Review of : 2011 . Evolution, Development & the Predictable Genome . Roberts and Co. Publishers , Greenwood Village , CO . ISBN : 9781936221011 . $45.00
When developmental biology reached the level of maturity and the technical capabilities to return to questions of organic evolution, after nearly a century of separation from evolutionary questions, the event was greeted with much anticipation. In the air was the expectation of a new level of synthetic understanding of evolution. Sadly, though, the meeting between evolutionary and developmental biologists was not a meeting of like minds. To the evolutionary biologists, the arrival of developmental types in their territory felt more like an invasion by space aliens with their transgenic monstrosities and glowing embryos. What exactly do we learn from a transgenic mouse without a head and flies with four wings? And the developmental biologist who ventured out of his/her homeland of experimental biology to evolution felt like he or she is visiting some strange tribe that practices a form a mathematical Voodoo with counting flies and bristles. How do the equations of population genetics relate to the molecular reality of heredity? There was little that the two groups really understood about each other, leading to skeptical voices, like that of Ron Amundson (2005), who questioned whether the chasm between evolutionary and developmental biology will ever be bridged. This of course is a cartoon, and does not do justice to the amount of positive work done by many at the interface of developmental and evolutionary biology. But nevertheless, at a deeper conceptual level, the rapprochement was and is still difficult. Is a new synthesis really necessary as, for instance, Massimo Pigliucci argues (2007), or is it perhaps even impossible, as Amundson suggests?
It seems obvious that a true integration of developmental and evolutionary biology will have to come from scientists equally at home in population genetic theory as in the experimental culture of developmental biology. Unfortunately for our science, the number of people so trained is still small, but David Stern in certainly one of them. Stern was trained as an evolutionary behavioral ecologist but then got postdoctoral training as a developmental Drosophila geneticist. Hence it is a good idea to listen closely when Stern is summarizing his thinking about a developmental perspective on evolution. To cut a long story short, Stern's approach is based on two key insights. The first comes from population genetics and concerns the factors that determine the chance that a mutation will go to fixation. Those factors include population size and structure, as well as genetic factors, such as pleiotropy, dominance, and epistasis. The less pleiotropic and epistatic a mutation and the more dominant it is, the higher the chance of it reaching fixation. These latter factors are not intrinsic to the mutant gene itself, but instead depend on the position of the gene in the gene regulatory network. Stern's central claim is that the likelihood of successful mutations to occur in a gene can be read off the structure of the gene regulatory network. This premise is motivated by the empirical fact that some genes are more likely to be the target of evolutionary change than others. For instance, the shavenbaby locus in Drosophila is more likely than any other gene related to trichome development to be changed in the evolution of this phenotype. The Pitx1 gene in stickleback also comes to mind (Chan et al. 2010). Similarly, the gene Fridgida is a privileged target of evolution of the vernalization pathway of Arabidopsis. These genes are, in Stern's language, evolutionary hot spots, and whether a gene can be a hotspot is determined, according to Stern, by the structure of the gene regulatory network. In order to facilitate the analysis of complex gene regulatory networks Stern proposes a device that he calls “pathwork.” A pathwork is that part of the gene regulatory network that acts within a certain cell to achieve a certain phenotype. In such a simplified network, it is easy to identify bottlenecks in the information flow that, according to Stern, are the foci of evolutionary change. The poster child of this approach is the trichome development pathway in Drosophila larvae. Stern's lab has made important contributions to the developmental evolution of this phenotype. To his credit, another example of an evolutionary hotspot, that of the Arabidopsis Frigida gene, is also presented, but it is not explained by the method of pathworks.
Maybe it is worth reflecting at this point, given that the perspective works in some cases, but, as Stern also shows, not in all cases, what the potential reach of this model is. In the case of the trichome cell, the phenotype is determined by whether a cell decides to assume the fate of a trichome cell or not. This is a qualitative decision between two possible cell types. I agree that it is plausible that, for these kinds of phenotypic differences, bottleneck genes are the most likely targets of evolutionary change. The question, though, is whether all evolutionary change will be of that kind. Clearly adaptive change includes a lot of continuous quantitative phenotypic change, and it is not clear whether genetic hotspots are also responsible for these changes. There is also the question of how novel cell types arise. Will the hotspot idea also help us understand what is going on in these cases? Or will we find an analogous principle that links developmental mechanisms to evolutionary mechanisms? I certainly hope so! That it was possible for Stern to make this step forward gives hope to rest of us that more will be coming and a deeper understanding of other kinds of evolutionary change will be arriving, may be even soon.
In what way does the perspective of evolutionary hotspots and pathworks advance the conceptual development of evolutionary biology? One theme that Stern is pushing in the title of the book is the claim that genomic evolution is predictable from the structure of the genome. I do not find that particularly convincing. One may note that there is no context in which the method of pathworks can effectively make “predictions.” Maybe experimental evolution is a possible approach, one that will require large investments. But in all the examples given in the book, the perspective of pathworks is an explanatory framework for already known patterns of evolutionary change. Some philosophers of science, though, question the logical distinction between “true” predictions and “retrodictions” and even question whether true predictions are more important than the ability of a theory to accommodate a large amount of knowledge, e.g., Scerri (2007). It seems quite unlikely that predictions will be the most important research program resulting from these ideas.
There is however a more convincing and perhaps more important implication of Stern's ideas and that is the chance to put to rest one of the controversies that do not seem to be able to die, namely that of the role of small mutations rather than macromutations. The importance of Stern's proposal is that it changes the frame of reference for this debate. No longer is the question whether macromutations play a role and how small mutations ought to be if they are to be called small mutations. Instead, the perspective is one in which the empirical facts that inspired macromutationism are accommodated and even expected. That is, we accept as fact that there are genes that can mutate with large coordinated effects, like homeotic transformation of whole body parts. The existence of so-called macro-mutations reflects the fact that the developmental regulatory networks have bottlenecks, where a few genes control a large amount of information flow. Under certain circumstances, these bottlenecks can be evolutionary hotspots accumulating many mutations, each of them small and with low pleiotropy, so goes the theory. In the comparison between species, these loci then look like ones that can transform the phenotype of one species into that of another or even affect whole bodyplan features. In that way, we can accommodate the facts that inspired macromutation theory, but do not need to buy into the untenable implications of that theory, large mutational steps with likely extensive pleiotropic effects. Stern's perspective is also different from the classical neo-Darwinian model, which assumes that adaptive change is based on a large number of genes with small mutations, with no preferred loci of evolutionary change. This is a truly important and productive new start in understanding morphological evolution.
Finally, I want to say a few words about the style of the book. This is certainly a book that makes a serious scientific point, as outlined above, but it is not a scholarly book in terms of its style. Most of the chapters are an informal introduction into evolutionary biology and development, at a level that is targeting biology freshmen or the educated public. The intellectually exciting parts are two of the last three chapters, or more precisely chapter seven, called Pathworks, and the historical note at the end, in which Stern outlines the explanatory scope of his ideas. Hence the point of the book can easily be made in a review article and in fact has been made by Stern and Virginie Orgogozo in a February 2009 article in Science. References to the sources of data are hidden in an appendix that one has to find by accident. I think I understand what Stern wanted to achieve, namely to make the book easier to read and accessible to a wider readership. But understanding this does not make me like the approach, as much as I like Stern's ideas and work. Yes, it may help sell more copies of the book, nothing wrong with that, but it sends a bad message to students, who, already spoiled by Google sourcing, are not given a sense of the foundation our science is resting on. A comparison with the recent books by Eric Davidson is informative (Davidson 2001; Davidson 2006). I may disagree on many issues with Davidson, but I like the style of his books much more, since I learn a lot of hard-core biology from them, even biological facts that Davidson may not want me to focus on: for instance, the evidence that transcription factors are not functionally equivalent between mouse and fly (Ranganayakulu et al. 1998) or that the famous Pax6/fly eye experiment is probably an artifact. From Stern's book, one only learns what Stern wants you to know. The desire to send simple messages has to have its limits. Einstein reportedly once said that in science we have to simplify things as much as possible, but not more. Nevertheless, this book is required (easy) reading for both, students as well as researchers in the field of developmental evolution, for years to come.