The target article displays a thorough and impressive gathering of cellular phenomena that cannot be accounted for by “classical” genetics. Such phenomena are then rallied against the fact that “it has been a key dogma of molecular genetics that DNA is the sole biological agent of heritability, and it is still commonly treated as such.” Such state of affairs, the author intimates, implies a Kuhnian paradigm shift towards a “postgenomic view” which, nevertheless, Charney does “not call … a paradigm because it has not yet coalesced around a core set of principles or assumptions characteristic of a paradigm.”
Two lines of study are identified in the target article: heritability studies and gene association studies. Out of both these lines, a feeling oozes that the implicit debate underlying the contrast between dogmatic biology and the more permissive postgenomics has been replacing the now largely obsolete nature–nurture dispute. Günther Stent (Reference Stent1975), through his dissatisfaction with the unique agency of the genome and his assertion of an implicit role of the cell, anticipated the essential uneasiness of the actual epigenetics movement.
So what is it this further element, over and above the dogmatic centrality of the DNA, which is alleged by the epigenetics community? We think it is the computational nature of the cell.
This runs contrary to Charney's statement: “DNA does not contain a determinate genetic program (analogous to the digital code of a computer) from which we can predict phenotype.” And it might well be true that the phenotype is not predictable, but rather more because generally program behavior is undecidable (Davis Reference Davis2004), hence unpredictable, than because the genome is a program, or rather a computational structure.
There is a lack of appreciation that the computational and information theoretic talk about the cell, current since the very beginning of molecular biology, is not metaphoric but literal, and that cellular processes are primarily computational rather than biochemical ones, in the same sense that the CPU chip of a PC is primarily a computational apparatus rather than a solid-state physics device.
As stressed ubiquitously in the epigenetics domain, the gene does not express the protein by itself. The author states as much in Note 6, and indeed, the fact that the strings of genetic material are not self-operating, autonomous entities has been overlooked by the dogmatic attitude of molecular biology (but not by Stent). We analyzed this empirical fact (Garzillo & Trautteur Reference Garzillo and Trautteur2009) and proposed that what those biological strings needed in order to beget expressive behavior was what in computer science parlance is an interpreter.
The core cellular machinery appears to consist of two parts: (1) the sequence of bases of the coding genes, which constitute a text specifying the structure of the proteins, and (2) the basic machinery of the cell, which we identify with the minimal gene set (Koonin Reference Koonin2000; Gil et al. Reference Gil, Silva, Peret and Moya2004; Glass et al. Reference Glass, Assad-Garcia, Alperovich, Yooseph, Lewis, Maruf, Hutchison III, Smith and Venter2006) together with the expressed enzymes coded by the minimal gene set itself.
Such an entity, that is, the minimal gene set encompassed with the expressed proteins that allow its own operation, is a biological forerunner of the universal Turing machine (Herken Reference Herken1988; Turing Reference Turing1936) and of the essentially equivalent stored program computer, soon to become the CPU chip of the PC. The core interpretive machine is made up of the coalescing within the same cytoplasm of the minimal gene set DNA together with the expressed enzymes coded for by the same minimal gene set. How it came to be that an active structure (ribosomes, etc.) evolved together with its own effective description (the genes coding for them) is at the core of the problem of the origin of life and of the genetic code. Suffice it to say that at present things are this way and allow for a computational interpretation. It is intriguing to note that this copresence of both the genic specification of the enzymes and the enzymes themselves had been “naturally” devised in the artificial, and computational, self-reproducing machine by John von Neumann (Reference von Neumann and Burks1966), a consequence of the 2nd Kleene recursion theorem (Rogers Reference Rogers1967). As a further remark, the epigenetics possibility that an initial self-reproducing genome may keep “improving” itself indefinitely had already been delineated by John Myhill (Reference Myhill and Burks1970).
The gist of our commentary is that the cell being an universal Turing machine entails that the genome, besides coding for structural proteins, does also code for “functional” proteins, whose enzymatic activities amount to program execution: Some fulfill control actions on metabolic pathways, thus interfering with the external milieu – the nutrients or signaling molecules – and some act upon the genome itself. For instance, in the case of the transposases, the enzyme recognizes the ends of the transposon and cuts it off a chromosome. That freed-up piece of DNA, with the transposase still attached, does move to a new place thus preparing on the fly a new genic string, which gets expressed in turn, always through the operation of the universal Turing machine–like minimal gene set. Such activities parallel common programming practices such as code displacement, code modification, or creation at execution time.
We discussed elsewhere the action of the genome upon itself (Garzillo & Trautteur Reference Garzillo and Trautteur2009) with examples overlapping many examples of the target article. In particular, we considered the treatment of the transposable elements and some further examples concerning the ciliates.
If systems biology concludes “that causation in biological systems runs in both directions: upward … and downward,” that “there are feedback and feed-forward loops between different levels,” and that “DNA is no longer privileged as the sole carrier of information” (all quoted from target article sect. 10, para. 1), it need not worry about causation, levels, and the nebulous notion of information more than it is done in computer science.
No further entity or phenomenon, beyond the genome, needs to be looked for in epigenetic research. Nor is a paradigm shift in the offing, unless one would call paradigm shift the interdisciplinary usage of established concepts in a different discipline. Epigenomics' travail (Ferguson-Smith et al. Reference Ferguson-Smith, Greally and Martienssen2009) consists in the painful and, let us not forget, therapeutically invaluable unraveling of an exceedingly complex computational machine.
You have
Access
The target article displays a thorough and impressive gathering of cellular phenomena that cannot be accounted for by “classical” genetics. Such phenomena are then rallied against the fact that “it has been a key dogma of molecular genetics that DNA is the sole biological agent of heritability, and it is still commonly treated as such.” Such state of affairs, the author intimates, implies a Kuhnian paradigm shift towards a “postgenomic view” which, nevertheless, Charney does “not call … a paradigm because it has not yet coalesced around a core set of principles or assumptions characteristic of a paradigm.”
Two lines of study are identified in the target article: heritability studies and gene association studies. Out of both these lines, a feeling oozes that the implicit debate underlying the contrast between dogmatic biology and the more permissive postgenomics has been replacing the now largely obsolete nature–nurture dispute. Günther Stent (Reference Stent1975), through his dissatisfaction with the unique agency of the genome and his assertion of an implicit role of the cell, anticipated the essential uneasiness of the actual epigenetics movement.
So what is it this further element, over and above the dogmatic centrality of the DNA, which is alleged by the epigenetics community? We think it is the computational nature of the cell.
This runs contrary to Charney's statement: “DNA does not contain a determinate genetic program (analogous to the digital code of a computer) from which we can predict phenotype.” And it might well be true that the phenotype is not predictable, but rather more because generally program behavior is undecidable (Davis Reference Davis2004), hence unpredictable, than because the genome is a program, or rather a computational structure.
There is a lack of appreciation that the computational and information theoretic talk about the cell, current since the very beginning of molecular biology, is not metaphoric but literal, and that cellular processes are primarily computational rather than biochemical ones, in the same sense that the CPU chip of a PC is primarily a computational apparatus rather than a solid-state physics device.
As stressed ubiquitously in the epigenetics domain, the gene does not express the protein by itself. The author states as much in Note 6, and indeed, the fact that the strings of genetic material are not self-operating, autonomous entities has been overlooked by the dogmatic attitude of molecular biology (but not by Stent). We analyzed this empirical fact (Garzillo & Trautteur Reference Garzillo and Trautteur2009) and proposed that what those biological strings needed in order to beget expressive behavior was what in computer science parlance is an interpreter.
The core cellular machinery appears to consist of two parts: (1) the sequence of bases of the coding genes, which constitute a text specifying the structure of the proteins, and (2) the basic machinery of the cell, which we identify with the minimal gene set (Koonin Reference Koonin2000; Gil et al. Reference Gil, Silva, Peret and Moya2004; Glass et al. Reference Glass, Assad-Garcia, Alperovich, Yooseph, Lewis, Maruf, Hutchison III, Smith and Venter2006) together with the expressed enzymes coded by the minimal gene set itself.
Such an entity, that is, the minimal gene set encompassed with the expressed proteins that allow its own operation, is a biological forerunner of the universal Turing machine (Herken Reference Herken1988; Turing Reference Turing1936) and of the essentially equivalent stored program computer, soon to become the CPU chip of the PC. The core interpretive machine is made up of the coalescing within the same cytoplasm of the minimal gene set DNA together with the expressed enzymes coded for by the same minimal gene set. How it came to be that an active structure (ribosomes, etc.) evolved together with its own effective description (the genes coding for them) is at the core of the problem of the origin of life and of the genetic code. Suffice it to say that at present things are this way and allow for a computational interpretation. It is intriguing to note that this copresence of both the genic specification of the enzymes and the enzymes themselves had been “naturally” devised in the artificial, and computational, self-reproducing machine by John von Neumann (Reference von Neumann and Burks1966), a consequence of the 2nd Kleene recursion theorem (Rogers Reference Rogers1967). As a further remark, the epigenetics possibility that an initial self-reproducing genome may keep “improving” itself indefinitely had already been delineated by John Myhill (Reference Myhill and Burks1970).
The gist of our commentary is that the cell being an universal Turing machine entails that the genome, besides coding for structural proteins, does also code for “functional” proteins, whose enzymatic activities amount to program execution: Some fulfill control actions on metabolic pathways, thus interfering with the external milieu – the nutrients or signaling molecules – and some act upon the genome itself. For instance, in the case of the transposases, the enzyme recognizes the ends of the transposon and cuts it off a chromosome. That freed-up piece of DNA, with the transposase still attached, does move to a new place thus preparing on the fly a new genic string, which gets expressed in turn, always through the operation of the universal Turing machine–like minimal gene set. Such activities parallel common programming practices such as code displacement, code modification, or creation at execution time.
We discussed elsewhere the action of the genome upon itself (Garzillo & Trautteur Reference Garzillo and Trautteur2009) with examples overlapping many examples of the target article. In particular, we considered the treatment of the transposable elements and some further examples concerning the ciliates.
If systems biology concludes “that causation in biological systems runs in both directions: upward … and downward,” that “there are feedback and feed-forward loops between different levels,” and that “DNA is no longer privileged as the sole carrier of information” (all quoted from target article sect. 10, para. 1), it need not worry about causation, levels, and the nebulous notion of information more than it is done in computer science.
No further entity or phenomenon, beyond the genome, needs to be looked for in epigenetic research. Nor is a paradigm shift in the offing, unless one would call paradigm shift the interdisciplinary usage of established concepts in a different discipline. Epigenomics' travail (Ferguson-Smith et al. Reference Ferguson-Smith, Greally and Martienssen2009) consists in the painful and, let us not forget, therapeutically invaluable unraveling of an exceedingly complex computational machine.