Click for next page ( 52


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 51
PART iiI

OCR for page 51

OCR for page 51
The Evolving Relationship Between Instrumentation and Research-- A panel discussion1 Panel Participants: William Ballhaus,2 Robert Gaensslen, Leroy Hood, Gabrielle Long, John Roberts, Michael Roukes, Vincent Shankey, Wm. A. Wulf3 William Wulf: I'm a computer scientist, and my career has pretty well corresponded with Moore's law. For 45 years it feels as if I've been sitting on the 50-yard line watching an incredible increase in technology. One byproduct of such rapid change has been that I have heard some really bad predictions about what's going to happen in the future. When tech- nology is changing as rapidly as it has for 45 years, making predictions about the future can be dangerous. That said, I've suggested to the panel that we spend some time looking into the future. Gabrielle Long: At the end of my talk I was discussing the linear coherent light source, which is essentially the x-ray laser. Such an instrument will create many opportunities, which I would put into two categories: chemistry and biology. The x-rays from this kind of device come in pulses and will be extremely short, on the order of a few femtoseconds. Chemists will be able to activate reactions and then watch the products go through vari- ous intermediaries to get back to the ground state. That kind of experiment is going to be very much possible with the linear coherent light source. It will open up an area of chem- istry that we are only beginning to glimpse today. Regarding biology, we're used to doing structural work on biological materials that have been crystallized, and crystallography has been enormously successful. The dream, how- ever, is to be able to do structural analysis on a single molecule. When we finally have enough coherent photons in a single pulse, the molecule may fly apart from the impact, but it will not fly apart until we find what its structure is. Understanding these kinds of 1 As prepared by Steven Olson. 2 William F. Ballhaus, Sr., former president, Beckman Instruments, Inc. Dr. Ballhaus (Ph.D., California Institute of Technology, 1947) began his career in aircraft design and engineering at Douglas Aircraft, Inc., around the time of World War II; he then joined Northrop Corporation in 1953 as chief engineer and became executive vice president in 1961. In 1965, Arnold O. Beckman recruited Dr. Ballhaus as his successor to operate Beckman Instruments, Inc. Under Dr. Ballhaus's leadership, Beckman Instruments, Inc., made the transition from government contract work to the emerging markets of medical and biotechnology. He remained president of Beckman Instruments until 1983, when the company merged with SmithKline Corporation. Since then Dr. Ballhaus has served as president of International Numatics, Inc. He is an elected member of the National Academy of Engineering and is a fellow of the American Institute of Aeronautics and Astronautics. 3 Wm. A. Wulf, president, National Academy of Engineering, and vice chair, National Research Council. Dr. Wulf (Ph.D., University of Virginia, 1968) is on leave from the University of Virginia, Charlottesville, where he is AT&T Professor of Engineering and Applied Sciences. Among his activities at the university are a complete revision of the undergraduate computer science curriculum, research on computer architecture and computer security, and an effort to assist humanities scholars exploit informa- tion technology. Dr. Wulf has had a distinguished professional career that includes serving as assistant director of the National Science Foundation; chair and chief exec- utive officer of Tartan Laboratories, Inc., Pittsburgh; and professor of computer science at Carnegie Mellon University, Pittsburgh. He is the author of more than 80 papers and technical reports, has written three books, and holds one U.S. patent. INSTRUMENTATION FOR A BETTER TOMORROW 53

OCR for page 51
structures at a level never before possible will enable us to explore areas of biology that are not now accessible. Leroy Hood: Biology has its own Moore's law, which is that all types of biological information will undergo an exponential expansion. The fascinating question is how we can use this information to gain some understanding of organisms. I think the bottom line is systems biology, which will dominate the twenty-first century in powerful ways. It is imperative that we develop techniques that will enable us to do global analyses of biological systems. One of the enormous challenges in doing such analyses is dynamics. Biology is all about transitions, whether developmental, physiological, or even evolutionary. Can we understand these dynamic transitions? Another big challenge is discovering how cells interact in complicated ways to create the properties seen in living organisms. At the institute we're using immune system cells because you can put them together in various ways, which makes an ideal model system. This enables us to digitize biology. We can interrogate individual molecules and cells. We will be able to put combinations of cells together and see how they create emergent properties. Another thing is controversial but I am convinced it will happen. Through computational methods we will be able to fold all proteins and see what their likely behaviors are. Furthermore, we'll be able to deduce the circuitry of life.We'll be able to recreate protein net- works and gene regulatory networks. What we won't know from these computational networks is how the environment impinges on biological systems. So there always will be a marriage of experimental and theoretical sci- ence. In fact, all biology needs to be grounded in experimental data. Within 10 years or so we will have aspects of the predictive medicine I've been describing. And in 15 years or so we'll be in the exponential phase of the advance of preventive medi- cine. Once we develop these tools, we'll be able to say that if you start taking these pills at age 54 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 51
38, you won't have to worry about this disease. I'm a skeptic about training people to stay out of the sun, stop smoking, and quit eating too much. Finally, a lot of the big problems in the world have to do with food. I think these approach- es will absolutely transform agriculture. Nutrition is really in the dark ages. If there was ever a discipline that desperately needed a systems approach, it's nutrition. These are enormous applications. William Ballhaus: One thing that most people don't know about Arnold Beckman is that his "If I could company had the first bioengineers. When I was working with Douglas Aircraft, one of the finest aeronautical engineers I worked with was Curt Miller. We had both received scholar- choose one ships to go to Caltech--I got my doctoral degree, and he got his master's degree. When it was over, I asked Curt what he wanted to do, and he said he wanted to be a medical doctor. And word that I so he went to medical school for 3 years and then did a residency for a couple more. would attribute In 1965, when I became president of Beckman Instruments, I happened to see a mutual friend of ours, and I asked where Curt was. He said that he was working at Aerojet, and so I called him the next day and asked him whether he would like to be director of medicine to Dr. Beckman, for Beckman Instruments. He was a medical doctor and an aeronautical engineer, so I had him sit down with our chief scientist and tell him what the medical profession needs. it would be I said that I wanted them to develop a set of instruments that medical doctors can use. That created a medical instruments department that became one of the fastest growing measurement." parts of the corporation. --WilliamBallhaus If I could choose one word that I would attribute to Dr. Beckman, it would be measurement. But the demands made of measurement vary. When you look at what the electronics people have been doing, they have been packing more and more capability onto smaller and small- er things. Pretty soon the electronics people are going to have everything on nothing. But if you look at biology, billions and billions of proteins have to be analyzed and measured, so there is no question that bioengineering is going to be the future. John Roberts: I feel very much like an anachronism in this discussion, because my focus on chemistry is entirely different. After I was a provost, I had to find some way of getting back into science. I did a little work on MRI, but I had to compete with patients for an MRI machine, and that didn't work out well at all. I also could not get the kinds of graduate stu- dents and postdocs that other people have in abundance. I had to work primarily with undergraduates. So I had to think about research in a different way, because you can't give an INSTRUMENTATION FOR A BETTER TOMORROW 55

OCR for page 51
undergraduate some of these tasks and get very much done in the course of a summer fel- lowship. I had to go back to some of the elements of organic chemistry and work on prob- lems that were not well known. One of these problems involved the structure of proteins. People don't always realize that the interior of protein is not filled with water but with something else, and the forces are very different. We started to study some of the simplest things you can study, the most "Nano is a favorable conformations around single bonds. In doing that, we discovered some quite unexpected things. In some compounds, two negative charges tend to want to be close methodology. . . . together rather than far apart in aprotic solvents. These things are interesting, and they may have applications. The difference is My point is that we shouldn't abandon basic chemistry completely. Lots of what's there has not been exploited well enough, and we don't understand it well. I'm enjoying this work that nano is a very much. We had nine people working this summer, so there is a role for this kind of research activity. methodology Michael Roukes: I'm not going to propose that the next big thing after nano is pico. Nano that allows is a methodology. Chemists could claim that they have been doing nano for a couple of centuries. The difference is that nano is a methodology that allows us to interact predic- tively with individual molecules. us to interact Many things that are glimmers in the eyes of scientists who are in the field today are going predictively to become robust in the same way that measuring tools Beckman built were brought into routine laboratory use. For example, one exciting possibility is doing dynamical measure- with individual ments at the nano scale, really understanding at the level of single molecules the interac- tions in individual cells. We could follow this process in real time and understand its molecules." properties by building up statistical aggregates of individual entities. --Michael Roukes I also believe that the Norman Rockwell picture of the family practitioner is going to fade into the sunset. Instead, you're going to plug your real-time attributes into a cold, calcu- lating, massive database and out will pop a quantitative assessment of your physiological state, your proclivities for the future, and the preventive measures you should take to avoid disease. Part of the transition to that era will be the ability to mass produce these nano devices robustly. Ultimately I believe that these devices will be implanted in our bodies and will monitor our physiological state in real time. This will be transformational. 56 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 51
Robert Gaensslen: I'm at the applied end of this discussion. One trend will be to get a lot of the measurement, technology, and scientific testing out of the lab and into the field. There's no reason why a very sophisticated machine cannot be put in the hands of a cop, as long as the data go back to the lab and the cop isn't responsible for figuring out what the data mean. If some of these technologies can be boxed up, they can be complete black boxes for crime scene investigators, which would create huge efficiencies. We need to learn how to individualize trace evidence. That would make trace evidence much more valuable. Showing that two red fibers have a common origin does not come up in every case, but where it does come up it's important. Regarding biological evidence, if all of the predictions I've heard come true, you can bet that the FBI is going to have one of the receivers for these devices that are implanted in people. How much information are you willing to give to law enforcement, or are you will- ing to let law enforcement have in a databank in order to keep you safe? These are impor- tant public policy questions that ultimately the citizenry is going to have to answer. The freer the society, the more crime it will tolerate. There's little crime in Singapore, but there's not much freedom either. Some law enforcement people are already saying,"Why don't we put all suspects in the databank right now?" The British do this, and they get a lot more hits on suspects. Are you going to have half, three quarters, or all of your population data- banked? That's a public policy question. You could catch every criminal who leaves bio- logical evidence. But is that what we want? Do you want the FBI to know about your health status and when you're likely to have that next heart attack? I don't. It's a matter of how much you trust your government and how much you think the government, on the basis of its past behavior, has given you reason to trust it. Arnold Thackray: I have a question for the panel. If you look at innovative organizations, probably the most successful long-term innovative organization in the world has been the Catholic Church, though we don't usually think of it in those terms. The last time I checked it had more members and a longer history than most other organizations. Universities are about the only other organizations that are serious competitors. Closer to home, General Electric set up an industrial lab in 1901, and the last time I checked GE was doing okay as a company. DuPont, too, has reinvented itself two or three times. So I don't think you can just push innovation entirely off to small companies. Leroy Hood: All of the big companies I know, such as GE and IBM, tend not to innovate. They buy innovations that other people have created and then take that innovation to the INSTRUMENTATION FOR A BETTER TOMORROW 57

OCR for page 51
The Arnold and Mabel Beckman Center of the National Academies. Courtesy of the Beckman Center of the National Academies. point of proof. Don't get me wrong. Big companies are very good at adapting to opportuni- ties and integrating things that have been proven.And organizations like the church and uni- versities are enormously innovative in the context of a bureaucracy that has been honed to do what they set out to do. My point is not that big organizations can't innovate. But they innovate around the conditions that their administrative structures have set in place. I do have a wonderful model for academia and industry that is based on what Jack Whitehead did with the Whitehead Institute at MIT. What persuaded him to establish this institute at MIT was David Baltimore--he found a soulmate who had a common view. He was tough with MIT, because the university was anxious to take this $150 million and go with it. He said, no, my institute is going to be entirely independent from MIT except for a couple of things. One was jointly recruited faculty, which was important for quality con- trol. The second was that the provost had veto power over the choice of the director, which was also a quality control provision. But in all major regards--managing the budget, per- sonnel, the ability to make decisions about how to use resources--the Whitehead Institute was totally independent. I would argue that the Whitehead is the best example of a really successful research insti- tute in the world today. Even more important, MIT has enormously benefited from the Whitehead. A large percentage of its biology faculty has come via the Whitehead. And the institute could be enormously flexible. When Eric Lander was considering leaving the Whitehead, the institute decided to make an enormous contribution to his work in a mat- 58 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 51
ter of less than a week. Can you imagine any kind of academic institution making that kind of decision? I think educational institutions are great and I don't want to replace them. But we need to think of innovative ways for them to create new kinds of opportunity. John Roberts: I've been a consultant with DuPont for 56 years and I know something about what happened. They got into the instrumentation field with protein synthesizers and so on, "The judge is and they had good instruments but did not have the commitment to finish them. They did some very innovative things, but they are generally not very good at doing the kind of thing willing to admit that Lee Hood is talking about. that the sun will With respect to advances in forensic chemistry, I have a question. I want to hear about how these esoteric and complicated new methods can be presented to juries. I'd also like to hear rise tomorrow, about preventative work, rather than solving evidential problems afterward. Robert Gaensslen: There are two schools of thought about how to testify. The O.J. Simpson and you don't trial is philosophy A. You try to teach everyone everything you know about the subject. It doesn't work. Everyone's eyes glaze over, and nobody knows what you're talking about. The have to bring an second philosophy is that you give the jury the bottom line. It's his DNA. That's all they need to know. You can say a few words about how you came to this conclusion. They're not going expert to testify to understand exactly what you did. But if you're credible and come from a credible labora- tory, they're going to believe you. about that. But The other thing to understand about an expert witness is that science does not fall within the everything else is realm of what courts call judicial notice. The judge is willing to admit that the sun will rise tomorrow, and you don't have to bring an expert to testify about that. But everything else is your opinion. So when you go to court, you're up there by yourself, and this is your opinion, your opinion." and the jury can totally disbelieve you, no matter how good the science is. Look at what hap- pened in the Simpson trial. I'm convinced that the DNA was right, though there was sloppi- --Robert Gaensslen ness in the case, but the jury simply disregarded the DNA evidence, and they are entirely entitled to do that. Prevention is a good idea. The biggest cause of violent crime in the United States is the drug problem. It's huge. Nothing else makes a dent. I've been in this field for 35 years now, and the problem was about the same in 1970 as it is now. The drugs may change, but the problem doesn't change. This is something that is hardwired into human nature that we have to do INSTRUMENTATION FOR A BETTER TOMORROW 59

OCR for page 51
something about. It's obvious that we can't just burn down coca fields in Bolivia and solve this problem. There's demand, and in every generation the demand seems to stay the same. It eats up resources like you can't imagine--laboratory resources, prosecutors, jails, prisons. Michael Roukes: I want to say something about the invasiveness of having our medical profiles accessible to large numbers of people. It seems inevitable to me. It's not a matter of whether it's going to happen--it is going to happen. Futurists talk about two different "Does the person populations--people with technology, and people without. I have great faith in humankind that we have the collective will to make things better, and I think science can who's identifying harness that will. a drug need to Leroy Hood: The predictive part of medicine does bring tremendous ethical dilemmas. But they're ethical dilemmas only if you can't prevent disease. If you can, who cares whether know how a gas you have that defect, because you can take care of it. Audience member: I'd like to ask about sophisticated instruments being black boxes with chromatograph/ no user-serviceable components. Is that a good or a bad thing? mass spectrometer John Roberts: I once wrote a book that essentially said, "If you want to learn about NMR, buy this book. If you're not interested and you're willing to let things be black boxes, works? It's an don't buy it." We're willing to accept the computer as a black box, even though it may freeze up on us from time to time and we get mad at it. But with NMR, and with I don't open question." know how many other modalities, I don't know how we're going to handle the problems of people not understanding what's going on inside these things. Maybe you can make instruments so reliable and so able to interpret the data that all you have to do is have it --Robert Gaensslen print out a complete diagnosis of what you've done; maybe it will even tell you where you've made a mistake. I've had a scientific argument with a man who claims that he can measure NMR splittings to hundredths of a hertz. But according to theory, that would require listening to the free induc- tion decays (FIDs) for 100 seconds. It turns out that you can't listen to an FID for 100 sec- onds because it seldom lasts more than 4 or 5 seconds. So how do you resolve that question? But when you run a spectrum, you get printouts of results to 0.001 Hz. That's the digital res- olution, but it doesn't correspond to reality. I don't know how to deal with the problem of people believing what the computer tells them, even if the accuracy of the results does not correspond to the printouts. This is a major problem with quantum calculations. 60 INSTRUMENTATION FOR A BETTER TOMORROW

OCR for page 51
Michael Roukes: There are at least two camps of people. There are the experts who are developing instrumentation and want to open up the hood and supercharge whatever's in there. But once the car is built, there's a large population that wants to use it to go some- where. Both groups have happily coexisted in the past, and both will in the future. If your explicit question is whether I think that 20 years from now we'll have single-atom MRI and it will be robust, the answer is, "I hope so." Leroy Hood: There are three stages of instrument technology. The first stage is the proto- type. Then there is a stage when an instrument is made robust. And then there is the peri- od when you move an instrument into an automated context. You need much more understanding of an instrument at the early, immature stage. If you think about the three kinds of major technologies in genomics, for example, they range from immature to teenage to relatively mature. Proteomics is very immature, and you have to understand it very well before you know what to believe. DNA arrays are intermediate-- you need to understand the statistics. DNA sequencing is now a very mature technology. Robert Gaensslen: There's another angle of this that comes into play in forensics labs. Forensic labs use these machines in their mature stages to look at specimens for which the machines were not designed. It's fine if you're looking at cell lines and if you know you have a single source. But most of the forensic people doing these analyses don't fully understand what these machines are doing, even if they have Ph.D.s. Furthermore, when something is owned by a company, the company is not necessarily going to tell you what's going on in the box beyond a certain point, when things become proprietary. This also becomes an issue in court involving how much INSTRUMENTATION FOR A BETTER TOMORROW 61

OCR for page 51
an analyst who is testifying needs to know. Does the person who's identifying a drug need to know how a gas chromatograph/mass spectrometer works? It's an open question. Vincent Shankey: What is our role as scientists and what is the role of scientific institutions in trying to make our culture more scientifically sophisticated? Leroy Hood: My philosophy has always been that one of the obligations of a scientist is to transfer knowledge to society through K-12 science education. In Seattle we've set up a K- 12 science education program that is focused on innovation, inquiry, and teacher training. The program includes the entire set of elementary schools in Seattle--72 schools, 1,100 teachers, and 23,000 students. Almost 100 percent of the teachers have been instructed, and test scores for science in the elementary program have changed in a striking way. We have a similar program for middle school. In high school it's more difficult. What we've done there is to develop modules for systems biology. The kids work with these graphical networks, and they love it. And the teachers love it, because it's something new and not recycled. If every scientist made a commitment to helping K-12 science education in his or her own community, it would change a lot of the suspicion and hostility. NSF is designing a series of kits that are really great. Our inquiry-based teacher training is organized around those kits. I always argue that the most important thing students should acquire is the ability to do analytic thinking. But one thing I worry about is how deep the understanding of the teachers and kids is. Another problem is that in the early stages our program was supported by a $2 million NSF grant that has since expired. Where does the money come from to continue to support the program? The school district has more than enough money to support this program, but it's all entitled. So I spend a lot of time rais- ing money to keep this going. But in the long term, if you can't integrate these programs into the school system, they won't be viable. I've looked at a bunch of programs in systems biology, and you really have to have one pas- sionate leader who is willing to spend a lot of time and make it happen. For example, at Princeton, David Botstein has single-handedly fostered tremendous relationships among physics, engineering, and mathematics, and I think Princeton is creating a really terrific program. He came in with a mandate to do that. How you make the environment recep- tive for that kind of change is an issue many schools are facing today. William Wulf: Well, I think this has been a wonderful symposium. I believe this is the first time I have heard the prefix yocto used, which I think stands for times 10 to the minus twenty-fourth. With that, I'd like to adjourn this session and the symposium. Thank you, and let's thank the panel! 62 INSTRUMENTATION FOR A BETTER TOMORROW