Name: Leroy Hood
Title: President and co-founder, Institute for Systems Biology, Seattle, since 1999
Experience and Education:Professor, Chairman, and Founder, Department of Molecular Biotechnology, University of Washington, 1992-1999
Professor of Biology, California Institute of Technology, 1970-1992
Senior Investigator, National Cancer Institute, 1967-70
PhD, Biochemistry, California Institute of Technology, 1968
MD, Johns Hopkins School of Medicine, 1964
BS, Biology, California Institute of Technology, 1960
BS, Biology, California Institute of Technology, 1960
Lee Hood has a long history in genome sequencing: as a faculty member at Caltech in the 1980s, he and his colleagues invented the automated DNA sequencer — along with the DNA synthesizer, the protein synthesizer, and the protein sequencer. He also co-founded Applied Biosystems, which commercialized the DNA sequencing instrument.
Hood, who won the 2002 Kyoto Prize in Advanced Technology, and the 1987 Lasker Prize for his studies on the mechanism of immune diversity, will be inducted into the National Inventors Hall of Fame this spring.
In Sequence spoke with Hood last week to get his take on the new sequencing technologies.
What biological — or systems biology — questions will the new DNA-sequencing technologies allow us to answer that current technology cannot address?
The new technologies are all oriented toward higher throughput and lower cost. I think the really big questions that we did not answer before are two: One, can we do genomes a lot more inexpensively, [for example] animal genomes. If we can get a $1,000 genome for humans, we could do lots and lots of animal genomes very inexpensively, obviously. That would open up an arena of comparative genomics in a dimension far beyond anything we have done before, where you could think of thousands or tens of thousands of genomes that you could do.
And of course the second area, which is certainly going to be the goal for many of these companies, is to do individual human genome sequences for under $1,000. There you are talking about, in principle, billions of genomes. These are the genomics applications that one can think about. On the other hand, there are transcriptome analyses of animals, of organs, of single cell types, and eventually, of single cells. And being able to analyze transcriptomes dynamically with these new sequencing techniques, and quantitatively and rapidly, is obviously going to be a very exciting dimension of what one can do with genomics.
It goes without saying, if you can do whole genome sequences, you automatically cover all of the analyses that are associated with human genetics and variability. I think the bottom line is, once we develop these new technologies — and I have no doubt that they will emerge from one of the four or five companies that are really pushing this effort — then there are going to be enormous opportunities in developing mathematical and computational tools that will deal effectively with all of this information. And we don’t begin to have the tools to do that effectively today.
Where do you see the greatest challenges for these technologies, or for analyzing the data that comes off of these machines?
There are two challenges: One is the technical challenge of how can sequencing be scaled up so you can really get down to the $1,000 or $500 or a few-hundred-dollar human genome, or the equivalent of that. That will involve important surface chemistries, it will involve very important imaging techniques, it will involve requirements for massive parallelization. It will raise questions on read length, and how you can assemble the information. Those are all technical problems on the sequencing side. And then on the analysis side, it’s ‘How do you compare with one another billions of genomes that have billions of nucleotides?’ If you think about doing all-by-all comparisons, which are the most rigorous kinds of comparisons, we don’t begin to have the tools nor the computational power to do that, so how are we going to do it? Those are the kind of questions that people have to deal with.
How do you think we will sequence DNA 10 years from now?
I think the human genomes will be done quickly, rapidly, relatively automatically, and for well under $1,000. And I think it’s really going be done by developing techniques that A), can massively parallelize the sequencing process and B), are elegant and simple in themselves, so that the automation and sample prep can be streamlined so you can do lots of them simultaneously.
You have had a history of developing new technologies, not least the first automated DNA sequencer. Do you see a parallel between the time that you did that and today when the new sequencing technologies are coming online?
I think there is a parallel in it, when we went from manual sequencing to the first automated sequencer. By the time we had a robust instrument, we had increased the sequencing throughput by three or four orders of magnitude, and I think exactly the same is going to happen with the new techniques. From the capillary sequencers that Applied Biosystems had at the turn of the century, I think we are going to see another three or four orders of magnitude increase in throughput and corresponding decrease in cost.
On the one hand, it’s a very parallel situation; on the other hand, the technical problems are quite different. [Researchers] are pushing down into microfluidic nanotechnology techniques, assembling small samples, small volumes, massive parallelization, things that we didn’t worry about with the initial automation.
According to the ISB’s website, the institute is not currently developing new sequencing technologies, but partnering with academic collaborators and companies. Can you tell me a little more about that?
One of the companies that I am an adviser to is Helicos [BioSciences] in Cambridge. That’s a company that is doing single-stranded DNA sequencing. It’s one of the four or five companies out there that are pushing for next-generation sequencing. I think in the short run, there are clearly other companies — 454, Solexa — that are going to bring out, or have brought out, instruments that excite a lot of people. But the question is, who in the end is going to have the under-$1,000 human genome? And I think that question is very open, and I think probably Helicos is a good candidate.
One, their technique is elegantly simple; single-stranded sequencing requires no PCR, no cloning and vectors, and two, it’s potentially capable of massive parallelism. For those two reasons, I think Helicos is certainly in a strong position. On the other hand, they have got surface chemistry and enzymology and other issues that they have to deal with, but all the companies have those to a greater or lesser extent. I know [Helicos’] technology much better, I know the other technologies in a general sense, [but] I don’t know them in a deep technical sense. I think everybody is struggling with optimization. It’s not going to be a trivial process.
Has the ISB bought any of the new platforms yet?
We are in the process of thinking about which of the instruments to buy. In the short run, the instruments you might get may be very different from the instruments you will get in the long run, but we have not made a decision on that yet.
How will cheap and fast DNA sequencing techniques compete with other techniques, like genotyping or gene-expression arrays?
I think they will completely replace all of them. I think they will totally replace arrays and things like that because, why get bits of information when you can just as easily get all the information? I think it’s the systems approach to thinking about these things — namely, you do global analyses where you can, which means totally comprehensive, which in this case means complete sequence comparison rather than partial information.