Forward, Toward the Epigenetic Future! Column in KomputerraOnline #87

Dmytro Shabanov

Ontological Adaptationism

Column in KompyuterraOnline #86 Column in KompyuterraOnline #87 Column in KompyuterraOnline #88

In my previous column I boasted that Alexander Pavlovich Rasnitsyn, a researcher with whom I had some organizational involvement, participated in a conference at the Kharkiv Faculty of Biology. I recounted some ideas from his address, delivered at the opening of the conference. Now I would like to discuss the epigenetic theory of evolution, which became the subject of the roundtable conducted by Alexander Pavlovich. [IMG_1] This segment of the roundtable with A.P. Rasnitsyn at V.N. Karazin Kharkiv National University on 21 November 2012 opens with a characterization of the epigenetic theory of evolution. A disclaimer is necessary here (as has been the case more than once in my columns). Evolutionary biology is under attack from propagandists who attempt to substitute dogma for science. All manner of "scientific creationists" (even their self-designation is a contradictio in adjecto — a contradiction in terms, or, in other words, an oxymoron: science is incompatible with the acceptance of religious dogma) are very fond of citing disputes among proponents of different theories of evolution. What I write about here has no bearing whatsoever on attempts to refute the well-established fact of evolution; rather, it concerns debates about the insufficiently understood mechanisms of this process. I am speaking of an evolutionary theory that its principal creator, Mikhail Alexandrovich Shishkin, named the epigenetic theory of evolution of I.I. Schmalhausen and C.H. Waddington. It should be said that in the works of Schmalhausen and Waddington one can find ideas central to the epigenetic theory of evolution, while at the same time those same works contain much that is incompatible with it. Here I will simply discuss ETE (the epigenetic theory of evolution), contrasting it with STE (the synthetic theory, which would more accurately be called the population-genetic theory, since there can be many different syntheses). If you wish to become better acquainted with ETE, read the articles tagged "Epigenetic classics..." here. That collection includes not only papers by Shishkin but also key ETE works by A.S. Rautian and A.P. Rasnitsyn. — The epigenetic theory rests on several well-established facts and generalizations. Let me attempt to enumerate some of them. — Natural selection (differential chances of survival and reproduction) selects or eliminates not individual traits but organisms as wholes. — Complex organisms are the product of individual development — ontogenesis. — Ontogenesis is influenced not only by the hereditary dispositions of the organism but also, to a significant degree, by environmental inputs; consequently, what is inherited is not the state of a character but the norm of reaction — the set of its possible states, realized under different environmental conditions with varying probability. — The overwhelming majority of genes affect not a single organismal character but many; the overwhelming majority of characters depend not on a single gene but on many (in the limiting case, on the genotype as a whole). — Organismal variability is not infinitely plastic; analysis of the diversity of individuals in nature allows one to assign them to a finite number of normal variants and some finite number of deviant types (aberrations). — The genotype is not a blueprint of the organism; genes contain "recipes" for proteins that perform specific functions, together with a set of switches that determine which of the possible developmental pathways will be chosen by the organism with greater probability. — The same changes in ontogenesis (and the characters arising from them) can be induced by either genetic peculiarities or environmental influences. — During ontogenesis, the most robustly realized state of a character is the one corresponding to the norm; genetic analysis of outwardly normal individuals from natural populations reveals their unexpectedly high genetic diversity. [IMG_2] Having compiled this list, I observe that accepting these propositions impels one toward a quite specific conception of the evolutionary process — toward agreement with the epigenetic theory of evolution. Alas, this is far from obvious to everyone. Moreover, biologists are divided into a minority that readily accepts this view and a majority that either fails to understand it or is unwilling to accept it. To accept it, one must rethink what all of us were taught — some in school, others at university. What consequences follow from the premises I have enumerated? Any developing organism possesses a definite set of possible developmental pathways. Selection may maintain one or several norms. Such selection causes normal development to become increasingly robust; in the case of several norms, genetic switching systems between these pathways evolve (typical Mendelian genes). When the ecological situation changes, selection begins to operate against the former norms. Organismal development is destabilized, and diverse aberrations begin to arise. If certain aberrations are supported by selection under the new conditions, selection begins to stabilize their development in turn, making them the new norms. In this way, evolution proves to be an alternation of stabilization and destabilization of specific developmental pathways by selection. Let me attempt to formulate the alternative viewpoint. I should state immediately that it will turn out somewhat caricatured. Not all proponents of STE think this way — only the most orthodox among them. Alas, even those proponents of STE who would contest the assertions I cited are compelled to rely on them implicitly in their conceptions of the mechanism of evolution. Without leaning on these, in fact easily refutable, assertions, STE does not work. The properties of an organism are determined by its genes. Each allelic state of a gene corresponds to a specific character. Some alleles are beneficial, others deleterious. New alleles arise as the result of spontaneous and undirected changes to existing ones — mutations. If a mutation significantly alters the genetic "text," a new allele arises with its own beneficial or deleterious phenotypic expression. Selection removes carriers of deleterious characters while preserving carriers of beneficial ones. Although individuals in natural populations differ from one another in many genes and many characters, the action of selection gradually increases the frequency of beneficial alleles in the population and thereby ensures evolutionary progress. Are you familiar with such ideas? Can it be that they even feel habitual? What should one call them? One apt designation is "beanbag genetics" (which plays on the fact that STE treats each character individually, like separate peas shuffled in a bag). It is almost a pity: such an elegant picture — and an incorrect one. Two circumstances are particularly devastating to it. The effect of each allele depends on the genetic background in which it finds itself and on the conditions under which the organism develops. There is no direct correspondence that would allow one to determine an organism's genes from its characters — the same favorable variants favored by selection can develop on the most varied genetic bases, or simply as the result of specific environmental action. The more complex the organism, the more complex the connections influencing the development of its characters. I have described the increasing complexity of the gene-character relationship in considerable detail in one of my earlier columns. Do you think that is all? There is yet another problem. Even with a one-to-one correspondence between genes and the characters reflecting them, selection operating simultaneously on multiple genes proves extremely inefficient (a point touched upon in yet another old column). This last assertion is often called "Haldane's dilemma." This issue was first articulated in the paper "The Cost of Natural Selection," published in 1957 by John Haldane (1892–1964), one of the architects of STE. I disagree with the formulation of Haldane's dilemma offered by Wikipedia; its essence lies precisely in the inefficiency of selection operating simultaneously on many genes, not in any particular assessment of the rate of evolution. Unlike the author of the Wikipedia article in question, I (following Rasnitsyn) am astonished not by how rapidly evolution proceeds but by how slowly it typically proceeds (relative to the examples that demonstrate the rates of change available to it). One could go on comparing STE and ETE at length. I will offer just one further detail. For STE, heredity is a property of living matter whose origin is inexplicable. First, genes arose that were capable of determining characters; then, selection among characters led to the evolution of genes. For ETE, heredity is the result of selection. Systems capable of copying (reproduction) proved successful if, in place of units of selection that met environmental demands, entities similar to them appeared. Heredity is a consequence of selection for the reproduction of a norm preserved by selection. Why, then, does ETE remain the province of an intellectual minority of biologists while the overwhelming majority continue to rely by default on STE? The reasons are numerous. STE is built on linear cause-and-effect chains, for the analysis of which our cognitive architecture is well suited. The task of analyzing the properties of the developmental control system of a complex organism is far more challenging than the three-body problem, which I discussed in the column I linked to in the previous sentence. The roundtable addressed some of these matters. Naturally, Alexander Pavlovich emphasized the problem to which his contribution to the development of ETE is connected — the problem of adaptive compromise as an evolutionary trap (which I discussed in my previous column). In an ecologically stable situation, aberrant individuals have virtually no chance of survival: they will almost inevitably be inferior to the compromise norm in some important respect. Herein lies the root cause of the stability of many species (which would otherwise be entirely puzzling, especially in species without sexual reproduction). Aberrant individuals will acquire the chance to survive only if the character of natural selection changes substantially and the former demands on their organization are lifted. Thus, the distinctive feature of ETE is its attention to the organism as an integral system viewed phenomenologically rather than analytically. This approach is diametrically opposed to the typical reductionist logic of molecular biology. Do you know how difficult it is for molecular biologists to appreciate ETE? This was evident even at the roundtable. I had the pleasure of observing how specialists in epigenetic inheritance listened to Rasnitsyn. In this context, the word "epigenetic" bears no relation to ETE: it refers to a mechanism of gene expression regulation through "marking" of the DNA chain, achieved by chemical modification of individual nucleotides. And so the reductionist molecular biologists were genuinely suffering! Rasnitsyn's arguments must have seemed to them like dreadful metaphysics, so unlike three-dimensional reconstructions of macromolecules. This did not, however, prevent the most expressively suffering among them from going to be photographed together with Alexander Pavlovich after the event concluded. [IMG_3] Listening to the exposition of unfamiliar ideas is difficult, but being photographed with a classic is interesting. How can ETE be made intelligible and acceptable to the broad biological community? It seems to me that it must be subjected to maximal formalization and expressed in models. The possibility of formalization is a test of the logical consistency and non-contradictoriness of any system of views. STE has been partially formalized: from the outset it rested on the mathematical theory of selection developed by brilliant mathematical biologists — Ronald Fisher (1890–1962), Sewall Wright (1889–1988), and the already-mentioned John Haldane. The models they created worked only for a limited set of simple cases, but for those cases they worked well. For complex, evolutionarily significant cases, it has not been possible to formalize STE, not least because of Haldane's dilemma. In my assessment, this is a manifestation of the deep internal contradictoriness of STE. It seems to me that in the situation of choice between the norm and aberrations, it may be possible to escape the constraints that Haldane's dilemma imposes on the course of evolution. Alas, I have not yet been able to understand how to formalize this intuition. And why has the logic of ETE not yet been captured in comprehensible models? The reasons are numerous. First, this is an extremely difficult task. Second, as Alexander Pavlovich emphasized, the number of ETE proponents who genuinely understand its essence is regrettably small. Third, the time has simply not yet come: we are not yet skilled at describing the properties of complex systems and may not yet know all the mechanisms of information transmission during ontogenesis. There is much work to be done.