Leroy Hood is one of the world’s leading scientists in molecular biotechnology and genomics. He holds numerous patents and awards for his scientific breakthroughs and has a life-long commitment to making science accessible and understandable to the general public. One of his foremost goals is bringing hands-on, inquiry-based science to K-12 classrooms. Hood earned the M.D. (Johns Hopkins, 1964) and the Ph.D. in biochemistry (CalTech, 1968) and has also been given eight honorary degrees and numerous other awards. His research has focused on molecular immunology and biotechnology. His interests also include autoimmune diseases, cancer biology, and mammalian development. He has published more than 500 peer-reviewed papers and co-authored textbooks in biochemistry, immunology, molecular biology and genetics. He also co-edited Code of Codes, a book discussing scientific, social, and ethical issues raised by genetic research. Hood is a member of the National Academy of Sciences, the American Philosophical Society, and the American Association of Arts and Sciences.
Hood’s professional career began at Cal Tech, where he and colleagues pioneered four instruments that constitute the technological foundation for contemporary molecular biology. One of these revolutionized genomics by allowing the rapid automated sequencing of DNA. He was one of the first advocates and a key player in the Human Genome Project. In 1992, he moved to the University of Washington to create the cross-disciplinary Department of Molecular Biotechnology, bringing together chemists, engineers, computer scientists, applied physicists, and biologists. In 1999, he became founding president and director of the Institute for Systems Biology in Seattle. He has also played a role in founding several biotechnology companies, including Amgen, Applied Biosystems, Systemix, Darwin, and Rosetta.
◊ ◊ ◊ ◊ ◊
Ethix: People have usually thought of biology and information science as very different fields. You emphasize their convergence. What is going on here?
The process by which DNA computes have given computer scientists new insights into logic and information processing and handling.
Leroy Hood: There has been a real paradigm shift in the way people view biology, catalyzed in large part by the Human Genome Project. The central thesis of this shift is that biology is really an informational science, all about a complex type of information, some of which is digital in nature. DNA is basically a digital four-letter language, inscribing information on what we call chromosomes. Genes themselves make products like nucleic acid and protein which serve like molecular machines executing the software programs of the chromosomes. Information at the most primitive DNA level is digital in nature, and information at the protein level is basically three-dimensional in nature.
At still higher levels, a series of these molecular machines work together to catalyze a particular kind of informational pathway that tells our body, for example, how to capture energy from sugar or how to trigger some aspect of human development. Within a cell, many interconnected informational pathways form networks. To understand biology then we have to capture information at all these different levels and form an integrated understanding of these systems and how they work.
Despite this convergence, does there remain some irreducible difference between the information systems of living biological organisms and those of non-living things? Is the boundary sharp or fuzzy?
I think it will end up being a very fuzzy boundary. We can now begin to conceptualize how to build cells which carry out all the common reactions of living organisms and how to synthesize them from basic building block components that aren’t themselves alive. Some claim that within a few decades we’ll have computers able to match the human brain in terms of raw computational power.
Could these creations be self-replicating?
Potentially, yes. But even more importantly they could be built with a judgment capacity that rivals and exceeds our own — at least for certain kinds of analytic procedures and questions. I think there will be an enormous convergence over the next fifty years between what living organisms and what inanimate objects and computers can do. Whether we will ever have to worry about these entities taking over the world is a fascinating and complicated question with no clear answer.
Will this convergence and fuzzing of the boundaries threaten our valuing of human life? For example, if I throw out my old computer I don’t worry about it except from a recycling standpoint. Will we some day throw out inefficient or old human beings with the same cavalier attitude if this convergence continues?
I think we will probably move in the reverse direction, respecting computers more and not treating them as stuff to be thrown around! The hierarchical nature of biological information is really fascinating. From the primitive level, the DNA digital program of life, it goes up to higher and higher levels. Each of these higher levels generates systems or emergent properties that are quite remarkable — such as memory, consciousness, and the ability to learn in the brain. We can’t even begin to conceive how those kinds of emergent properties are created. They arise because the brain is an incredibly complicated network of 1012 neurons with 1015 connections.
As computers get more and more complicated emergent properties may arise perhaps displaying a distinction between the capacities to think, to analyze and to learn. Will computers ever feel? I suspect they will at least be able to create or imitate many of the fundamental capacities that we have today. We might respond today by saying “that’s just impossible — we can’t imagine machines being able to that.” But this isn’t necessarily a question of twenty or fifty or two hundred years; the evolution of life took a very long time. All such development is constrained, of course, by what we do to ourselves along the way — how we get along within and between nations. Issues of population growth and what we are doing to the environment create major uncertainties and variables about our future.
So is that another level on the systems hierarchy? One level is the systems within a human being; another is multiple human beings and their interactions with each other?
Yes. The ecological network of interactions of humans with the environment (which we’re in the process of screwing up in a variety of ways) is a still higher level. In the past, those who studied population, ecology, and so forth, were very different from those who studied molecular and cellular kinds of things at the other end of the spectrum. The new systems biology may be the pathway by which there will be a convergence of these two points of view.
Edmund O. Wilson’s Consilience argues an even bolder vision for a grand unification of knowledge, a convergence of the arts and humanities, the social and natural sciences, and the physical and biological sciences. He argues that the convergence point will lie in an understanding of the human brain and a new common language that can bridge the barriers and gaps that exist between these fields.
So information science helps us understand biological systems. Is there a flip side in which biology influences computer design and computers themselves have biological components?
DNA is basically a digital four-letter language, inscribing information on what we call chromosomes. Genes themselves make products like nucleic acid and protein which serve like molecular machines executing the software programs of the chromosomes.
Many of the ways living organisms process and handle information are going to be captured and used in the information technology arena. We have already begun to see this in terms of genetic algorithms, neural nets, and so on. The idea that DNA might be used as a computer is fascinating. The processes by which DNA computes have given computer scientists new insights into logic and information processing and handling. Building living computers is further down stream. I don’t see how to get there from here but I’m certainly open.
Biology also offers wonderful model systems for testing new approaches to computation. One of the major problems in IT today is integrating heterogeneous information in databases. There is no information system better than biology for dealing with this problem. In biology there is a reality against which you can check your algorithms and your approaches in the end. With IT, when you get all the way to the end you don’t know for sure that something works. You haven’t had a reality to test it on that’s quite as effective as this biological reality.
The future of biology is now irrevocably intertwined with computational and mathematical tools. How we think about biology, how we deal with, integrate, and model an explosion of information — these are fundamentally IT problems. Future biologists will need to have a very detailed understanding of IT to do their biology. We don’t think genes or proteins per se, anymore, but about systems interacting with other systems. Our multi-dimensional problems can only be resolved in computational terms.
If a system is large and complex enough though, don’t you get emergent behavior that is unpredictable and can’t always be completely modeled and understood in this way?
I think what we grapple with today is that we can’t yet see the technological limitations.
Is this because our analytic methods are intrinsically incapable of handling the complexity? Or do these emergent properties merely appear impossible to analyze because they are more complicated than anything we’ve yet developed the tools to analyze?
I would argue that over the next ten or twenty years we will be able to define a lot of the biological systems pretty completely through new kinds of modeling and mathematics. These models will allow us to describe the nature of the system and also to predict emergent properties. Some argue that we are never going to be able to deal with the 1015 connections of the brain and the possible emergent properties that arise from that. Maybe that’s true but I have faith that we will be able to understand even the most complex systems by simplifying assumptions and by simplifying models. Eventually, of course, we get to the philosophical question of whether reductionism can give us all of the truth.
As computers get more and more complicated, emergent properties may arise perhaps displaying a distinction between the capacities to think, to analyze and to learn. Will computers ever feel? I suspect they will at least be able to create or imitate many of the fundamental capacities that we have today.
Early in my own research on solving large sparse systems equations, three thousand equations was about the boundary of what we could think about. Then it was fifty thousand and then a hundred thousand and now it is in the millions of equations we can solve. Computing has enabled the data processing to do this; the algorithms are there. Now on smaller systems we didn’t worry about the parameters and their variation. But on large systems we have huge impacts. We can’t say simply “this is the data.” We must say “this is the data with variation.” So it seems to me that complexity and emergent behavior come from the variation of the information — not just from inadequacies in the computation model.
I wouldn’t argue with that at all. People often ask if chaos is actually going to be important in the end for understanding about biology. At the levels we’re thinking about in biology right now, we are orders of magnitude above the phenomena of chaos. Will we get down there? My intuitive feeling is that we probably won’t in biology because the properties that are really interesting lie in molecular interactions way, way above what would be talking about with chaos. What doesn’t go away is the importance of stochastic processes. In biology, cause and effect can be slightly dissociated by probabilities of what’s going to happen, what the collisions and interactions will do, and here the uncertainty is very real. Predictions in this context operate within boundary conditions of uncertainty beyond which we cannot push.
Weather modeling is an example that says this is a harder problem than we thought. We keep pushing and pushing but we’re still not very good at predicting the next day.
I agree, and biology is much harder than weather! I mean 1015 connections in your brain is an absolutely staggering number. But if we have computers in twenty or thirty years that far exceed the complexity of the human brain, we’re going to be able to do things that we can’t even imagine now, in terms of analysis, modeling, and prediction.
How does all of this play out in the marketplace? Does this biological systems understanding merge with the practice of medicine in the future or does it form another field? Does this become the new medicine or is it a rival to medicine?
The systems approach can be applied equally well to basic problems in biology, and to fundamental problems in medicine. It could really transform medicine. Medicine today is in a reactive stage: you get sick, you go see a doctor, and they try to do something for you. Over the next twenty years there will be enormous progress defining in great detail genes that predispose to diseases. We’ll be able to write out your predictive health history, complete with the probabilities that you’ll get, for example, cardiovascular disease by the time you’re sixty-five and so on. Of course, the ability to predict a disease, but then not to be able to cure it, creates enormous ethical, social, and medical issues.
The next stage will be to understand the informational pathways and how to circumvent the limitations of those genes. We need to design drugs that you can take as pills. You can’t convince people to stop smoking or not to eat certain kinds of food, or to stay out of the sun. The key is preventive medicine which gives people pills …
… that allow them to sunbathe while smoking and eating deep-fried snacks all day!
If that’s what they decide they’d like to do, that’s right. The systems approach would usher in preventive medicine. At some early age you’d have your genes analyzed for a predictive health history, and then you would take the two or three or four things that would wipe out the most likely scenarios for disease that you would otherwise face.
If this happens, people on average will live significantly longer and will probably remain quite capable both physically and mentally. How will society deal with people in their eighties and nineties who can really be effective contributors to society? We don’t treat older people very well now so that’s going to cause an enormous amount of change.
This will require a revolution in the education of physicians. Patients already have access through the Internet to detailed information about any disease they may have. When they come into physicians’ offices now they may know more than the physician does about all the latest things regarding rheumatoid arthritis, or whatever, even if they don’t have much context and deeper understanding.
For over ten years, we have been hearing that cystic fibrosis is a primary target for genetic therapy. Is the failure to cure cystic fibrosis genetically just a bump in the road or an indication of deeper problems?
For cystic fibrosis gene therapy has tried to substitute a good gene for a bad gene. Gene therapy has enormous technical difficulties that we’re not anywhere near solving at this point in time. Nor is this gene therapy a systems approach. It will probably be five years or so before systems approaches are applied in many areas and can be assessed.
We’ve studied prostate cancer for the past three or four years using some of these systems approaches and we have made remarkable progress in thinking about how to deal with this disease. There are implications for treating all varieties of cancer. So the skeptics have a right still to be very skeptical because for most correlations between genes and disease we have been in a predictive but not a preventive mode. We can say this is giving you the disease but we don’t know how to deal with it.
I think we will develop a whole range of capacities to deal with genetically caused diseases. Some may be really straightforward and achievable in a matter of five years or so; others may take ten or twenty years. It will depend on the complexity of the system the disease is embedded in and the detail of our understanding. For most systems there are a few nodal points that have enormous impact. If we can devise techniques for identifying and manipulating these key points we may be able to do a lot of good things, even short of a detailed understanding of the whole system itself.
Some studies came out of Princeton four years or so ago on making transgenic mice with a gene that facilitated memory. These transgenic mice with just this one gene modification apparently showed a remarkable increase in memory and ability to learn mazes and things like that. Some people are really skeptical about this, but it may be the case that we can get quite remarkable kinds of effects.
Actually I think the human brain may be the best candidate for early manipulations because there are marvelous possibilities for dealing with neurological disease in the messaging system by which nerve cells communicate with one another (neurotransmitters and receptors and so forth). We already know that the genes controlling those phenomena can have a profound impact on aggressiveness, thrill seeking, and a variety of other things. Still, it is absolutely true that there are very, very few examples of effectively dealing with genetically identified diseases.
I think there will be an enormous convergence over the next fifty years between what living organisms and what inanimate objects and computers can do. Whether we will ever have to worry about these entities taking over the world is a fascinating and complicated question with no clear answer.
The IT world has been troubled by the transmission of child pornography, hate group propaganda, and the malicious introduction of viruses into systems. Are there parallel threats in the biological systems area? Creating a few nine-foot tall basketball players? Biological terrorism? Does creating all this knowledge make us more vulnerable to its misuse? Is there any sort of watchdog or any way of creating public policy? How do we deal with this?
One issue you raise has to do with genetic enhancement and that can be divided into two categories; you can genetically modify people to take care of defects or you can genetically modify for improvement. That’s an enormously big and interesting debate.
The second issue is biological terrorism. Can our increased knowledge be used for terrorist purposes to manipulate populations and things like that? Probably so, but no fancy genetic engineering or gene splicing is really necessary. Classic microbiology can make unbelievably deadly biological reagents and any terrorists who were really smart would use the simplest kinds of techniques to contaminate food supplies or water supplies. This is an enormous problem already, with or without our research.
Of course, you could always send Nazi literature or child pornography or poison through the regular mail. But nobody predicted that pornographers or hate groups would exploit the Internet as they have. Should we anticipate parallel developments in biology as these fields converge?
I have faith that we will be able to understand even the most complex systems by simplifying assumptions and by simplifying models. Eventually, of course, we get to the philosophical question of whether reductionism can give us all of the truth.
Knowledge can be used for good or bad. For some time to come, I suspect that terrorists will use what’s been around for a hundred and fifty years, simple microbiology rather than the more complex genetics. The potential of biological terrorism is a very serious challenge because the tools for doing it are pretty simple compared to making an atomic bomb or other kinds of things.
Terrorism aside, does government play a role in overseeing radical biological interventions? If scientists fundamentally modify our planetary biodiversity the impact on people’s lives in a generation or two could be massive and negative. Should we have government regulation or should scientists themselves set up watchdog organizations to reflect on the possible consequences of modifying the species and ecosystem?
Government agencies already control and oversee commercial applications of genetic engineering and things like that. There are hoops that people must pass through. Parties like the Greens would say there aren’t sufficient controls and constraints. These are matters that could be put up for debate. I’m not convinced that everything the Greens say is true but I do think some governmental oversight is unquestionably important. Scientists also have to have a sense of personal responsibility for the environment, people, food, and these kinds of things. We have to look carefully at what we’re doing and what the implications are.
In the early Seventies when genetic engineering was first created, a group of scientists got together and talked about the implications. Out of that discussion came safeguards that the government then implemented quite effectively. Biology has enormous opportunities for dealing with disease or creating food or animal products But this is accompanied by some significant ethical, social and a legal challenges.
Do economic interests affect your field in a significant way? University researchers, the theory goes, search for pure knowledge, not for profits and marketable products. Do business interests ever make it difficult to pursue truth and the interests of humanity?
Basic and applied research are actually on a continuum with no sharp distinction between them. In fact, very often the applied stuff is more interesting because you can get at it more effectively and there is the excitement of it having a bigger impact on human beings. Society does pay for scientists to do scientific research but I think society deserves scientists to play a responsible role in effectively transmitting to society what they learn.
I’ve had a lot of projects that were so very basic you couldn’t see how to apply them. But I’ve also done a lot of applied research, so I’ve lived in both camps. I don’t see a big distinction. Some people define as applied things that I would see as quite fundamental in nature. Academics today really span that entire spectrum. They are not just sitting in some pure basic research ivory tower …
The question of how commercial interests influence scientists is a serious one. If you get support from a company for a particular project, you have to be very careful about how you carry out that research. For example, it is not appropriate, in my view, to have graduate students working on things you want to get done for a certain company. Grad students should almost never be involved in academic centers that have such contracts …
There is also a potential for contamination, for example, in medical tests of drugs where the medical researchers have equity positions in the companies that are selling the drugs. Companies getting negative results from the researchers may urge that the findings not be published. So the interface between business and academia does raise serious questions and really needs oversight.
If a company has invested hundreds of millions dollars in a research project developing a certain pharmaceutical, they need to market the product to recoup their investment. If you get to a point where you realize your product really isn’t so great . . .
You have to be scrupulously honest. The reason why a drug costs six hundred million dollars to get through clinical trials is that a hundred other drugs only got part way through and had to be abandoned. That’s why these tests end up being so expensive. You hope that there is intellectual honesty but over and over we see drugs getting out in the market that are pretty marginal. The research community has to be very conscious of these issues.
What was your involvement in the Human Genome Project?
I went to the first meeting about the Human Genome Project in 1985. Bob Sinsheimer, the Chancellor of UC Santa Cruz, had thirty-five million dollars and was thinking about setting up an institute to sequence the human genome. So he invited eight or ten of us to come and talk about that for two or three days. I went there slightly skeptical, mostly on technical grounds. But I came away excited for two reasons. One was the excitement over the enormous technical challenges that the project was going to pose. I was utterly convinced that we could solve them in time.
The second was much more profound in some ways. It was the realization that the Human Genome Project was bringing to biology a completely different way of thinking about science. I’ve since come to call this the “discovery” approach to science. The idea of discovery science is to take an object and define all of its elements and then use that as infrastructure on which to do the classic kind of biology that is hypothesis-driven science. There is no question that having the human genome sequence available now is revolutionizing all sorts of hypothesis-driven science interested in human medicine and human biology.
What was both tragic and interesting during the first five years, from about 1985 to 1990, was the bitter opposition to the Human Genome Project from most biologists, based on a fundamental misunderstanding of what discovery science was about. They said it was just “stamp collecting,” wasting money on trivial information and threatening their own funding. The hardest years of my life were going out and making the pitch for the Human Genome Project and running into desperately negative audiences.
Subsequently all that we said has proven to be true. This whole business of paradigm changes is fascinating. Until a group is ready to hear and understand, you can talk until you’re blue in the face and they won’t believe you. In 1989 the National Academy of Science appointed a committee to look into the whole question of the Human Genome Project. Half of the committee members were opposed — half in favor. But after they sat down and really thought about it, every single opponent came over to the other side, and the committee unanimously endorsed the Human Genome Project that started in 1990.
We also developed the critical instrument for carrying out the Project, the automated fluorescent DNA sequencer and we developed a lot of the major research strategies. We were one of sixteen major centers that worked on the project. We have recently published the first description of the human genome.
I think I’ve also been the one who has seen most clearly the consequences of the project. There are three fundamental consequences. First, it introduced to biology the idea of discovery science and today this is an integral part of how sophisticated biologists everywhere think about biology. Discovery science is taking a biological object and defining all of its elements, like taking the human genome and determining the sequence and order of each of its three million nucleotides or taking all of the DNA transcripts present in a particular cell and quantifying the expression patterns, their ratios of expression. These analyses don’t tell you anything in themselves but you can use them to do incredible hypothesis-driven biology.
Second, the Human Genome Project has given us a sort of periodic table of the elements of life. It defined the genes and gave us access to the variation in the human genome that is the basis for disease predisposition in many cases. It gave us the sequences around the genes, called regulatory regions, which actually dictate how they’re turned on and off. The regulatory region is a second very fundamental informational system in human chromosomes. Now we have the information to decipher the regulatory code in the future.
Finally the project led to a whole series of paradigm changes that ended up with this view I’ve described — systems biology and the hierarchical levels of information. This will be the future of biology. We played a major role in articulating these ideas and our Institute for Systems Biology is dedicated to this whole approach to both biology and to medicine.
The work that you’re describing is fundamentally interdisciplinary but universities by their history have been set-up along fairly strict boundaries with funding sources and review processes that follow those strict boundaries. It seems to me there’s going to be a shift in how universities think about research. A professor with two graduate students can’t do much on a very large, complex problem.
One major problem is the existence of departments. Departments are enormous barriers to cross-disciplinary science and the diffusion of many different kinds of disciplines into another discipline — physics, chemistry, and computer science into biology for example. A second problem is that academic institutions tend to be enormously bureaucratic and constrained in the kinds of partnerships that they can set up and in their timely responsiveness to new opportunities. A third problem is that they’re enormously limited in resources. A good university is only a franchise. In most cases scientists have to raise all the money they need to do any kind of science.
One enormous barrier in universities is tenure. That’s not only because it often ensures the preservation of mediocrity, but because the process of making the tenure decision imposes a value system on young professors. First, the tenure system takes young professors at the most creative stage of their careers and compels them to do straightforward, publishable research. Real innovation is considered very dangerous.
Second, the tenure system counts and rewards only your own individual work. Any collaborative research you do is unsatisfactory. So the tenure process itself is inimical to the cross-disciplinary approach that is the future of the science of complexity.
Those are the reasons that I resigned from the university and set up an institute where we would have the freedom to organize without departments, to be able to set up academic and industrial partnerships and to bring together cross-disciplinary scientists in a single, unified environment so the collisions would lead to interesting opportunities and possibilities. The universities that are going to be serious players in this post-genomic world are going to have to modify their organizational structures to deal with these opportunities.
I just finished a project with the National Academy on the future of information technology and we proposed a book called Making IT Work. We agreed that the issues in IT are fundamentally interdisciplinary, complex, and of a large systems character. The environment of a single department with a couple of grad students and a typical funding cycle won’t work. But academic institutions seem to be structured in concrete. Do you see any change happening?
I think the elite, private universities will have the flexibility to set-up structures that can encompass these kinds of opportunities. The state universities often will have a much more difficult time generating this kind of flexibility. Cross-disciplinary institutes probably need to be set up with large degrees of freedom. The problem of resources needs resolution because what we’re talking about is not inexpensive science. While you can do pretty well with grants and so forth you do need fundamental basic infrastructure of contributions to set up the kinds of things you need. So I think there will be, especially in the private sphere, universities that adapt in time but it will take some time.
The other problem with many academic centers is that you need a leader who has a clear vision to make this happen. That’s pretty hard to do in academia. There are an awful lot of constraints on leaders with vision. Even if you have your own institute there are lots of constraints.
Universities must ask what they can do well. Maybe research is not what they are able to do best for all the reasons stated here. Major cutting edge research that affects society and the world may need to move off-campus to places where there’s more flexibility.
But the universities still have to fundamentally change their teaching to train cross-disciplinary students. In the past, research has driven teaching changes. If you divorce the sciences of complexity from universities they will become far less effective in teaching. That’s my worry. I don’t think there are simple solutions to any of these problems
Do certain ethical values need to guide the kind of work that you’re doing?
Our institute has three missions. One is to do systems biology and systems approaches to medicinal research. The second is to develop completely new approaches to teaching biology, centered on this informational point of view (we’re writing a textbook on that). The third is to bring science to society. One of my missions is to educate society so that it can think and make rational decisions about these issues without being influenced unduly by the Jeremy Rifkins and demagogues of the world.
Most of my career I have had a major commitment to K-12 science education. We work with school districts, focusing on professional teacher education. We’ve set up programs for the seventy-seven elementary schools in Seattle, providing all of their 1400 teachers with 100 hours of instruction in inquiry-based science in summer institutes and in-service training. If we can set up similar programs in middle schools, we will produce kids with an ability to think in an inquiry-based, analytical manner. They will have a deep enough knowledge about science so they can think about some of the great environmental or biological questions.
We have for example a program in high school now where we’re teaching kids how to sequence DNA and that’s an enormously exciting thing for them to do. One of the modules divides the class into groups of four. Then they each pretend they’re a family with Huntington’s disease. We give different members diagnoses as to whether they have the gene or not. And then we teach them how to analyze this ethically and think about their reactions to the disease, to other family member’s and so on. After three or four sessions kids walk away with a deeper understanding of these ethical issues.
Later this year we’ll be setting up a workshop for leaders in industry, science, medicine, and journalism, to talk about how to educate the public so we don’t have the kind of thing that went on in Europe with the Greens and Monsanto and genetically engineered food.
So my mission is to try to educate people on the nature of these issues where science impacts society. I co-authored a textbook on the Human Genome Project that dealt in considerable detail with the social, ethical and legal issues. I believe that every scientist should spend five to ten percent of their time in educational communication. That is as fundamental to being a scientist as doing good research.