At A Glance
Name: Robert Williams
Title: Professor of development genetics, University of Tennessee Health Center
Education: PhD in physiology, University of California, Davis; BA in psychobiology, University of California, Santa Cruz
Background: Assistant professor, Yale University
Robert Williams, a professor of developmental genetics at the University of Tennessee Health Center, is mid-way through a 4-year NIH grant to run a genotyping core lab that focuses on the genetics of alcoholism.
Pharmacogenomics Reporter caught up with him to find out where his research stands.
Tell me about the research you’re doing at your genotyping core lab.
We’ve been doing high-density genotyping of the following mice: The BXD panel of strains, which were generated by crossing C57BL/6J (the sequenced strain) and DBA/2J (sequenced by Celera); and the LXS panel of strains, which were generated by crossing ILS/Ibg (inbred long sleep) to ISS/Ibg (inbred short sleep). These strains have been used by the alcohol community for so long as models for human alcoholism. There’s a new set of mice that has just been completed by a group at the University of Colorado, Boulder. The PIs there are Beth Bennett and Tom Johnson. They’ve made another set of recombinant, inbred strains. They’ve made a total of 76 of them, and we’ve also genotyped them; we’re also doing that.
Our genotyping core has been genotyping these permanent immortal mapping populations of recombinant, inbred strains.
In the case of the BXDs, they’ve been used for more than 20 years for alcohol research. In the case of this new set, called LXS, [they] were intentionally produced by selecting mice that either sleep a long time in response to an alcoholic injection, or snap right out of it.
We have two wonderful systems of mouse models that have been built up for over the last 20 years to explore the genetic basis for alcoholism for humans. Now, of course, there are going to be some very important differences between mice and humans, but there are going to be enough similarities at the synaptic level — the basic levels of what’s going on in the brain — that these models will be extremely useful.
Why are you studying the genetics of alcoholism?
Because, I guess there’s sufficient data at this point to suggest an important genetic component for alcoholism. So if we can understand this component, we can understand-in principle, and, we hope, in practice-we’ll be able to devise new interventions, or preventative strategies. To the extent that people are willing to understand their own genomes, they’ll know whether they are vulnerable or not.
Does the core lab have any plans to replicate these studies in humans?
We’re heading, pretty much, in the other direction. Because we have such good control over the mouse environment, we can do experiments that we can’t do in humans. What we can do is find vulnerability alleles in mouse populations, and then we ask, ‘Do these same genes seem to be associated with vulnerability in human populations?’
Define preventative strategies. Would you partner with pharmaceutical companies or other academic centers? What would you do?
I think you’d do all of the above, but I think prevention means that you get there before a pharmaceutical company intervenes (laughs). By prevention, I mean something like a campaign similar to how we try to dissuade kids from smoking cigarettes. I’m not sure to the extent to which people are comfortable to this, but there is a lot of utility in knowing what your genetic vulnerability to certain traits are. There’s certain a lot of nervousness on the part of people because of what they fear insurance companies might do. …
That’s where we’d like to be. We know that the cure for alcoholism is not to drink. The trick is, how do you induce people to control their intake?
How would pharmacogenomics technologies lead to this?
Everybody’s got relative risk for diseases, so if you understand those relative risks, if they could be computed reasonably accurately, if the statistics are relatively firm, then you’re providing people with important knowledge. And the question is, ‘Will people use that information to their own knowledge?’ Obviously, there are many instances where we fail dismally. We know that cigarette smoking is bad, we know that alcohol consumption is bad. So the question is, ‘Will more information help?’
And I think it would. I think if you knew specifically that because of your genetic inheritance you have a certain vulnerability, you might take a little more responsibility for your behavior than if you thought it was all environmental.
Your project is now two years — or roughly one-third — into an NIH grant. How’s the progress been?
Really good. As a core service, it’s not our job primarily to drive the science; we simply provide the core services. A lot of the genotyping involves a mutagenesis project … and we just crank through the genotypes. We also genotype transgenic mice where we have to figure out if an animal is a carrier or not. So that also works very smoothly.
What are some the biggest challenges you’ve encountered.
For a lot of genotyping cores, the issue is, ‘Can we crank through 200,000 genotypes per month. For us, that hasn’t been an issue. The issue has really been, ‘Can we handle your particular samples efficiently, and get you the results in a week or two?’ And usually the answer is yes. But there are, inevitably, start-up problems in finding the right set of markers for a particular project. But [these days] it’s easy to design markers [thanks to available genome sequences].
Has there been any commercial interest?
Yes. We’ve been collaborating — informally — though it has prospects of becoming formal-with Affymetrix. They have a human SNP chip, and they’re very interested in designing an equivalent mouse chip. Their human chip handles 10,000 genotypes on a single array, and they’re coming out with one that handles substantially more. They’ve been talking with us about testing the mouse equivalent, which would have around 10,000 SNPs on a single microarray. The problem here is really the marketing — the needs of human geneticists and mouse geneticists are surprisingly different in terms of the density markers you needs.