Genetics: Stephen Fodor

By Jeff Minton, Deborah Franklin|Thursday, November 25, 2004


biochemist, Affymetrix

He grew up in Seattle, where he enjoyed blowing things up in the backyard, and worked on a potato farm before deciding to go to college.

On success in life and in science:

“The main part is to recognize when you do something that opens up into new areas and keep going in that direction. It doesn’t matter what part of life it is. . . . Things are not so far down the road in science that you can’t make impacts, and yet I don’t think science is really taught in that way. It’s taught in a more constraining way, so that you wind up thinking about how you’re going to fit into the existing structure that everyone has built for you. It’s not taught from the perspective of how you could create something new, so you should just go for it.”


DANIEL BRANTON, professor of biology, Harvard University

DAVID COX, chief science officer, Perlegen Sciences


MICHAEL STRATTON, chairman of the Cancer Genome Project, Sanger Institute



In the late 1980s, long before the human genome had been sequenced, Stephen Fodor and a few scientist pals in Silicon Valley dreamed of putting the entire human genome on a single glass test chip the size of a thumbnail. Late last year, Affymetrix—the company that Fodor founded and heads as CEO—announced the creation of the whole-genome chip. Researchers and doctors can use it to scan the entire library of human genes and, in a single semiautomated experiment, zoom out and spot the interactive activity of many far-flung bits of DNA in a particular type of cancer, for example, then track the tumor’s response to treatment. Last spring the company created a slightly different tool that promises to be at least as transformative: two tiny glass grids, each containing 50,000 SNPs (pronounced “snips”), or single nucleotide polymorphisms, which can be thought of as all the slightly different spellings of many genetic “words.” Because geneticists use SNPs as markers throughout the genome, this new tool can tease out the link from a particular pattern of widely spaced genes to such complex diseases as autism, diabetes, and cancer. Fodor’s big idea—to wed crisp computer-chip technology to gooey biology—started it all.

You developed a technique that never existed before. Tell me a little bit about that process.

F: The idea behind Affymetrix was to try to build chemical diversity in different ways. We were exploring different techniques, and one was how to build arrays. Arrays had been around since at least the mid-1960s. What we did was develop methods to shrink everything down to a really, really small size. We went very quickly from concept to demonstrating different aspects of it within a couple of months.

Combining techniques are extremely efficient. Say you take pieces of DNA that are 25 units long. It turns out that with A,C,T, and G there are 425 different possible combinations of these nucleotides. That’s about 1015 combinations. The human genome is only about 109 nucleotides. Combinatorially, you can make this entire set in 100 steps. With standard chemistry, if you asked how many 25 nucleotide combinations can I make with 100 chemical steps, the answer would be four.

You make it sound easy, but to make these techniques work, you have to have good hands, right?

F: It’s like anything else. Look at cooks and quiltmakers and artists—I mean, there are a lot of really skilled people out there. They just focus on something, and they make a contribution. Science is the same way.

Are there good reasons to continue to go smaller and smaller?

F: Sure. Two things happen. First, it gets cheaper, and  you also get more information to look at.

Will this be useful in doctors’ offices for diagnostic tests?

F: Once we have a way of looking at the human genome, we should be able to screen multiple copies of the genome and find common polymorphisms [gene variants] and then put those polymorphisms on chips. So we did that with about 50 human genomes and found the spots that vary through a complex set of people. So we think of it very much as, “How do you take a snapshot of the human genome and represent it on these arrays?”

Where are you now in terms of correlating this information with various diseases?

F: That’s one of the really cool things about this work—there is this tremendous scientific fun. But the second aspect is that as people actually start to use the technology, we really start to touch humanity in new ways. The SIDS example with the Amish and the Mennonites is a wonderful example. Because of the relatively closed nature of their communities, they are not very outbred populations, and that has led to a higher-than-average incidence of a form of sudden infant death syndrome. We supplied equipment for the Translational Genomics Research Center in Phoenix and for Holmes Morton of the Clinic for Special Children in Strasburg, Pennsylvania, who collected genetic samples and within a very short time isolated the genes responsible. Modern genetics will be extremely powerful for telling us a lot about health and disease—that will be the main push obviously, but at the same time it will start to tell us more and more about ourselves, about the migration of man, and about our individual quirks.

Do you worry about the ethical implications?

F: You have to. One issue is, do you really want to know about things that you can’t do anything about? The second is, what happens if people start to use genetic information to discriminate against others in terms of health insurance, jobs, and so forth? We simply must have genetic nondiscrimination laws. You should not be able to discriminate on the basis of genetics, because everyone will have different genetic predilections—some combination of all the things that have been generated through mutations for millions of years.

But my genetic information also tells you something about my brother or my parents or my grandparents, right?

F: That’s a problem even today, without genotyping. Insurance companies make assumptions that there are diseases that run in families. But personally, I don’t think these rates can or should be calculated based on genetics—mainly because beyond the question of whether that’s a bona fide use, the knowledge base is not good enough. On the other hand, I think there is a real obligation to do this sort of research. Ethics isn’t just about what not to do; it’s also about what to do. A lot of really, really good things are going to come out of this work—findings like those turning up in the SIDS work, like diabetes, Alzheimer’s disease, and all sorts of things, you name it. We’re going to get a good handle on what some of the genetic markers are and then next, hopefully, what some of the underlying biochemical mechanisms are that will allow us to treat these things.

Comment on this article

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar

Log in to your account

Email address:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »