There is no requirement that the code support operations in Galois vector space.
So a TTC DNA codon acts as input to a biochemical channel and yields an RNA UUC codon. That codon acts as input to another biochemical channel that yields phenylalanine. Under Shannon information theory this is an information pathway with codes and channels. It even has some noise and (slightly imperfect) error correction mechanisms.
Geneticists universally use the term codon to describe nucleic acid triplets. There is a reason they chose that word.
I fully agree with the caution that this terminology should not carry metaphysical implications. Shannon information theory certainly does not support any metaphysics.
@Dan_Eastwood ā Youāre forgetting that an analogy is a three-legged horse. The information is stored in the DNA, not the operations. And the operations are physical, not logical. (Even in digital electronics, the actual operations are physical, not logical ā a change of state from low to high is symbolic to us, not the hardware.) The āoperatorsā youāre looking for are primarily enzymes, and they are way more complex than symbolic and physical binary.
Another horse short a leg would be to think of the operators as CNC machines manipulating and modifying materials.
So it still stands that DNA is essentially a quaternary digital ROM. And it is already being used that way, synthetically, or maybe more accurately, as an EPROM.
On the music box, the drum is like a ātemplateā in a way that the, for example, ASCII code is not. This is, in this sense, a closer analogy to DNA. However, it does not have a paired self-templating capacity. The original phonograph technology might be an even better analogy in that sense. Those old phonographs, however, were analogue, not digital (discrete), so that analogy falls apart in a different wayā¦
Iām a little baffled that this discussion has gone on for 104 posts and no one has once brought up the 64 amino acids combinations (codons) that map onto 20 different proteinsamino acids
. THAT is exactly the relationship that I understand Perry et. al. to be referring to when they say discuss DNA as code. As I chowed my way through popular science books, I used to wonder, why do the 64 map to the 20 in the way that they do? Is it arbitrary? Is it deterministic? Perry and others seemed to refer to it as if it was arbitrary, but either he didnāt describe why in his book or I didnāt have the background to understand.
Simon Conway Morrisās book Lifeās Solution also spoke of it as if it was arbitrary, marveling at the incredibly high āefficiencyā of the ācodeā, referring (I think) to the way in which it handles possible combinations of one-off errors that still map to the same codon, literally the first or second most efficient possible out of millions of possible mappings.
It wasnāt until I read Godel Escher Bach that I stumbled on an explicit explanation of how the mapping itself is determined - maybe it was Hofstadterās computer science background that helped illuminate things for me - I forget the nomenclature - trans-something? - but the mapping has something to do with the available connecting ends of the created proteins which in a beautiful recursion (the main theme of the book) is essentially hardcoded in the DNA itself. So mutations there can change parts of the 64-to-20 mapping, which would be just like āarbitrarilyā changing the agreement for the ASCII code for āBā from ā1000010ā to something else, and apparently there are very minor variations of that mapping in a few organisms, but obviously mutations there would generally have catastrophic effects and tend to be highly conserved.
Anyways, it is obvious to me how that apparently arbitrary 64-to-20 mapping relates to something like the similarly arbitrary ASCII mapping. Now as to whether I actually understand any of that correctly, or as to whether that actually has any meaning to the overall conversation of whether āDNAā is or is not like ācodeā in general, I will leave that to the experts here as I continue to learnā¦
@joshuahedlund, that is (somewhat) how the genetic code is like a code. That does not make it a code, because it is also different. Do you know how it is different?
Thanks for the title, Gƶdel, Escher, Bach! An intriguing combination. It is not in our public library (I should request it), but is, as I expected, in our universityās. I may have to do some calisthenics to check it out with my wifeās emeritus ID card but no current .edu email address, but it looks worth the effort.
That is what we are exploring here. The different ways DNA is different than merely a code.
Also, the association between codons and amino acids is not arbitrary. There are strong chemical relationships that govern the code. If that is the definition āarbitrary relationshipā then the genetic code is not a code.
Great - I only read it this year but it seems to be a sort of cult classic among a lot of computer/engineering/philosophy/science-minded folks, and it holds up well despite being a couple decades old - Itās extremely long but overall both amusing and intellectually stimulating throughout, and even though I do not share the authorās priors about the nature of the soul it certainly gave me a lot to consider.
I would object, as itās not a good analogy. For āconventionalā (the term is an artifact from when they were discovered) myosins like those of muscle, they form helical filaments with the heads poking out at regular angles. Thus, very few of them are bound to actin at any given time, while the dancers would all have at least one foot on the ground. For āunconventionalā dimeric ones, one head would be bound most of the time, with interesting mechanisms to maintain the binding. For unconventional monomeric ones, they generally work in bunches with most heads detached.
AFAIK, it is not a requirement of Shannon information theory that a code mapping function be arbitrary. Thatās why it is valid to refer to a āgenetic codeā under the theory.
Of course, Shannon information theory is not well understood by the general public, so I can sympathize with someone like yourself, Joshua, who is working hard to educate the lay folk. You face a dilemma: use Shannon terminology and be widely misunderstood vs. find a different but less intuitive terminology. Not an easy choice.
When Stephen Meyer and the ID crowd assert that the genetic code necessarily implies a codesmith, they certainly muddy the waters IMO.
Shannono information theroy has nothing to do with this. What is said is āIf a code is defined as an arbitrary mappingā, refering to what Perry and @joshuahedlund mentioned said:
@joshuahedlund found out later on reason the code is not arbitrary, and there are others. Each of these reasons (IFF CODE = ARBITRARY MAPPING) makes it less and less like a code.
In this thread, maybe not. But it has everything to do with how and why geneticists created the āgenetic codeā terminology about 50 years ago. My source on this is James Gleickās book, The Information A History.
