The Human Genome project is widely considered the most revolutionary scientific achievement of modern times, and Dr. Leroy Hood was one of its first advocates and a key player in its success. He continues to lead the way in challenging traditional methods of scientific research, and he is now championing the new discipline of systems biology, which treats biology as an informational science and attempts to solve fundamental problems in biology and medicine by taking more of a big picture approach.
Dr. Hood started down this path when he created the cross-disciplinary Department of Molecular Biotechnology at the University of Washington, bringing together chemists, engineers, computer scientists, applied physicists, and biologists. He has recently expanded these efforts and moved them outside the walls of academia by creating the unique and independent Institute for Systems Biology in Seattle, Washington. We took him away from his busy schedule to discuss his work in systems biology, bioinformatics, and the ethical questions raised by the advance of predictive and preventative medicine.
Stewart: What is systems biology?
Hood: Systems biology is basically the ability to study complex biological systems looking at all genes or all proteins, in terms of perturbations and model organisms. It's studying systems by looking at all the elements in the system rather than looking at things one at a time. That's a simple explanation.
Stewart: What are some of the promising areas of application for systems biology?
Hood: It's absolutely global. That is any biological system that you're interested in can really be approached with these technologies. So at the Institute we're studying microbial organisms and how their systems for energy metabolism and things like that work. We're studying the immune response and how, for example, one can come to understand how different types of immune cells talk to one another so you can make them more efficiently generate vaccines. We're interested in taking a systems approach toward cancer, and we've studied prostate cancer for the last six or seven years using this approach and I think it's been remarkably revealing.
Dr. Hood will be delivering the closing keynote, on Integrative Systems Biology: Genomics, Proteomics, and Computation, at O'Reilly's upcoming Bioinformatics Technology Conference,.
We're interested in stem cell developments and again, using a systems approach to understand the multiplicity of choices that a stem cell makes in going down its pathway to differentiation. We're studying stem cells in the bone marrow, hematopoietic stem cells. We're also interested in applying systems approaches toward looking at issues of genetic predisposition to disease, and this will move us toward more of a predictive kind of medicine. And then ultimately, to take genes that predispose to disease, put them in the context of the systems in which they operate using the systems approaches, and learn how to circumvent whatever limitations they impose.
This is moving to the final phase, which is preventive medicine. That is you can actually design ways that will prevent people from getting the predispositions to which their genes have made them susceptible, if you can figure out how to circumvent whatever limitations those genes are causing. So there is a broad spectrum of different kinds of interest that we have.
Stewart: How is systems biology changing the face of biological research?
Hood: The revolution that systems biology has caused is really an interesting one because whereas when molecular biology arrived in the late '60s and early '70s, it was a paradigm change in biology that fit in nicely with both academic structures and funding structures. It was a discipline based on individual investigators and small laboratories, and it was all based on looking at genes and proteins one at a time.
The systems approach is really a striking contrast with that because if you're to take these more global analyses that systems approaches require it means two things: one, that you have to have the tools for looking globally at RNA and how it changes when you do perturbations, or protein and how it changes when you do perturbations, or complex cellular responses and how they change. So it means you have to have lots of technologies for sequencing DNA arrays, genotyping various aspects of proteomics, multiparameter cell sorting, single-cell assays, all of these kinds of things. It means that systems biology requires an infrastructure that is significantly greatly than what molecular biology, or biochemistry, or cell biology requires.
The second thing it requires, which is very unusual, is sophisticated computational technologies. Because if you're to acquire these global data sets, you have to have the tools for being able to capture and store them, and ultimately to display them, analyze them, and mathematically model them.
|Tell us what you think about systems biology and bioinformatics.|
Stewart: How does bioinformatics fit into systems biology?
Hood: The integration of bioinformatics with these systems approaches is an integral, essential feature. One of the things that we stress is that in the future it's going to be increasingly important for people in bioinformatics to be intimately associated with data producers, because no matter how smart you are you can't model biological complexity--it's just too complex. The only way we're going to understand it is through the integration of these global experimental observations, together with powerful computational tools for analysis, and ultimately, for modeling.
A mistake that a lot of people in bioinformatics have tended to make is thinking that you can set up a bioinformatics center and it can work in isolation from the biology, and it can study all these great databases and learn lots and lots about biology. In vitro biology and in silico biology are all popular terms, but it isn't true, and it isn't going to be true in the future.
This whole iterative process of systems biology means you do a global set of experiments, you model it, you find that theory and experiment don't quite jive so you have to do more experiments, and you have to change your model and it goes around and around in a continuous iterative kind of process. The people in computational biology that aren't very closely allied with data producers and those people who don't have the capacity to integrate experiments with their events are going to be left behind in the world of systems biology. Now, that doesn't mean you can't search genomes and find interesting things, or you can't search proteoms and find them, but if you want to understand the systems and how they work you're going to have to integrate biology with computation.
Hood (continued): I think those relationships are absolutely critical, as I indicated before, and they're critical in a number of ways. First, it means people that come to biology from computation, or theoretical physics, or electrical engineering, must understand biology in a deep sense, because I think their ability to contribute is going to be directly proportional to the sophistication of their understanding of biology. Second, they have to be in an environment where they can directly feed their their computational insights back into the kind of biology that's being done in this iterative kind of cycle. That is, the computational people, as well as the experimental people, must communicate and talk with one another in formulating the next round of iterative experiments so they can juxtapose theory with experiment, and move from more descriptive aspects of biological systems to their graphical formulation and ultimately, their mathematical modeling, and so forth.
I would emphasize as strongly as I could the intimate interrelationships that are really going to determine the success of those groups that are practicing systems biology. That is, those who do it only with the biologists or those who attempt to do it only with the people in either computational biology or bioinformatics, will fail.
Stewart: What are the biggest challenges you've faced in applying a systems-based approach to biological research?
Hood: The biggest challenge was that it didn't fit into the classic academic infrastructure for doing science. We found that bringing together the true, cross-disciplinary scientists was rendered difficult by the fact that our academic center, and most academic centers, live in a world of departments. And the departments tend to create barriers both in how their students are educated and what the expectations are for faculty.
We found that there were real limitations in the resources we needed to raise to create both our technical high-throughput facilities and our computational facilities. We also found there are real difficulties in salary scales. That is, you need software engineers and indeed other kinds of engineers, and you need to be to be able to compete with the best, at least on a par with industry, and in academia this is obviously extremely difficult to do.
And then there are simple things, like systems biology is a teamwork-type of process, and that runs into tenure where the expectation is that when you're young and most creative you do really safe things all by yourself. That doesn't fit nicely into the teamwork that's necessary for systems biology.
So I think the biggest challenge we met, which took me three and a half years to realize, was that it wasn't going to work in the classic academic infrastructure and that we had to strike out and do this thing independently. We did so about a year and a half ago and it's been strikingly successful. We've been just absolutely delighted. I wish I'd realized this years earlier and not wasted a considerable amount of effort trying to basically fit a square peg into a round hole.
Stewart: You've spoken quite a bit to the motivation for starting up the independent Institute for Systems Biology, and I understand it has a unique organizational philosophy, and it is really prospering at a time when many of the venerable research institutes are suffering. How is the ISB model different from the academic and biotech pharmaceutical models of research?
Hood: It's different in many dimensions. One, we have a true cross-disciplinary faculty. We've got eight faculty members and among them they represent physics, computer science, chemistry, and engineering, as well as biology. Two, we make an enormous effort to have this faculty be interactive and to educate ourselves with regard to the deep problems in biology. Three, we have put in place and are integrating together beautifully all of these high-throughput facilities that capture information at the DNA level, and at the RNA level, at the protein level, and at the interaction level, and cellular levels, and so forth. Very few academic institutions have any means whatsoever for this kind of sophisticated, broad integration of technology.
Fourth, we've made a lot of key industrial partnerships, for three purposes. One, in some cases we take on very long-term problems and companies that have substantial resources are willing to go along with us and help us take on these long-term problems. Two, we have collaborations in at least one case where we, together with a company, are creating a very high throughput platform for proteomics. This is going to cost a substantial amount, so sharing between ISB and the company to do this very challenging thing is working out extremely well. Then finally, and maybe most importantly, we've got a lot of collaborations with companies that have leading-edge technologies that we aren't working on ourselves. ISB is developing a lot of new technology, but we can't develop everything. So we see ourselves as an integrator of technologies by making partnerships with small companies, and bringing their leading-edge technologies into these integrative platforms we have. In doing so, we are giving them biological reality and a type of benchmarking they just can't get any other place. I think this ability to integrate all of these technologies together, and to bring them in both from industry and academia, is one of the unique aspects of the Institute.
The real constraint in academia with these kind of industrial partnerships is the challenges they face in dealing with intellectual property, and the fact that in general, there are not very good people that are dealing with this at universities. So, it takes a long time, and never, ever gets done very well. We certainly ran into that in spades as well.
|I think society doesn't realize that science education is the real basis for inquiry-based thinking, and inquiry-based thinking, I would argue, is equivalent to the three R's.|
If you look at it overall, the Institute also is really interested in how this new systems biology is going to fundamentally change education in biology. We think it's going to push it toward a view of biology as an informational science. In fact, I'm, with others, writing a textbook on this area, and this is something the Institute is going to push. We're also cognizant of a strong need to bring science to society, so we're interested in Third World medicine. We're interested in ethical questions of modern genetics. We're interested in how intellectual property needs to be changed in biology--in the context of these very new views of biology. We have major programs on kindergarten through twelfth-grade science education and I think we've been a real pioneer and leader in changing the educational system in Seattle.
Stewart: I'd love to talk just a little bit more about that. You've been active in education your whole career, including writing many textbooks, and you've been very involved in kindergarten through twelfth-grade programs. How do you feel we're doing as a society at teaching science, both at the university level and in the lower grades, and what changes would you like to see specifically?
Hood: I think in general we aren't teaching science well at all, particularly in the kindergarten through twelfth-grade arena, and the reason for that is that society doesn't understand how difficult it is to teach science. The teachers themselves, particularly at the lower levels, aren't very well educated in science. For example, it's estimated that 2 percent to 3 percent of the elementary teachers have not had much, if any, science background at all. So how can somebody who's never had any background in science teach science to kids?
I think society doesn't realize that science education is the real basis for inquiry-based thinking, and inquiry-based thinking, I would argue, is equivalent to the three R's. As these kids move out into a world of communication and information it's going to be critical that they can think analytically and position themselves for reasonable opportunities and options in the future. The simple fact is, and it's been documented 50 different times, that we're failing in that endeavor. I think we're basically failing because we don't understand how to teach science, which I think you need to teach by hands-on, inquiry-based approaches.
At the college level, in general, the teaching is better. At least you can argue people understand the topics much better. But I still think much of the science teaching at the college level fails for similar reasons. The real essence of teaching is really teaching inquiry-based thinking. It isn't didactic lectures. It isn't sitting up and giving them a thousand facts. It's not having them recite back in rote memory form the formulas for all the amino acids. Rather, it's getting students engaged in an educational process, where they actively think and query and analyze what it is you're teaching them. There are some colleges that can do that pretty well, but most places don't.
In terms of biology, I hope that when we bring out this new textbook about biology as an informational science it will be a real lead in moving the teaching of biology, frankly at all of these levels, away from biology as a classification discipline--where you've got 5,000 words that have been defined and used once or twice throughout the entire text--to very much more of a conceptual, analytic, inquiry-based kind of teaching. That's really what we propose to do. The text we're going to write is going to be for the upper undergraduate level, and for cross-disciplinary scientists initially, but we envision doing something for kindergarten through twelfth-grade science later on, too.
Stewart: Is this the same thing as discovery science and how that's different from traditional hypothesis-based science?
Hood: Yes. Discovery science is basically all about defining the elements in biological objects. The Human Genome Project is a classic example of discovery science because what it's really about is sequencing all 3 billion base-pairs in the 24 strings that constitute the human genome. Once you've done that discovery project, then you obviously open up a multitude of opportunities for hypothesis-driven science. It's why the genome project has revolutionized biology in virtually every way.
Another example of discovery science is analyzing the quantitative expression of all the messenger RNAs present in a particular cell type, or all the proteins present in a particular cell type. Again, the discovery is just the delineation of those elements in a quantitative sense; it doesn't say anything about science or biology. But once you have those things, then you can pose and formulate and integrate together the hypothesis-driven science with the discovery science. So systems biology actually mandates the integration of these two very different approaches to science, discovery science with hypothesis-driven science.
Stewart: You brought up the Human Genome Project, which I know you were involved with from the beginning. What do you see as the paradigm shifts that this project has unleashed on us?
Hood: I think the genome project has changed the practice of biology in medicine in a lot of different ways. The two fundamental paradigm changes I see coming out of it are first, this whole approach of systems biology, and second, in medicine it's going to make possible first predictive and then preventive medicine. So in 20 or 25 years we'll be able to take a snippet of your genome and analyze it and give you a probabilistic health history for the future. Then we'll be able to give you preventive medicines that circumvent the limitations of your genes. So we can keep people healthy, effective, and functional for not just 50 or 60 years, but 70 or 80 or 90 years. And, you know, that gets into very interesting social and ethical questions too.
Stewart: That's leads right into my next question. You've spoken about medicine evolving from a reactive science to a predictive and finally to a preventative science. As this is happening a whole host of new ethical issues arise. What do you consider the most important social and ethical questions our society will be faced with during this transition?
Hood: You know I don't think you can say "most important." I think there are a whole a series of important ones. If we do get to preventive medicine, you know we're going to have productive, creative individuals in their seventies and eighties and nineties, and society's going to have to restructure its attitudes toward older individuals to take advantage of their potential creativity and productivity. That's going to be a major change, just in retirement and all those kinds of things.
A second thing is we're really in the midst of dealing with issues of genetic privacy. If we can analyze your genome, then who has access to it? Insurance companies? Employers? Family? How do we deal with those issues? I think that's a tough one, and it's one that is terrifying in how badly it potentially could be done.
A third area that is somewhat related to things that are happening is this whole misunderstanding of stem cell research and stem cell biology. I think stem cells are one of the greatest potential preventive therapeutic agents in the world of medicine, and it's just ridiculous that the religious right has put these pathetic constraints on what we can do in sorting out how to deal with the human condition, and treat people, and everything. The whole confluence of the religious with science--and you see that in terms of creationism too--is something that America as a society, as an educated society, has responded to worse than any other educated society, by far. And it is due to a small minority of very vocal people, unfortunately.
Questions of germ line genetic engineering and whether humans should take a hand in their own evolution and direct the specification of intelligence, or of physical ability, or memory, or all of those kinds of things, are interesting questions. I think from opportunity always comes challenge, and what you have to do is balance the challenge and the opportunity. I think there are very rational ways for doing that, that don't inhibit and prevent, but our society can be pretty reactionary at times, and not deal well with these kinds of subjects.
Stewart: Do you see any computational research projects on the horizon that will have as dramatic ramifications as the Human Genome Project has had?
Hood: Not computational alone, but computational plus biology, yes. One of the things we're really interested in is can we use the systems approach and the computational approaches to understand how the innate and the adaptive immune system interacts to create vaccines? You know the fact is, in the 100 years since we started making vaccines no change has been made in how we do it, and that's because the systems interacting with one another are so complicated. Studies of one gene or one protein at a time were never ever revealing. With the systems approach we can fundamentally change this. Suppose that we can create at-will vaccines that are effective for emergent infectious diseases, and AIDS, and all of these kinds of things. I think that will have an enormous impact, in fact, in terms of what it will do for people throughout the world, a far more striking effect than the genome project, at least in the short term. There are things like that that I think are going to be really key.
Stewart: Although the ISB has been in existence for only a short while, a lot of exciting science has been produced there. What achievements at ISB are you most proud of?
Hood: One is how we took a simple model system in yeast and demonstrated that this whole systems approach, and the integration of these different levels of biological information, DNA to RNA to protein to interactions, is going to work absolutely beautifully. This was kind of the proof of principle up to that time, which had been largely lacking.
I think a second thing is how Ruedi Aebersold has really pushed the whole field of proteomics forward in a very dramatic fashion, developing two or three major new kinds of technologies for proteomics that will have an enormously broad impact. Also the work that Alan Aderem has done on the innate immune system in defining the nature of its receptors, things called the toll receptors, and how they communicate with one another, and beginning to define the nature of their signal transduction pathways so we can manipulate innate immunity. I think all of this is going to have just an enormous impact on, for example, how we deal with infectious diseases, and so forth. Finally, I would say we've developed a lot of very powerful computational methods for beginning to integrate the complexities that are required of systems biology, which I think are going to be tools generally effective for everyone. Those are four areas where I think we've made an enormous impact.
Stewart: You've been a prominent researcher in molecular biology and biotechnology for many years now, and you've witnessed firsthand the evolution of bioinformatics as a discipline. What do you see as the major or grand challenge facing bioinformatics? Are new approaches needed for analysis?
Hood: I think the grand challenge is how you effectively integrate bioinformatics with biology. I think that is going to be by far the hardest thing. I think in terms of new algorithms, in terms of databases that can integrate heterogeneous kinds of information, in terms of tools that can graphically display complexity, and ultimately model it, and so forth, I think we really can see how to do all of these kinds of things. With the enormous increase in hardware capacity I don't think there are going to be any technical limits on what we can do. So I think the grand challenge is the successful integration of computational biology, or bioinformatics, with biology itself.
Stewart: What kind of training is necessary for people entering the field? What kind of research do you think will be most critical in the next few years?
Hood: Well, I think it's great to have people come in from computer science and theoretical physics with these very strong computational tools, but the training that is really critical is they have to learn biology in a deep sense--what it's all about, and understand what relevant applications of computational biology can be pointed toward biology. For example, probably not this next summer, but by the following summer, we will be running summer courses on biology as an informational science where cross disciplinary scientists can come here and in two or three weeks get a grand introduction, not only didactically to biology as an informational science, but experimentally to what does it take to do biology? It's very important to give computational people a feel for experimental aspects of biology. We have all of our computational people do experimental work for a month or six weeks just so they gain some appreciation of the other side of the world.
Stewart: Do you feel the funding agencies are responding adequately to bioinformatics needs?
Hood: I think bioinformatics is kind of a gimmicky word right now. I think in general there's been a pretty good response in terms of the dollars that are out there. Whether they're being spent wisely or not is another question, but I think both at NIH and NSF, and even DARPA and DOE as well, there are dollars for bringing computational biology into biology.
Stewart: What is your opinion of the work being done on the Cardiome and Physiome projects? Do you think these attempts to create sophisticated computer models of entire organs are moving biological science in the right direction?
Hood: I think the ultimate question that you have to ask when you model biological complexity at the higher level of the whole organism, or even whole organ level, is whether you ultimately can tie those models back to the most fundamental elements, the genes, and the proteins, and things like that. I think as long as there's a gap between the whole organ modeling and the specifics of the molecular and transcriptional networks then you're not going to be able to obtain a really deep understanding. But I think it's very worthwhile to attempt to do these things, and then, very pressingly, to attempt to bridge the gap between downstream phenotypic descriptions and the detailed molecular descriptions.
Stewart: A lot of what we're talking about here with systems biology seems to boil down to this one question: Is it possible to produce a complete mathematical description of a complex biological system? Is that a yes or no question?
Hood: You know it isn't really a yes or no question. I think part of the reason is we'll be able to generate pretty effective mathematical descriptions of many of the complex biological systems in higher organisms. Where I'm less certain about in terms of how effectively we can do that is where the environment and stochastic processes play a major role in defining the informational media, and the brain is such an organ. Whether these more reductionistic methodologies can take you to a more complete, precise definition of how the brain works, and consciousness, and memory, and all those kinds of things, I don't really know and I don't think we'll probably know for awhile. But I do think we will be able to have enormously deep and sophisticated modeling insights and an understanding of the immune response, for example.
Stewart: Thank you so much for your time Dr. Hood, it's been fascinating. I look forward to your keynote at the upcoming Bioinformatics Technology Conference.
Hood: You're welcome.
Bruce Stewart is a freelance technology writer and editor.
Copyright © 2009 O'Reilly Media, Inc.