Martyn Amos
GENESIS MACHINES
The new science of biocomputing
320pp. Atlantic Books. £18.99.
978 1 84354 224 7
Robert Frenay
PULSE
How nature is inspiring the technology of the twenty-first century
576pp. Little, Brown. Paperback, £15.99.
978 0 316 64051 4
US: Farrar, Straus and Giroux. $30.
978 0 374 11327 8
Predicting the future of technology is a mug’s game, but that doesn’t stop people from trying to do it. I still enjoy looking through my 1902 copy of The Romance of Modern Invention, which devotes as much space to the telautograph as to the telephone, and predicts that the horseless carriage will solve the problems of city congestion. The two very different books by Martyn Amos and Robert Frenay share the premise that biology is going to be extremely important in twenty-first-century technology, taking over from the electronics that has dominated recent decades. There are good reasons for expecting change, one of these being that there are physical limits to how small silicon-based electronic circuits can be made. Smaller means better in the world of computing, and computers have been shrinking in size for the past forty years. Moore’s Law states that the number of components that can be packed onto a given silicon chip doubles every eighteen months, but the end is in sight for this steady progress. In contrast, biological systems manage to store and manipulate information more compactly than any silicon-based device can achieve. In particular, DNA holds information in digital form, just like a computer, and uses fewer than fifty atoms to store one bit.
Martyn Amos is a computer scientist who became interested in the possible use of DNA as a computing medium, and as a result claims to have earned the first ever doctorate in biocomputing. In Genesis Machines, he provides a readable and engaging description of his and others’ adventures in the overlap between two major realms of current technology, computation and biotechnology. Along the way, he provides lucid explanations of both the computing and the biology, though some of his examples are let down by bad typesetting, and he makes a few minor mistakes in biology.
Genesis Machines has three main strands. First is the use of DNA to make tiny computers, which is where Amos’s main expertise lies. For solving some computational problems, it is possible to encode the problem in a complex soup of many different DNA molecules. Complementary DNA molecules are able to find and pair with one another, so with the right tricks a unique solution can be extracted from the soup. Gradually, more and more complicated problems are being attacked by this kind of approach, which is radically different from conventional computing. The second strand is the use of DNA in micro-fabrication, to construct minute structures by exploiting the ability of DNA to fold up in specific patterns dictated by base pairing. This process, sometimes called DNA origami, is tremendous fun, and has led to recent advances such as the synthesis of bits of DNA that spontaneously fold up into two-dimensional shapes like smiley faces, or maps of America. The smiley faces are a few millionths of an inch across, so they can only be seen with an atomic-force microscope, but they can be made by the billion, because it is so easy to replicate DNA. As a result, the scientists involved joked that they had been responsible for “the most concentrated happiness ever created”. Other scientists have gone on to make tiny motors attached to a DNA scaffold, or stable three-dimensional objects made of DNA. There seems to be no end in sight for ingenious creative developments in this area.
The third strand is synthetic biology, which is the use of whole cells, rather than DNA molecules, to carry out simple logical operations, and hence to act as core elements in a biological computer. Again, the advantage lies in compactness: bacterial cells are only about a micron across, so huge numbers can be packed into a tiny volume. Furthermore, bacteria are very good at taking care of themselves and are easy to manipulate genetically. Synthetic biology is largely based on our accumulated knowledge of the control circuits that bacteria naturally use to regulate the activity of their genes and proteins; synthetic biology allows scientists to subvert these circuits to respond to new inputs and generate new outputs. For example, one team developed bacteria that oscillate between states, generating fluorescent protein on each oscillation, so they behaved as microscopic blinking fireflies. Others have constructed light-sensitive bacteria that can be used to make high-resolution photographic film.
So far, most of the applications of biocomputing have been to “toy” problems, like devising biocomputers to play noughts and crosses. This is a necessary start, providing proof of principle; but scaling up to take on larger and otherwise intractable problems is not so easy. A simple DNA computer occupies only a tiny drop of liquid, whereas attacking a real problem might need a whole bathtub full of DNA, which is clearly impractical. Amos is well aware of these problems, but remains optimistic. Development is also not moving very rapidly: it is now over ten years since the first report of DNA computing, and it took two years to build the bacterial oscillator, which is an extremely long time from the point of view of the average genetic engineer. At present, the field is still looking for the “killer application”, which would employ these new technologies to achieve something previously impossible, and preferably also useful and marketable. Once this happens, the speed of development will pick up enormously.
There is a fair amount of digression in Genesis Machines, most of it entertaining and relevant, such as the biography of bad boy Kerry Mullis, inventor of the Polymerase Chain Reaction and general maverick, but some, like Amos’s account of his failure to appear on television, unworthy of inclusion. There are also some significant omissions: it is surprising that he makes no mention of quantum computing, which is heralded as another technology that might change the computing world, and would probably render DNA computing obsolete. Again, the practical applications of quantum computing lie in the future, but advances willcontinue at an encouraging rate. The author also fails to discuss aptamer development, which is one way of playing with nucleic acids in test tubes that has resulted in tangible benefits.
DNA, or better yet its cousin RNA, can be made to undergo accelerated evolution in vitro, by selecting for desirable kinds of molecular recognition or chemical reaction. When this works successfully, repeated cycles of selection and amplification generate novel RNA molecules known as aptamers. One of these RNA-evolution methods has already yielded a clinically useful aptamer drug for treating retinal macular degeneration, a common disease of old age. These developments are relevant to Amos’s general theme, so it is a shame that they are omitted. But he does an excellent job in conveying the excitement of working at a remarkable new frontier of human ingenuity and invention.
Robert Frenay, in Pulse, covers a far wider canvas, far less successfully. He deals with everything from thermodynamics, evolution and artificial intelligence to farming, town planning, global economics, advertising and multinational corporations. Inevitably, the coverage is shallow, and his mistakes, such as confusing genes with base-pairs when writing about the human genome, are not reassuring for his general reliability as a guide. His basic thesis is that what he calls the New Biology can change not just technology, but also economics and politics, by the emulation of natural living systems. This is not a surprising position, given Frenay’s background as a writer for the environmental magazine Audubon, and much of Pulse reads as a standard environmentalist tract. The combination of so many different themes might have led to an inspiring and persuasive vision, but his analyses and suggested solutions seem simplistic and utopian. For example, he implies that money is the root of all economic evil, and all would be well if we switched over to barter economies. Another of his hobby horses is that measures of national prosperity have been perverted by tracking GDP rather than GNP. And he believes that the rise and dreadful behaviour of multinational corporations stems from an 1886 court decision in America, which awarded corporations the same Fourteenth Amendment rights as persons. Even if this were convincing, overturning the decision wouldn’t put the genie back in the bottle.
Pulse is not helped by being written in mushy journalese, interspersed with lusher passages rhapsodizing over snails on “the famed White Cliffs of Dover”, and so on. There is also a great deal of vacuous name-dropping and repetition, with frequent invocation of “the sweet spot” between order and chaos, as the only place to survive. The problem with sweet spots is that there may not be much room on them, especially if everybody decides to move there at once, and they rarely stay sweet for long.
The book takes its title from two meanings of the word: on the one hand, in the sense of a heartbeat, a rhythmical signature of life; and on the other, as a kind of bean. Actually, Frenay repeatedly writes “a pulse is a seedhead”, in his desire to invest the word with seminal depth and resonance, but “seedhead” is scarcely a standard definition of “pulse”, and this odd usage reveals some of the flaws in his thinking. Seeds are not the same as the plants they can give rise to, and a seedhead contains many seeds, not just one. Nature works by creating a vast excess of new potential organisms at each generation, only a few of which will survive and be successful, as a result of chance or better adaptation to new circumstances. This is not what happens with human societies or economies, so their evolution has to proceed differently. Frenay’s belief that economies should be based on natural systems is also unrealistic. Economic complexity is profoundly different from the complexity of the natural world, not least because the agents studying the system are themselves part of it, and can influence it by their own predictions and by fashions in economic theory; human beings are also frequently irrational in making financial decisions. Economic forecasting therefore remains a very dodgy business.
Both Amos and Frenay write with reverence about the Santa Fe Institute, a think tank for the study of complexity, and particularly about one of its gurus, Stuart Kauffman, who has become celebrated for his suggestion that interactive systems tend to self-organize to a state close to chaotic behaviour, because that is the most productive place to be. This is certainly interesting, but it is not clear how or why things should be that way. Self-assembly is a familiar and fundamental organizing principle, of course (Amos provides a charming demonstration, using a bowl of milk and a few Cheerios), but self-organization of a complex interactive system, to a point just the right distance from disorder, is another matter. Kauffman may be on to something, but he has a track record of intellectually appealing ideas that turn out to be mostly wrong. Long ago, he published various clever theoretical models to explain the development of a fly egg, which gets subdivided into a pattern of stripes on its way to becoming a segmented maggot. In Kauffman’s models, the stripes would all be generated in similar ways, but we now know that this is not what happens in reality. Instead, each stripe is formed by using a separate set of almost arbitrary instructions; it is very inelegant, but it works.
This example illustrates one of the profound differences between natural engineering, which is what evolution does, and human engineering. Biological structures are frequently kludges – to use the engineering term for machines or computer programs put together by a series of ad hoc modifications. The end results may be magnificent and superbly adapted organisms, but their underlying construction can be far from optimal. Moreover, the quirks of evolution, acting via sexual selection, can also drive the formation of cumbersome ornaments like a peacock’s tail or the antlers of a moose. So, slavishly copying Nature may not always be a good idea. There are certainly cases where living organisms have provided the inspiration for new technology: Velcro, inspired by cockleburs, is a favourite example. But there are many others where the natural precedent was only recognized in hindsight: engineers did not need to know about bats in order to invent sonar.
How much can be achieved by imitating the natural world also depends on two large questions. The first is whether there is anything that natural selection cannot do, once the necessary conditions of variation, inheritance and replication are in place. Developers of nanotechnology are frequently criticized for running the risk of creating “grey goo”, tiny artificial replicating robots or nanobots, which could run riot and multiply to engulf the world. Those who conjure up this Doomsday scenario rarely point out, though Amos does, that natural tiny robots already exist, in the form of bacteria, and have flourished on this planet for billions of years. Any artificial grey goo would find itself competing with natural brown scum, which it would be unlikely to displace. This is because bacteria have been around for so much longer and have already explored such an immense volume of phenotypic space, efficiently colonizing all possible niches from ice floes to kitchen sinks and nuclear reactors. It is unlikely that human design could ever improve on bacteria, on the scale of single-celled microscopic life. Things get more complicated at the level of multi-cellular organisms and ecosystems because there may indeed be room for improvement. How hard this will be remains to be seen.
The other big question is whether computer simulation of natural systems and even economies can get sufficiently good to be really reliable, as guides to the real world or as testbeds for new ideas. Much research money is currently being thrown at systems biology and other attempts to model biology in computers. Computers continue to get more powerful, so maybe these simulations will become effective and usable, but there are formidable difficulties involved. Significantly, Frenay makes no mention of Biosphere 2, a spectacularly unsuccessful attempt to construct a large artificial eco-system in Arizona, complete with human occupants. This failure shows how hard it is to design ecologies from scratch. On the other hand, one fascinating development in human society, which Robert Frenay does write about, is the rise of virtual worlds or cybersocieties. Millions of people now indulge in alternative lives, thanks to on-line role-playing games like AlphaWorld and EverQuest. We have even reached a point where the proliferation of parallel cybersocieties is allowing accidental tests of sociological theories. Again, the validity of those tests depends on how well cyber-societies can mimic real ones, but in terms of important socio-economic developments, this is undoubtedly an area to watch.
Martyn Amos appositely quotes one of the last remarks of Richard Feynman, the great physicist and prophet of nanotechnology: “What I cannot create I do not understand”. True enough, but it does not follow that creating something means fully understanding it, and one can only be confident that new technologies, like the car and the mobile phone in the past, will have unpredicted consequences. How boring and pointless life would seem, if it were not so.
_________________________________________________________
Jonathan Hodgkin is Professor of Genetics at the University of Oxford.
