… There is a vision of the cell as a nugget of information suspended in a soup of dumb and formless goo …
This is from Processes of Life: Essays in the Philosophy of Biology by John Dupré (2012):
… Surely within the practice of genetics and genomics, as within any human practice with even a minimal intellectual content, there are arguments. For instance:
This gene codes for the Bacillus thuringiensis toxin
If we insert it into the genome of this plant, the plant will produce BT toxin
BT toxin poisons insect pests
Therefore, if we insert this gene into this plant, the plant will poison insect pests.
This argument is plausible, if a bit enthymematic. One premise that might start to flesh it out is:
If we insert a gene for x into a (living) genome then that genome will produce x.
This premise shows us that the argument, whether or not plausible, is not sound. For the missing premise is certainly false. There are lots of reasons for this falsity. One of the most interesting involves the familiar redundancy of the genetic code. Amino acids, the constituents of proteins, are coded for by as many as six different base-pair triplets. However, different organisms tend to use different triplets preferentially and will be disproportionately equipped with extra-nuclear equipment for reading the preferred codons. Consequently they may be very bad at transcribing a gene from a distantly related organism. More simply, whether a sequence is transcribed will depend very much on where it ends up in the genome, on its spatial relations to other genes, especially promoter and suppressor sequences, and even to other structures in the cell. Current techniques for inserting genes into alien genomes are thoroughly hit or miss as far as where the genes end up.
Another reason that the gene may fail to produce the toxin is that the plant may die before it has a chance to do so. If the inserted gene should land in the middle of a sequence of the genome vital for the plant’s functioning then the plant will not function. Inserted genetic material may also have a range of effects on the host organism distinct from those intended (pleiotripy), and these may be harmful or fatal.
The relevant moral of these genomic factoids is that genomic events are diverse and specific. One familiar model of scientific argument, that most closely connected to mathematical ideas of proof and demonstration, essentially involves generalization — traditionally thought of as scientific laws — and generalization is a risky business in biology generally and genetics in particular. The simple example just discussed illustrates the difficulty. The attempt to convert such simple generalizations into exceptionless laws would be extremely difficult if not impossible. Such considerations lead naturally to the conclusion that there are few if any laws that apply to genes.
Dupré fills out his argument with material that has already been covered in previous posts from this book (the text is a collection of what were originally independent papers). I’m not going to repeat those arguments here. Jumping to his conclusions:
… So my proposal is for an atheoretical pluralism similar to that which I advocate for [the categorization of] species: a gene is any bit of DNA that anyone has reason to name and keep track of. Genes may be proper parts of other genes; they may overlap; they may have non-contiguous parts, perhaps on two or more chromosomes. And, as illustrated for the case of developmental defect genes, ‘gene’ may even refer to a functionally connected class of DNA segments. My conclusion is that there are genes — an important point given how much people talk about them — but that the price of this is conceding that it doesn’t take much to be a gene. Not much, but not nothing either. I am assuming that genes are real material entities. Many of the genes discussed by behavioral geneticists for instance, may well not even meet this minimal condition.
… The great diversity of the subject matter of biology calls for the most central terms not to be those in terms of which laws can be formulated, but rather those which are tolerant enough in their reference to bridge the divides between the various phenomena in which local communities of researchers may be interested. There are, I suppose, some general truths about DNA that make it possible for DNA to constitute genes, but there are lots of ways for bits of DNA to be genes of various kinds, and all of these depend on the relations between bits of DNA and other things to which they are related.
It is, as I have noted, hardly a novel suggestion that the view of science as the search for universal laws is of little or no relevance to biology, but the extent to which this suggestion has been reinforced by recent developments in genetics has not yet been fully appreciated. Indeed, it is still sometimes imagined that the annoying failure of biology to generate law-like generalizations is a consequence merely of its continuing concern with complex and variable structures, and its concomitant failure to get down to the real action at the molecular, and ultimately even moree fundamental, levels.
One moral of my preceding remarks is just that no such consequences result as we investigate the inner structures of biological things. On the contrary, what we find as we become more familiar with molecular processes is a diversity of structure and action quite comparable with that which we find at more complex levels. We are far from approaching the few simple laws that earlier theorists imagined might reduce complexity and diversity to order and uniformity.
My argument here is not in any simple way anti-reductionist. It is clear in genetics that enormous illumination and insight has come from our ability to investigate and describe molecular processes. It is, however, anti-reductionist in the sense of rejecting the hierarchical view of nature often associated with reductionism. Knowledge of different levels of organization is complementary, not competing. The molecular view is not a superior view to, say, the cellular view, and one that in principle should render the latter obsolete. And the reason for this is simply that the molecular view is not even separable from the cellular view. There is no possibility of specifying the behavior or function of bits of DNA independently of a detailed description of the biological context in which they exist. Minimally this context will include further genomic and cytological information. Sometimes the relevant context will be much broader, including physiology, ecology, and even sociology. And of course this dependence on context is a large part of why what may look very similar — strings of DNA — may nevertheless prove to be so diverse.
There is a vision of the cell as a nugget of information suspended in a soup of dumb and formless goo, a notion that still seems common in popular presentations of biology, and this vision perhaps best represents the remaining aspirations of hierarchical reductionism. The extra-nuclear goo, in this vision,, is no more than the minimal context necessary for the expression of the structure inherent in the DNA.
But in reality the exra-nuclear goo is as structural, as rich in information, as is the nuclear DNA. The sorts of things bits of DNA can do involve diverse reactions with particular chemicals and structures in the cell. Biochemistry only becomes molecular biology when it is embedded in cytology. Lower-level knowledge cannot possibly displace higher-level knowledge.