Online Extra: A Genome Pioneer Looks Forward

Expectations of instant applications were unrealistic, says Dr. Francis Collins, but an outpouring of discoveries is coming soon

Before he became director of the Human Genome Project and the National Institute for Human Genome Research, Dr. Francis Collins earned a reputation as a leading gene hunter. He and his team at the University of Michigan helped discover the genes that cause cystic fibrosis, neurofibromatosis, and Huntington's disease. But now much of Collins' focus isn't on individual genes but on projects that follow up on the success of the human genome project.

Collins spoke with BusinessWeek Senior Correspondent John Carey in his office at the National Institutes of Health in Bethesda, Md. Edited excerpts of their conversation follow:

Q: Now that the Human Genome Project is over, what comes next?

A:

We put together a vision for the future on the advice of 600 or so of the best and brightest minds we could convince to come to NIH. Out of that document, one can point to a few flagship efforts. I think they're all groundbreaking and exciting and keep me inspired.

The genome project, as incredibly exciting and ambitious as it was, and having delivered on all its promises, was building a foundation. Now it's time to put the rest of the building together.

Q: For instance?

A:

Let's talk about some of these new revolutionary genomic initiatives, beginning with chemical genomics. What we're trying to do there is develop chemical probes for all the products of the genome -- all the proteins in their various forms. It's extremely ambitious.

A few hundred of those [proteins] -- a small fraction of the genome -- have been the backbone of the pharmaceutical industry for decades [i.e. as targets for drugs]. The pharmaceutical industry is expanding its interest in other targets. But they aren't going to be in a position to tackle the vast majority of targets that the genome project has now identified.

Q: So how will academic scientists be able to study all those other targets?

A:

The plan is to put in a network of high-throughput screening centers. In the first facility, we will have the capacity to screen 500,000 small molecules every 24 hours using a robotic system.

Investigators will bring assays [i.e. biological tests] to the center [to be screened against the library of chemicals]. The first version of the library of chemicals is already together, with 100,000 compounds, and it will grow to a half-million compounds.

The goal is to really understand biology. And some small percentage of these compounds that show promise may actually find their way into further development [as potential drugs].

Q: What other initiatives are helping scientists make discoveries?

A:

HapMap has already begun to make a significant difference. The HapMap was used to discover, really surprisingly, a major heredity factor of the most common cause of blindness in the elderly -- macular degeneration. It's one of the early home runs and presages a lot more to come.

Q: So what is the HapMap?

A:

The things that cause the greatest amount of suffering, morbidity, mortality, and economic costs are common variants [of genes], not rare ones. Each of those variants has a pretty modest effect on risk, but finding them is incredibly valuable.

Q: And where does the HapMap fit in?

A:

HapMap comes along to try to catalyze more of these discoveries. These genes [i.e. genes for which there are varying versions that are associated with disease] have been hard to find. People spent years trying to find genes for mental illness, cancer, hypertension, heart disease, or asthma.

Q: But in many cases they didn't succeed, or they found a candidate gene that didn't hold up to scrutiny, right?

A:

Yes. They were using the best approach we had, which was to study families [i.e. looking for families that have higher risk of a disease, then trying to find variations in their genes that might explain the higher risk]. The method works well for cystic fibrosis or Huntington's disease [which are caused by single genes]. But it's lousy for diabetes.

What you need is association. You have cases of a disease, and controls [i.e. people without the disease], and you look for places in the genome where there's a difference in the frequency of a particular [genetic] variant. With 1,000 cases and 1,000 controls for a common disease, you have the power to find variants that might contribute a very modest amount to the disease. But testing for every common variant in the genome -- about 10 million of them -- is a problem.

Q: So you need a way to test for fewer of these variants?

A:

That's why we need HapMap. It's a fantastic shortcut. The reason for the shortcut is that those 10 million common variants aren't really independent of each other. If you test one on chromosome 7 and know it's a T [one of the four "letters" in the genetic code] and not a C, you can go to the next variant and have a very good idea of what it is going to be. [Note: the regions of linked variants are called haplotypes. That's why this effort to map them in the genome called HapMap. Click here for more detail.]

That's what the HapMap is determining. Already you will be able to pick a set of a quarter-million of these variants and basically represent the whole genome. That saves you a factor of four. It takes an absolutely unthinkable project and makes it very thinkable.

With HapMap and genotyping costs coming down, I suspect there will be an outpouring of discoveries of genes associated with disease in the next 12 months.

Q: Tell us also about your idea for a human genome cancer project.

A:

The ability to get the complete sequence of an individual is an explicit goal for the future of medicine. As we move toward that, we would like to pick examples that are likely be most medically rewarding, and cancer immediately comes to mind.

If I've learned anything in the last 15 years working on the genome project, it's that you shouldn't settle for the 20,000-foot view if you could get down and look at every blade of grass.

Our current proposal is to sequence 250 tumors for each of 50 cancers. So we're basically talking about 12,500 genome projects. That seems wildly ambitious. But the way things are going, I think that's achievable if you allow something like six or seven years to get there, and if you push the technology really hard along the way to try to drop the cost.

Q: Of course, we've heard for years that the whole genome effort will lead to all sorts of new drugs and treatments. Now these new projects sound like they are making similar promises. Has all of this gone slower than expected?

A:

Do you think people bought into the idea that there would be immediate medical benefit the day the genome sequences was completed?

Q: If they did, it might be that many of the scientists involved -- as well as the media -- made it seem as if these advances were right around the corner.

A:

Maybe we're being punished for being careless in those original statements and not including the timetable. Anyone who knows anything about the timetable between discovery and application would have said, "Oh, come on."

Q: So the promise of a revolution in medicine is still there?

A:

If things go the way they should, you and I -- if we're still kicking by then -- should be able to have our genomes sequenced for $1,000 or less within 10 years. Then when you need a prescription, we can have a quick check to be sure that you're getting the right drug at the right dose.

By that point, HapMap will have catalyzed an outpouring of discoveries of [genetic] variants that are associated with common disease. So people who want to find out what their individual risk is will have a chance to do so at an increasing level. And some fraction of those discoveries will be associated with potential interventions, be they lifestyle or medical surveillance or diet.

I don't want to overstate the timetable either and make the same mistake we made earlier. But I will hold passionately to the idea that we're on the right track and that this is still the most compelling opportunity we have to change the equation of ourselves vs. disease.

Edited by Patricia O'Connell

    Before it's here, it's on the Bloomberg Terminal.
    LEARN MORE