person respond to a drug or will they have an adverse reaction? A lot of research is currently focused on mapping the whole genome. We're not just discovery that way; there's a very important, heavy technology information component to the company.
Is the variation randomly distributed, or do you find patterns based on geography or other factors?
We are seeing that the variation is, in fact, distributed. I wouldn't go so far as to say that it's randomly distributed, because some areas probably have a little bit more than others. We tend to see, for instance, amino acid changes being somewhat less prevalent than what are called wobble positions in the genetic code. But we are seeing substantial amounts of variation in all of the regions we're looking at. We're actually kind of surprised by the amount we're seeing. Underlying this is the idea that different variants might react differently with different drugs. The variants that we look at will have different outcomes. We know this from everything that's ever been looked at with genetics. As in all the classical fruit fly genetics, you want to look at something where there's a genetic component to it.
What kind of diseases are you concerned with in this research?
I just came up here from the National Cancer Institute, where I had spent the last nine years. We had a big project there looking for genetic influences underlying infectious diseases, including AIDS. That lab has discovered several genes now that have important relevance to AIDS and AIDS progression. Even things you think of as being not necessarily genetic, such as heart disease - which people think of as having a large environmental component - also have a substantial genetic component. So anything you want to look at, what as geneticists we call 'phenotype,' has a substantial genetic component to it. What we're set up to do is to be positioned to establish what that genetic component is.
How many people's genes do you look at to establish variation?
For every gene we look at approximately 100 individuals, so that we're getting essentially 200 chromosomes a gene. We have a real in-depth feel for what level of variation there is, gene by gene. In individuals, you're looking at four different world populations: equal numbers of Caucasians, African-Americans, Asians and Latino-Hispanics. It's not really what you would call an anthropological sample, but it's relevant to the pharmaceutical industry in terms of what kind of individuals you might see in the pharmaceutical market.
|
Go FirstGo PreviousGo
NextGo LastGo
to Index
|
|
