Moss - What genes cant do - 2003 (522929), страница 26

Просмтор этого файла доступен только зарегистрированным пользователям. Но у нас супер быстрая регистрация: достаточно только электронной почты!

Текст из файла (страница 26)

Kauffman thus shifts the locus of biological order and stability from that of solid-state bonds to the dynamics ofsystems whose state-space converges on attractors. That the latter canoccur without anything like a genetic code is given plausibility by thesimulations which Kauffman and Sante Fe Institute collaborators haveproduced. An independent group of investigators in Japan have likewiseused computer simulations of simple reaction-diffusion systems to showthat patterns resembling the heritable differentiation of cell types can bethe result of dynamic phase transitions in the absence of anything playingthe role of genes.

In addition, Kaneko and Yomo (1994) were able tocorrelate trends in their simulations with certain biological phenomenaobserved in bacterial cultures.What then is the relationship between DNA—i.e., the aperiodiccrystal—and steady-state dynamics? Kauffman emphasizes that theformer is neither necessary nor sufficient for the latter—which is to say100Chapter 3that homeostatic, autocatalytic systems with the capacity to mutate andbecome subject to natural selection are at least theoretically independent of the need for any noncatalytic, chemically inert template polymer(e.g., DNA).

Further, the presence of such a polymer cannot and doesnot specify (or guarantee) any particular dynamic regime. The presenceof DNA or any other aperiodic crystal, for example, cannot itself prohibit a slide into chaos. DNA can at best be considered a kind of “fellowtraveler” in the ship of life.Mutation, defined in the dynamic-systems context, refers to a fluctuation in the synthetic catalytic cycles resulting in the accumulation of anovel product.

Such a product may then become stably incorporated intothe cycles with an alteration in the catalytic dynamics. Such dynamicmutations were evidenced in simulation. Now clearly, biological entities,as we know them, consist of DNA templates (Genes-D) in the contextof steady-state dynamics so that attempts to distinguish between them,beyond just being arbitrary, run the danger of becoming a categorymistake. Genes(-D), as real biological effectors, are the result of dynamicprocessing on multiple levels (e.g., transcriptional regulation, transcriptprocessing, transcript transport, and translational regulation, etc.), so thevery concept of the gene brings with it dynamic presuppositions.

Andyet the challenge of how to simultaneously cognize sequence-based anddynamic aspects together persists. Just as in the case of our distinguishing heritable structural features of the cell, the sense of referring tosteady-state dynamics as an “epigenetic inheritance system” can properly be meant only to analytically distinguish concurrent and mutuallydependent aspects of integrated living systems.Kauffman attempts to elucidate this relationship through a theoreticalsimplification that appears to reap certain rewards. Following theconsiderations touched on above, pertaining to the number of types ofmolecules and the number of types of reactions, Kauffman centers onthe ramifications of these two parameters for self-organization.

In thefirst approximation, N represents the number of components of a livingsystem and K the number of connections between them. The trick forKauffman will be to parse the cell in such a way as to result in valuesfor N and K that have interesting implications. To ask how many com-A Critique of Pure (Genetic) Information101ponents there are in a cell is akin to Wittgenstein asking how many thingsthere are in a room. It entirely depends on how you divvy up the room—and thus decide what is going to count as a “thing.” And then depending on what is thus counted as N will follow how many interactions thereare between the elements of N, i.e., K.

Now if each component has onlytwo states—i.e., can be counted as a binary variable, but receives inputsfrom a K = N number of factors—then even where N = 200 the numberof possible states of the system would be 2200 which is so sufficientlylarge that even if the state changed only every 1 to 10 minutes, it wouldstill require more than the age of the universe for the system to sampleevery state.Such an attractor cycle could not undergo selection, for obviousreasons. Kauffman elects to limit his world of N-relevant componentsto genes, settling upon 100,000 as a plausible number.

As we’ve said, ifK = N = 100,000, i.e., every gene effects every other gene, then thenumber of possible states would be far off of any relevant scale, and thesuccession of states would be random and chaotic. Having selected genesas the basis of his N parameter, Kauffman can entertain the abstractionof an organism as a Boolean network with 100,000 binary variables.Order emerges from such a system if the value for K is 2 or slightly less.With K = 2 the system is highly interconnected and complex, such thatthe state of variables is far from independent of one another, yet itis simple enough for discrete patterns to emerge as opposed to beingcondemned to a wholly unwieldy chaos. In the Boolean model, theactivation state of each gene can be computed on the basis of one of 16possible Boolean functions (randomly assigned)—e.g., AND, OR, If, andso forth—as well as input received from (K =) two sources.

Now givensuch an approximation, Kauffman finds the very satisfying result(through simulations) that the number of states which make up the statecycle approaches the square root of the number of variables. So, with100,000 variables the predicted number of states in a cycle would be onthe order of 317, which is roughly the number of differentiated cell typesthat cell biologists have distinguished in humans. The convergence of thestate-space of this hypothetical system with 100,000 variables down to317 possibilities is what Kauffman refers to as “getting order for free.”102Chapter 3The pertinent question, however, will be if and to what extent the parameters that Kauffman has chosen can be understood to be of real biological relevance.Kauffman’s theoretical assimilation of the cell-organism to that of aparallel-processing genetic regulatory system has much of the characterof a Faustian bargain.

The idea that the biology of a living cell is morethan just a series of linear reactions—more systematic, interconnected,and complicated than even a compilation of ever so many RubeGoldberg schematics—is not so much controversial among biologists asit is seemingly intractable. Kauffman’s model provides the rare handlefor conceptualizing biological processes at a higher level of complexity.It brings with it a power evidenced in Kauffman’s ability to serve upexplanatory accounts of phenomena ranging from the origins of life toontogenetic differentiation and morphogenesis (Kauffman 1993). But atwhat cost? In modeling the cell-organism as a complex genetic regulatory system, Kauffman, somewhat ironically, contributes to the attemptsto make the real (extragenomic) complexity of life disappear.

Now surelyKauffman’s model is militantly anti-genetic-reductionist in the sense thatits basic unit is not a gene but rather a cell-state defined in terms of whichgenes are turned on and which are turned off. While the activation stateof a cell’s genome is considerably less reductionist than merely a focuson individual genes, the identification of a cell’s phenotype with the stateof its genome—and by extension the identification of the phenotype ofan organism with the genetic regulatory state of all of its cells—is insidiously seductive and patently false.

While the activation state of a cell’sgenes is inseparable from its phenotype, it by no means uniquely determines a phenotype, for all the reasons discussed in section 2 above aswell as many more. A pattern of gene expression, no less than individual gene expression, is only meaningful in the context of multilayeredlevels of organization, structure, and dynamics which are in no wayreducible to patterns of gene expression.

Kauffman, in his own way, isparty to the kind of “slippage” discussed in chapter 2. For the sake ofgetting a grasp on a powerful theoretical engine, Kauffman has made thereal biology of the cell, and thus biology itself, disappear into the virtualinterstices lying between the connections of (genetic) nodes in a parallelprocessing simulation.A Critique of Pure (Genetic) Information103The irony that I refer to above is as follows.

Kauffman’s slippage, hiscontribution to the disappearance of biology in the name of the genome,is the product of what is perhaps the most sophisticated and wellelaborated challenge to the “hegemony” of the genetic code-script.Against the heirs of Schrödinger, Kauffman denies the role of a solidstate set of instructions and offers instead a model whose order emerges,not out of the stability of covalent bonds at all but out of the higherorder “logic” of complex dynamics at the edge of chaos. He has simplylatched onto the genome for lack of any other theoretically amenableledge to grasp onto amidst the biological maelstrom.While the central theme of this chapter has been that of epigeneticinheritance systems—that is, stable sources of biological order whichare inherited in parallel with the genomic sequence and codeterminitiveof phenotype—Kauffman’s understanding of steady-state dynamics, orreally the convergence of the state space of a complex system onto aseries of numerically, highly delimited attractor states, is not about aparallel system of inheritance.

Dynamics is the big picture for Kauffman,and if we are to think of dynamics as epigenetic, then in the Kauffmanpicture, epigenetics always comes first. But perhaps a better characterization would be to say that genetics and epigenetics simply merge intoone and the same. But a merging of the two is hardly unique in twentieth century biology, as chapter 1 revealed. Nor would we want to suggestthat the distinction between genetics and epigenetics should be accordedthe status of a natural kind (a natural distinction?).At the end of the first chapter, I too suggested that ultimately a theoryof epigenesis which allowed for a recontextualization of the gene quaGene-D would subsume the legacy of Mendelian genetics and, thus,Gene-P.

Has Kauffman provided this step into the future—subsumingthe legacy of Mendelism, not into a more “reduced theory” but ratherinto a more holistic and expansive theory? Certainly not. Just as theMendelian tradition found practical utility in proceeding as if genesdirectly determined phenotype, Kauffman has found theoretical utility inproceeding as if the cell were constituted by 100,000 binary units andas if each of these received input from only two other components.Kauffman’s model is itself a form of instrumental reductionism. Perhapswe should consider it a brand of “instrumental dynamicism” as opposed104Chapter 3to the instrumental preformationism of the Mendelian tradition.

Характеристики

Список файлов книги