What Mad Pursuit

by Francis Crick

  We had planned to give our results a decent burial in the august pages of the Philosophical Transactions of the Royal Society. As we now had a result of some interest, we took the experiments out of the proposed Philosophical Transactions paper and made them into a separate paper that appeared shortly afterward in Nature. I was somewhat surprised to find my name on the draft paper, since the convention in our lab was that one did not put one’s name on a paper unless one had made a significant contribution to it. Mere friendly advice was not enough. “Why,” I asked Sydney, “have you added my name?” He grinned at me. “For persistent nagging,” he said, so I let it stand.

  One of the more laborious experiments that Leslie did was to put six +’s together in one gene and show that the result was like wild type. It is difficult to convey just how tedious and complicated such an experiment is. The required (+ + + + + +) must be put together in stages, testing at each stage to see that the gene does indeed have the structure it is supposed to. When the final combination has been produced and tested, it must still be taken apart, step by step, to make sure that it was what we thought it was. Even an outline description of all Leslie did took up several of the large pages of Philosophical Transactions.

  When we were going through the final manuscript, I told Sydney that I supposed he and I would be the only people in the world who would ever read through it carefully. For fun we decided to add a fake reference, so at one point we put “Leonardo da Vinci (personal communication)” and submitted it to the Royal Society. One (unknown) referee passed it without comment, but we had a phone call from Bill Hayes, the other referee, who said, “Who’s this young Italian working in your lab?” so reluctantly we had to take it out.

  The demonstration by genetic methods that the code was a triplet code was a tour de force, but in only a short time it was established by direct biochemical methods. Of more importance in the long run was the demonstration that acridine mutants caused small deletions and omissions. Even this was not unsuspected, since Leonard Lerman had produced very suggestive physical chemical evidence that acridines slipped in between the bases of DNA, and this could easily lead to additions or deletions of DNA when it was copied. Moreover, the theory had to be firmly established by direct biochemical methods. Both the biochemists Bill Dreyer and George Streisinger planned to do this though they were somewhat slow in getting the answer—at that time it was technically difficult to do the biochemistry. Each month or so Sydney and I would debate whether we should tackle it ourselves but we were reluctant to do this, especially as George was an “old boy"—meaning he had spent some time in our lab. Eventually George got the answer, working not on the unknown products of the two genes but on phage lysozyme. It came out exactly as we expected. In between the mutants a string of amino acids was indeed altered, and moreover, they fit well with what was known of the genetic code, which was just coming out.

  A little later I was at a meeting at the Villa Serbelloni on Lake Como, organized by the biologist Conrad Waddington (always called Wad by his friends). There for the first time I met the mathematician Rene Thorn. Almost the first thing he told me was that our work on the acridine mutants must be wrong. As I had just heard that our ideas had been confirmed biochemically, I was somewhat surprised and asked him why he thought so. He explained that if one made, say, a triple mutant, one necessarily got a Poisson distribution of single, double, quadruple, and so on, and so our arguments were not sound. Since we had laboriously put together our multiple mutants (and tested each carefully), I saw immediately that his objection had no force, being based on a misunderstanding. Either he had not read our paper carefully enough or, if he had read it, he had not understood it. But then in my experience most mathematicians are intellectually lazy and especially dislike reading experimental papers.

  My impression of Rene Thorn was of a good mathematician but a somewhat arrogant one, who disliked having to explain his ideas in terms nonmathematicians could understand. Fortunately another topologist, Christopher Zeeman, also at the meeting, was exceptionally good at putting over Thorn’s ideas.

  My other impression was that Thorn really understood very little about how science was done. What he did understand he didn’t like, and referred to it disparagingly as “Anglo-Saxon.” He seemed to me to have very strong biological intuitions but unfortunately of negative sign. I suspected that any biological idea he might have would probably be wrong.

  13

  Conclusions

  THE TIME HAS COME to try to pull all the threads together. In the episodes sketched earlier I have tried to suggest some aspects of biological research, both to illustrate its special character and also, by the way, to paint in a few glimpses of research as a human activity.

  What gives biological research its special flavor is the long-continued operation of natural selection. Every organism, every cell, and all the larger biochemical molecules are the end result of a long intricate process, often stretching back several billion years. This makes biology a very different kind of subject from physics. Physics, either in its more basic forms, such as the study of the fundamental particles and their interactions, or in its more applied branches, such as geophysics or astronomy, is very different from biology. It is true that in the latter two branches we have to deal with changes over comparable periods of time and what we see may be the end result of a long historical process. The layers upon layers of rock exposed in the Grand Canyon would be an example. However, while stars may “evolve,” they do not evolve by natural selection. Outside biology, we do not see the process of exact geometrical replication, which, together with the replication of mutants, leads to rare events becoming common. Even if we may occasionally glimpse an approximation to such a process, it certainly does not happen over and over again, till complexity is added to complexity.

  Another key feature of biology is the existence of many identical examples of complex structures. Of course, many stars must be broadly similar to each other. Many crystals in geological rocks must have a basically similar structure. But in neither case do we find masses of stars or crystals that are identical in many small details. One type of protein molecule, on the other hand, usually exists in many absolutely identical copies. If this were produced by chance alone, without the aid of natural selection, it would be regarded as almost infinitely improbable.

  Physics is also different because its results can be expressed in powerful, deep, and often counterintuitive general laws. There is really nothing in biology that corresponds to special and general relativity, or quantum electrodynamics, or even such simple conservation laws as those of Newtonian mechanics: the conservation of energy, of momentum, and of angular momentum. Biology has its “laws,” such as those of Mendelian genetics, but they are often only rather broad generalizations, with significant exceptions to them. The laws of physics, it is believed, are the same everywhere in the universe. This is unlikely to be true of biology. We have no idea how similar extraterrestrial biology (if it exists) is to our own. We may certainly consider it likely that it too will be governed by natural selection, or something rather like it, but even this is only a plausible guess.

  What is found in biology is mechanisms, mechanisms built with chemical components and that are often modified by other, later, mechanisms added to the earlier ones. While Occam’s razor is a useful tool in the physical sciences, it can be a very dangerous implement in biology. It is thus very rash to use simplicity and elegance as a guide in biological research. While DNA could be claimed to be both simple and elegant, it must be remembered that DNA almost certainly originated fairly close to the origin of life when things were necessarily simple or they could not have got going.

  Biologists must constantly keep in mind that what they see was not designed, but rather evolved. It might be thought, therefore, that evolutionary arguments would play a large part in guiding biological research, but this is far from the case. It is difficult enough to study what is happening now. To try to figure out exactly what happened in evol
ution is even more difficult. Thus evolutionary arguments can usefully be used as hints to suggest possible lines of research, but it is highly dangerous to trust them too much. It is all too easy to make mistaken inferences unless the process involved is already very well understood.

  All this may make it very difficult for physicists to adapt to most biological research. Physicists are all too apt to look for the wrong sorts of generalizations, to concoct theoretical models that are too neat, too powerful, and too clean. Not surprisingly, these seldom fit well with the data. To produce a really good biological theory one must try to see through the clutter produced by evolution to the basic mechanisms lying beneath them, realizing that they are likely to be overlaid by other, secondary mechanisms. What seems to physicists to be a hopelessly complicated process may have been what nature found simplest, because nature could only build on what was already there.

  The genetic code is a very good example of what I mean. Who could possibly invent such a complex allocation of the sixty-four triplets (see appendix B)? Surely the comma-free code (page 99) was all that a theory should be. An elegant solution based on very simple assumptions—yet completely wrong. Even so, there is a simplicity of a sort in the genetic code. The codons all have just three bases. The Morse code, by contrast, has symbols of different lengths, the shorter ones coding the more frequent letters. This allows the code to be more efficient, but such a property may have been too difficult for nature to evolve at that early time. Arguments about “efficiency” are thus almost always to be mistrusted in biology since we don’t know the exact problems faced by myriads of organisms in evolution. And without knowing that, how can we decide what form of efficiency paid off?

  There is a more general lesson to be drawn from the example of the genetic code. This is that, in biology, some problems are not suitable or not ripe for a theoretical attack for two broad reasons. The first I have already sketched—the current mechanisms may be partly the result of historical accident. The other is that the “computations” involved may be exceedingly complicated. This appears to be true of the protein-folding problem.

  Nature performs these folding “calculations” effortlessly, accurately, and in parallel, a combination we cannot hope to imitate exactly. Moreover, evolution will have found good strategies for exploring many of the possible structures in such a way that shortcuts can be taken on the paths to the correct fold. The final structure is a delicate balance between two large numbers, the energy of attraction between the atoms, and the energy of repulsion. Each of these is very difficult to calculate accurately, yet to estimate the free energy of any possible structure we have to estimate their difference. The fact that it usually happens in aqueous solution, so that we have to allow for the many water molecules bordering the protein, makes the problem even more difficult.

  These difficulties do not mean we should not look for the broad principles involved (for example, a protein that exists in aqueous solution folds to keep many of its water-hating side groups out of contact with the water), but it does mean that it may be better to try to go around such problems and not try to tackle them head on at too early a stage.

  A number of other lessons can be drawn from the history of molecular biology, though it would be easy to find examples in other branches of science as well. It is astonishing how one simple incorrect idea can envelop the subject in a dense fog. My mistake in thinking that each of the bases of DNA existed in at least two different forms is one such case. Another, more dramatic in some ways, was the assumption that the ribosomal RNA was the messenger RNA. And yet see how plausible this mistaken idea was. Jean Brachet, the embryologist, had shown that cells with a high rate of protein synthesis had large amounts of RNA in their cytoplasm. Sydney and I knew there had to be a messenger to convey the genetic message of each gene from the DNA in the nucleus to the ribosomes in the cytoplasm, and we assumed that this had to be RNA. In this we were right. Who would have been so bold as to say that the RNA we saw there was not the messenger but that the messenger was another kind of RNA, as yet undetected, turning over rapidly and thus probably there in small amounts? Only the gradual accumulation of experimental facts that appeared to contradict our base idea could jolt us out of our preconception. Yet we were acutely aware that something was wrong and were continually trying to find out what it was. It was this dissatisfaction with our ideas that made it possible for us to spot where the mistake was. If we had not been so conscientious in dwelling on these contradictions we should never have seen the answer. Eventually, of course, someone else would have spotted it, but the subject would have advanced less rapidly—and we would have looked very silly.

  It is not easy to convey, unless one has experienced it, the dramatic feeling of sudden enlightenment that floods the mind when the right idea finally clicks into place. One immediately sees how many previously puzzling facts are neatly explained by the new hypothesis. One could kick oneself for not having the idea earlier, it now seems so obvious. Yet before, everything was in a fog. Often it becomes clear that to prove the new idea a different sort of experiment is needed. Sometimes these experiments can be carried out in a remarkably short time and, if successful, serve to put the hypothesis beyond reasonable doubt. On such occasions one can go from muddled puzzlement to virtual certainty in the space of a year or even less.

  I have discussed earlier (in chapter 10) the importance of general, negative hypotheses (if one can find good ones), the mistake of mixing up a process with the rather different mechanisms that control it, and especially the importance of not mistaking a minor, subsidiary process for the main mechanism one is interested in. However, the principal error I see in most current theoretical work is that of imagining that a theory is really a good model for a particular natural mechanism rather than being merely a demonstration—a “don’t worry” theory. Theorists almost always become too fond of their own ideas, often simply by living with them for so long. It is difficult to believe that one’s cherished theory, which really works rather nicely in some respects, may be completely false.

  The basic trouble is that nature is so complex that many quite different theories can go some way to explaining the results. If elegance and simplicity are, in biology, dangerous guides to the correct answer, what constraints can be used as a guide through the jungle of possible theories? It seems to me that the only really useful constraints are contained in the experimental evidence. Even this information is not without its hazards since, as we have seen, experimental facts are often misleading or even plain wrong. It is thus not sufficient to have a rough acquaintance with the experimental evidence, but rather a deep and critical knowledge of many different types of evidence is required, since one never knows what type of fact is likely to give the game away.

  It seems to me that very few theoretical biologists adopt this approach. When confronted with what appears to be a difficulty, they usually prefer to tinker with their theory rather than seeking for some crucial test. One should ask: What is the essence of the type of theory I have constructed, and how can that be tested? even if it requires some new experimental method to do so.

  Theorists in biology should realize that it is extremely unlikely that they will produce a useful theory (as opposed to a mere demonstration) just by having a bright idea distantly related to what they imagine to be the facts. Even more unlikely is that they will produce a good theory at their first attempt. It is amateurs who have one big bright beautiful idea that they can never abandon. Professionals know that they have to produce theory after theory before they are likely to hit the jackpot. The very process of abandoning one theory for another gives them a degree of critical detachment that is almost essential if they are to succeed.

  The job of theorists, especially in biology, is to suggest new experiments. A good theory makes not only predictions, but surprising predictions that then turn out to be true. (If its predictions appear obvious to experimentalists, why would they need a theory?) Theorists will often complain that experimentalist
s ignore their work. Let a theorist produce just one theory of the type sketched above and the world will jump to the conclusion (not always true) that he has special insight into difficult problems. He may then be embarrassed by the flood of problems he is asked to tackle by those very experimentalists who previously ignored him. If this book helps anyone to produce good biological theories, it will have performed one of its main functions.

  14

  Epilogue: My Later Years

  IN JUNE 1966 the annual meeting at Cold Spring Harbor Laboratory was on the genetic code. It marked the end of classical molecular biology, because the detailed delineation of the genetic code—the little dictionary—showed that, in outline, the basic ideas of molecular biology were largely correct. It was remarkable to me and to most other people, both in the field and outside it, that we should have reached this point so quickly. When I started biological research in 19471 had no suspicion that all the major questions that interested me—What is a gene made of? How is it replicated? How is it turned on and off? What does it do?—would be answered within my own scientific lifetime. I had selected a topic, or series of topics, that I had assumed would last out my active scientific career, and now I found myself with most of my ambitions satisfied.