The good fairy godmother of science.

Abstract

There is a story that Einstein did not pursue a career in mathematics, not because he was not good enough, but because he did not have the ‘nose’ for recognizing the important problems. It is just as creative to define scientific problems as it is to solve them. As Peter Medawar put it, “Science is the Art of the Soluble”. So the trick is to define those problems that can be solved and to know what questions to ask. A fundamental change in biology occurred, for example, when the questions about proteins changed from asking where the energy to synthesize them comes from, to where the information comes from for putting the amino acids in the right order. Of course, this required Sanger's discovery that there is a right order.With the powerful techniques of molecular biology, genetics, electrophysiology, imaging and so on that are available today, there may be a tendency for science to be technology-driven: the techniques dictate the next experiment. This is not necessarily a bad thing. In my own field, developmental biology, I have argued that the principles are understood and all that remains is to fill in the details of how the thousands of genes controlling development work [1xDo we understand development?. Wolpert, L. Science. 1994; 266: 571–572Crossref | PubMedSee all References][1]. But am I correct? Perhaps there are many new principles to be discovered and I have not framed the questions correctly.But the more general issue is: what are the really interesting questions in biology that need to be answered? If the good fairy godmother of science (GOOFGOOS) allowed one to ask just one question to which she would give the correct answer, what would one ask? There are limitations — one cannot ask a very general question, such as: how does the brain work? It has to be more specific, and should be answerable in no more than 30 words. The nice thing about asking the GOOFGOOS is that she does not demand that we know how to solve the problem, or have the techniques to do it.The question I would like to ask the GOOFGOOS is this: is the egg computable? In other words, is there some computation that would enable us to predict the outcome of embryonic development? And if that were not possible, would it at least be feasible to simulate the whole process and so arrive at the outcome? Such a simulation might require the behaviour of every cell to be modelled. If so, how much detail in each cell needs to be included? And how ‘sharply’ must the initial conditions be defined? I await the answer with little optimism.I am greedy. I want to ask one other question. How is the left–right asymmetry of the embryo set up? We think it is based on an asymmetric molecule that becomes aligned with respect to the anteroposterior and dorsoventral axes and then provides a left–right bias [2xThe development of handedness in left/right asymmetry. Brown, NA and Wolpert, L. Development. 1990; 109: 1–9PubMedSee all References][2]. Are we right?