I was recently sent a strange article for comment…well, not that recently. It's 51 pages long, so I've kind of dragged my heels over it all. I have finally finished it, though, and it is weird.
There is a significant tradition in developmental biology that is currently a bit out of fashion: some names for it are formalism or structuralism. It's the idea that forms arise from the application of simple physical properties to tissues, and that it would be possible to describe an organism in some way by a set of parameters plugged into a mathematical formula. It's an interesting idea, it has a long history with some significant proponents (Goethe, for instance), and I'd even argue that there's a solid kernel of truth to it — we should not ignore the importance of physico-chemical properties in shaping organisms! However, it's an idea that has been eclipsed by the successes of molecular biology and genetics, which has moved the focus of research away from the general and universal, the fluid properties of membranes, for example, towards the particular and discrete, the actions of genes. A gene-centered research program has proven powerful, while the idea of exploring how global properties influence form…not so much.
I confess to a fair bit of sympathy for the idea that physical and chemical forces are significant, and that our current emphases in biology don't do the concept justice. The best of the modern structuralists is, in my opinion, Brian Goodwin, and I find his work thought-provoking. No, more than that; he describes many processes that must be addressed from a structuralist position, and for which genes are inadequate descriptors. The giant in this field is, of course, D'Arcy Wentworth Thompson, whose classic work is On Growth and Form. It was written before genes, genes, genes took over biology, so it has a very different perspective from modern work; it's also written in a wonderfully formal style that I find sublime but some may find a bit too old-fashioned. It would be required reading for all biologists, if I were the tyrant of the sciences. You'll never look at a field of cells in quite the same way after reading it.
Unfortunately, structuralism also attracts more than its share of cranks. There's something about reducing all the complexity of a mouse to an array of fields and vectors and mathematical formulae that draws in a certain kind of mind. It also sucks in a certain kind of person who doesn't actually ever want to get his hands sticky and slimy poking around in a real live mouse, but wants the elegant purity of math — something he or she can do at a desk, where it's clean and simple. Simplicity is also important to these people: all that fussy, nit-picky complexity of thousands of genes clutters their impeccably sterile mathematics. This means that the field is littered with kooks.
I reviewed a book of that sort from Stuart Pivar. It had all the characteristics of this genre: a shocking ignorance of actual biology, a weird obsession with a single physical cause (in his case, all organisms were toroids, or as I pointed out, balloon animals), and a willingness to ignore the petty problems of biology actually contradicting his conclusions.
This latest addition to the structuralist literature is a long paper by Vincent Fleury, called "Clarifying tetrapod embryogenesis, a physicist's point of view". It is nowhere near as crazy as Pivar's work, so let's get that out of the way — Fleury has at least read the developmental biology literature. Boy, has he read it — the bulk of his paper is a huge survey of introductory developmental biology. Part of the pain of reading the paper is that almost all of it is completely irrelevant to his thesis, and reading it from the point of view of a developmental biologist, it was too much like reading a smart undergraduate's overlong term paper…a smart undergraduate who has decided that the appropriate subject is to attempt a complete overview of an entire field. This happens to me now and then; I try to catch it early, though, and steer the student into focusing on something manageable and specific. Too late for Fleury!
A voluminous overview of all of developmental biology (which, of course, fails — 50 pages is not enough!) is actually a distraction from making a point, unless perhaps that point is to flaunt some kind of misplaced erudition. So what is his point? Skip ahead to his last paragraph.
A possibility therefore appears, that the apparance[sic] of
tetrapods be generic, and that it follows a general law with few degrees of freedom, although it has lots of genetic parameters. In which case, Darwinian evolution plays with a very restricted set of shapes, with stringent internal (physical) correlations, and the known body forms might be unavoidable, in the long run.
This is not a very useful hypothesis. He's basically saying that all tetrapods are four-limbed because there are physical constraints that stem from the earliest processes in the embryo that impose a general form on them. It's rather tautological, for one thing, but it is also trivially true. I think all developmental biologists would agree that there are limits to the potential for morphological change in a lineage, but the question is the nature of the processes that impose the limits. Despite my sympathies for the structuralist idea, I'd have to say that what does that is the interlocking pattern of gene regulatory networks. Fleury wants to claim that it is hydrodynamics, and fluid flow, and shearing forces.
I don't doubt that fluid properties are important in cell behavior, but the paper has only two modes of instruction: a long and somewhat pedantic description of known developmental genes (which does not advance his thesis at all), and another long section of hydrodynamic theory (which is not applied to any experimental or observational work in biology). There is no synthesis! The biology section is like an exercise in which the author shows off that he has done some homework, and the physics part is to demonstrate that he has the math to apply to problems of problems of fluid dynamics, but the two don't meet. What I would want to see in order to be persuaded that this approach is viable would be a specific application of these principles of physics to a specific process in development, in either an experimental or comparative context. Show me how similarities and differences in form are better explained by differences in physical properties than by genetic differences (which then subsequently induce differences in physical properties).
The author is a physicist, and it shows. Not that I'm saying anything derogatory about physicists, but they have their own domain of expertise, while biologists have a different one, and I'm afraid you can't just blindly assume expertise in one translates into the other, or I'd be designing rocketships right now. So many little things in this paper clued me in to the superficiality of Fleury's knowledge of biology.
For example, and this is a very small thing that will grate on any biologist, is that he refers to single species as "specie". The singular of species is "species"; specie is money in the form of coins. The third time Fleury did this, it was driving me nuts.
Then there's this strangely indignant outburst.
True as well as extra limbs actually originate in the lateral
mesoderm of the flanks, in an area called "limb field" or
"lateral plate" or "limb plate". While preparing this
review, I found it impossible to find a description of where
the lateral plates come from. All existing work assumes an
already formed lateral plate for the extension of the limb, or for the early expression of limb markers (such as Tbx5).
His confusion is utterly baffling: a two minute conversation with a generic, minimally experienced embryologist, like me, could have settled this easily enough. Cells invaginate in the embryo to form a mesodermal layer, a relatively undifferentiated sheet that progressively separates into paraxial and lateral plate mesoderm on the basis of their distance from the embryonic axis. It comes from the same place as the other mesoderm; it's simply a sub-domain of the mesoderm that separates from the others.
It's this kind of thing throughout the paper that gives the impression that his knowledge of the subject is an inch deep, and further, that he isn't even willing to have a conversation with a biologist.
I'm very glad I wasn't asked to formally review this paper, because it would have taken ages and would have required so much commentary that it would have been like writing a 50 page paper myself. Or maybe not: I would have short-circuited the whole thing by simply rejecting it outright and recommending that it might be worth re-evaluating after the author cut out 45 pages and distilled it down to some discriminating tests of his hypothesis.
Strangely, or perhaps not so strangely, the reviewers of this work did not do that. It's a paper that was published in The European Physical Journal: Applied Physics! I am baffled again. Why it was accepted for a journal with such inappropriate content at all is a mystery, but at least we can presume that the reviewers were fellow physicists who were completely oblivious to the superficiality of the biology. Either that, or they were stunned by volume of basic biology shoveled up by the author, and signed off in surrender; anyone will confess to anything when waterboarded by that much introductory text.
All that aside, though, what ultimately kills the work is the maddening vagueness of its claims and implications. Here's the pocket summary provided by the author.
Reviewing the paleontological, genetic, developmental and
physical data, shows that a dynamic picture of embryo
development, resting on fundamental laws of physics, can
be proposed. In this view, development consists in a continuum deformation of an initial formless animal which
progressively changes shape by scaling up an initial symmetry breaking. Starting from such a symmetry breaking,
a uniform behaviour of cells (constitutive equation) suffices to induce a deterministic asymptotic form, by low
Reynolds flow. The dimensionless flow forms a general
law, whose parameters have a genetic origin. Modifying
the parameters shifts the animal forms along morphological diagrams which follow the streamlines of the flow. Although several aspects (phaners, metabolism, cognition,
etc.) add to the problem lots of features which we have
not addressed, there may exist a simple biomechanical rationale that explains in detail the global pattern of embryo
structure in vertebrates. The formation of a streak, the invagination of the yolk-sac, the back and forth motion of
Hensen's node, the formation of a body which has globally
the shape of "an 8", the lateral position of the limb fields,
are all very complex and important developmental questions, which in fact may be reduced to simple hyperbolic
viscous flows in the embryonic sheets. These flows require
an initial symmetry breaking and possibly genetic pools of
molecules which overlap mechanical fields. The rationale
of animal formation is that of a cellular flow which runs
away by producing its own factors of self-organization.
An equation suffices to define form? Show me. Show me this "simple biochemical rationale that explains in detail the global pattern of embryo
structure in vertebrates," because I don't believe it is there, and it certainly isn't shown in the paper. What hydrodynamic property makes a tetrapod different from an arthropod? Fleury doesn't say. What differences in cellular flow make a squirrel different from a mouse? Fleury doesn't say. Where are the measurements of forces in tissues in the gastrula that lead to the four-limbed state? Fleury doesn't give them. What does this formula provide to developmental biologists that is not matched or vastly exceeded by the tools of molecular biology? Where are the successful, productive laboratory experiments we can carry out to test and extend this model? Fleury has nothing.
Where Pivar had toruses and drawings that reminded me of ballon animals, Fleury seems to believe development proceeds by swirling forces spinning about in the embryo, and gives us fictitious diagrams like this one.
a drawing of an homo habilis skull, with a
pattern of presumptive stream lines superimposed. The head material rotates around the ear orifice (roughly), and above the
eye orbits.
Please show me cell movements or tensions in the developing skull that correspond in any way to those psychedelic lines. It's completely nonsensical.
Swirl this one right down the drain, please.
Fleury, V (2009) Clarifying tetrapod embryogenesis, a physicist's point of view. The European Physical Journal Applied Physics 45(3):30101.
Vincent Fleury has demanded the right to reply, and I'm always happy to help a fellow hang himself. Here's what he sent me:
your review on my article in EPJ (Clarifying tetrapod embryogenesis)
was pointed to me by embarrased colleagues, who are appalled by such incompetence (I mean yours).
I am sorry to say that you understand very little physics.
It is not a big problem in so far as you do not make blatant deffamatory statements.I am a bit choked by several statements which you make.
For example, I do give a hydrodynamic explanation of transiton from apes to humans.
I do give a hydrodynamic explanation of transition of lizards to snakes etc.
it is very easy from the examples I give to extrapolate to other cases.I do show how equations may generate forms, it is so simple (moving boundary problems).
Your statement about the "psychedelic" image are really pathetic for you, since apparently not only you do not understand what a deformation rate field is, you are unable to follow basic rationales in the article, but you do not seem to be aware of the structure of the stress field in young heads (cephalic vesicle).
I certainly do not want to enter into a useless argument with you, but as a compensation for all the harm that your incompetent review will do, I would kindly request that you put side by side with the "psychedelic" image the animation which you will find in attachment with the caption "Animation showing the progressive deformation of a skull in a dipolar field of stretch, the animation concatenates a dipolar winding towards the past, and towards the future, of the skull in the <<psychedelic>> image",
and a link to the pattern of shear lines in early brain vesicles :http://www.msc.univ-paris-diderot.fr/~vincent/englishthemecerveau2.html
That will suffice to satisfy me, and you will certainly agree that it is not a big demand; I shall not go any further with you.
With my best regards
Vincent Fleury
- Log in to post comments