SEATTLE--Medical genetics professor Maynard Olson, an award-winning geneticist at the University of Washington here, met with BioInform last month to discuss the future of genomics and the role of bioinformatics in the new millennium.
"The bottom line is two things," Olson said. "We really should be thinking about integration of genomics with traditional areas of biology. And we need to be thinking more about social implications."
Since 1992, Olson, 55, has earned numerous honors, including a Genetics Society of America Medal and an American Association for the Advancement of Science fellowship. His work has been published in Science, Genomics, and Genetics. His current research focuses on germ-line genomic sequences and the characterization of natural variation in yeast, bacteria, and, to a limited degree, humans. Natural variation, he said, is "really the key to understanding what the genome means."
Olson trained as a chemist at Cal Tech and Stanford, earning a PhD in physical organic chemistry in 1970, but a sabbatical leave in 1974 inspired him to move into genetics: He visited the University of Washington Genetics Department here and stayed for five years. He left in 1979 to teach at the Washington University School of Medicine in St. Louis, but in 1992 returned to the University of Washington, where he now holds a joint appointment in the departments of medicine and genetics.
BioInform: The field has evolved tremendously since you entered the arena in 1974. What do you consider the genomics highlights of the past 25 years?
Olson: The seventies was the decade of great innovation and experimental methods. It was at its prime for DNA sequencing and mapping. The eighties was mostly refinement of seventies techniques. In the nineties, I think, we finally have a clear picture of what it is that we should have been doing all along, which is to sequence these genomes. And we have some data-handling methods that have really changed the landscape.
BioInform: What are the obstacles facing the genomics field?
Olson: As a quick policy point, I think we need to worry about the incentives of our system. We're seeing large amounts of proprietary data, most of it redundant and at relatively low quality. It's good enough for most purposes, but the idea of getting reference databases of impeccable quality that are publicly available, no questions asked, is a somewhat endangered idea because of the kind of gold-rush mentality that prevails.
BioInform: What do you see as the solution?
Olson: We need a renewed commitment to identifying those reference databases.
BioInform: What about the scientific obstacles?
Olson: Despite all these difficulties, we are getting some good quality reference sequences that are publicly available. What are we going to do with them? A serious scientific obstacle is that we don't really know. The great challenge in bioinformatics is moving beyond that. We need new ideas.
The industry-government-university triangle needs a lot of nurturing. The private sector has become extremely powerful very quickly in this area. And that's basically good, but there is a role for universities and a role for government and we're not paying enough attention to that.
I gave some congressional testimony in June on this subject, that we really need to think clearly about what should be going on in the universities. What should the government be trying to do with tax money? What should be privatized? What should the upstart biotech companies be trying to do? What should the big pharma section do? I think those issues are pretty unstable.
Right now a lot of the action is in the biotechnology industry, not in the pharmaceutical industry, not in the public domain, not in the universities, as far as the hotspots go. We have to worry about how much sense that makes.
BioInform: Where does bioinformatics fit in?
Olson: The field needs to become much more biological. I don't think that means it needs to be turned over to the biologists, but that it is too invested in methods that have little biological input to them. It's not a surprising phenomenon really. Bioinformatics has much stronger roots in information science than it does in biology.
What we need actually are ideas, and ideas ultimately come from biology. The great challenge here is that everyone working in bioinformatics and genome analysis certainly feels they should be learning more biology. If they are studying diabetes they should be spending half their time reading about physiology, biochemistry, and molecular biology. And they also need to be more hypothesis-driven.
BioInform: There are many bioinformatics companies coming into the market. What factors will decide which come out on top?
Olson: There is a big service issue that is driving a lot of the bioinformatics companies. They need a lot of service support for better ways of getting the information from whatever proprietary databases to the user in a form that gets maximum value. That's a worthwhile activity, but I think bioinformatics is over-invested in this activity.
Leadership is going to come from nonbiological areas. There is probably too much effort to design very customized ways of taking genome data and functional genomics data and getting it out. Instead of being the backwater of informatics, bioinformatics should focus more on adding value to the data, real value, not just making it more accessible. That requires getting more involved with biological ideas. Data quality cannot be too strongly emphasized.
BioInform: Where do you see the genomics field in five or 20 years?
Olson: Five years from now the world is going to look a lot like it does now, except there will be more of a lot of things--more data, more chaos. It will be a very chaotic environment, with too many drug targets, not enough ideas on how to sort between them, too many competing user interfaces. So I don't think we've peaked yet in the chaotic transition to data-rich biology.
Twenty years from now things will have settled down greatly and there will be enormous change. Some of the principles will by then have been consolidated. We will have high-quality publicly accessible germ-line sequences and by one way or another we will have developed much cheaper ways of getting them. That activity will be picking up five years from now. It's at a rather low ebb right now.
I think the really big progress on a 20-year timetable is that we will start to understand natural variation in organisms. The leads will be with humans and agricultural plants. We will not only have one good germ-line reference sequence, but we will have 100 or 1,000 fragments. We will really understand, at a level we don't know now, what the variation is really trying to tell us. I actually think that in both drug development and agriculture, natural variation is where the real progress is going to come from.
For the future, the issues are not scientific as much as they are social. Most people working in this field are very thoughtful about social issues and have a slightly uneasy feeling about where all this is headed.
This is immensely powerful new technology, probably more important than nuclear energy in changing the ways of the world. We already have some pretty powerful ways to intervene with biology. The Dolly brouhaha was meaningful in the sense that it was a symbol that we are developing these powerful ways to intervene.
The biotech industry manifests the best and the worst of American society. It's very innovative and entrepreneurial, but it also has an extremely short timeline in its focus and not much social responsibility. All parties need to think more about the socially optimal use of these tools. We need to educate each other a lot more.