The Genetic Code: A Teleological Perspective


But is it the only possible set of amino acids that could show optimization?

Why can’t it be both? The frozen accident may have happened very early on (i.e. prior to evolution of a full set of codons and amino acids), and the genetic systems we see now have been optimized through evolution. If the earliest life had been somewhat different we may have seen a different optimum than what we currently see.

That would also be what we predict if the genetic code evolved.


As you haven’t spelled out your argument, only somewhat-condescendingly thrown out vague references, I guess I’m supposed to reconstruct it myself. As far as I can tell, your argument goes something like this:

  1. Life employs amino acid residues that aren’t specified by the genetic code
  2. It would have been better if those residues had also been specified by the code, in addition to the standard 20 residues
  3. Therefore, life does not exhibit the optimization that Freeland and others claim to have found

But it’s not at all obvious that the genetic code would be improved by the expansion to cover those currently unspecified residues. For one, it would mean that the genetic code lost some of its redundancy, as some mutations, which currently do not result in a changed amino acid residue, will now result in one of those additional residues.

You are more than welcome to make those arguments.

The falsification is only as strong as that of the phylogeny used to back it up. For example, phylogenies based on 18S rRNA sequences placed microsporidia at the base of the eukaryotic tree, supposedly a precursor to mitochondria-bearing eukaryotes. Today, thanks to more lines of evidence (an insertion in microsporidia shared only by animals and fungi, phylogenies based on tubulins, and mitochondrial genes in the microsporidia genome), we know that microsporidia nest within or close to fungi, and that they once possessed mitochondria but subsequently lost them.

The concept is defined mathematically in the papers I cited in my opening post.

No, Ilardo et al. found six sets of amino acids (oops, sorry, amino acid residues) with better properties than the set specified by the genetic code. Now, it’s possible that the number of those sets drop as more properties are considered, but I don’t have the model to test that, so I don’t assert that. However, the set of residues specified by the genetic code appears to be extraordinarily good, or pretty darn good.

So if something is the product of evolution, we should expect it to be exceptionally good? In his post above, @Haywood claims that the translation machinery is “pretty darn bad”. Is this also something evolution predicts?


This raises the question of why we only see one set of amino acid residues in proteins if there are many optimal sets to choose from. So why doesn’t teleology also predict that we should see many sets of amino acids in different types of life?

We should see fit adaptations since this is what evolution does. Mutations that increase fitness are kept.


Each novel set would require a novel genetic code, and as I explained in my original post, there is a good engineering reason for reusing the genetic code.

Indeed they are. So what does the insight that “mutations that increase fitness are kept” lead us to expect regarding whether the translation machinery is “pretty darn good” or “pretty darn bad”?


That’s what I find interesting. Here is what you said in the opening post:

“Before answering this question, let us ask a counter-question: Why not? What would the point be, from an engineering perspective, to reinvent the wheel? Making multiple codes is extra work and increases the risk of mistakes when genes have to be designed in different languages.”

I would agree, that makes sense for human engineers. We reuse code because it is extra work to start from scratch and come up with something new that is just as optimal. Of course, that doesn’t stop humans from using different code to do the same thing, such as the Chrome web browser that I am currently using which looks the same on PC and Apple, but has very different code underneath. Nonetheless, I would agree that for designers like humans this makes sense.

However, for an omnipotent and omniscient designer who has unlimited time, resources, and power it would be equally difficult to reuse designs as it would to use different ones. Therefore, teleology would not necessarily predict the reuse of designs simply because it is easier for humans to do so.

If there were mutations that improved the genetic machinery then those would be kept, so we would expect to find examples of optimized genetic machinery. There is also the possibility that there were bad designs which no mutations could improve without destroying the system. This would result in the freezing of bad designs. So we would expect a mix of bad and optimized designs if evolution is true, and that is what we often see both at the genentic and organismal level.


I don’t know that the designers are omnipotent or omniscient. I’m an agnostic, not a Christian.

Is there any way to know in advance which structures are predicted to be either bad or optimized? Or do we need to know if the structure is bad or optimized before we can tell?


I understand that. I was simply pointing out the implications of your argument.

Not really. Evolution appears to be a blind process that can only find local optimizations if they are available. This contrasts with a teleological process which knows what the optimizations are and can overcome valleys of lower fitness in order to find other fitness peaks in a fitness landscape. With evolution, the path between the current state and an optimal state might require mutations that initially reduce fitness which makes optimization nearly impossible.

(Haywood Clark) #28

Well, in my defense, it’s hard to demonstrate to someone who claims to understand biology better than 99.999% of biologists that he’s ignoring critical evidence. How would you recommend I do it less condescendingly?

My hypothesis is that you are ignoring vast amounts of evidence that don’t support your argument. Secondarily and less charitably, I hypothesize that you are doing so intentionally. I therefore asked those questions to test those combined hypotheses, and your evasive reply did not falsify them.

Perhaps you should start over and simply answer my questions about the number and the difference? It would be helpful if you knew the number before answering. :wink:

I’m sorry, but what does any of that paragraph have to do with my pointing out that predictions have to predict what will be observed directly, not our opinions of what we observe?

Me: > The same goes for your inclusion of “robustness” in your second prediction.

Krauze: The concept is defined mathematically in the papers I cited in my opening post.

Then you should have no problem expressing the prediction as an observation! Would you kindly do so?

So if something is the product of evolution, we should expect it to be exceptionally good? In his post above, @Haywood claims that the translation machinery is “pretty darn bad”. Is this also something evolution predicts?

Golly. I suggest that you reread more carefully. I wrote, “I can show you one of life’s solutions in the translation machinery that’s pretty darn bad, because it was constrained by evolution!”

To me, “one of life’s solutions in the translation machinery” is clearly not synonymous with “the translation machinery,” as I am offering to point out a very stark contrast between one and another within the set of solutions.

Is there some way I could have expressed that more clearly?


I haven’t claimed, nor do I believe, that I understand biology better than 99.999% of biologists.

As for the rest of your condescending demands, I have nothing further to add. I’m not looking for your approval. If you have some evidence you think I should be aware of, you’re more than welcome to share it. Or find something else to do with your time.

(Haywood Clark) #30

I did.

  1. Posttranslational modifications.
  2. Scientific predictions have to be about what will be directly observed, not how anyone will interpret said observations.

You’re clearly not interested. I understand completely. :wink:

(Stephen Matheson) #31

Hey, I’m just dumb enough to want to jump in here, but I think @Krauze deserves better. She/he is wrong about a lot of things but I don’t think s/he is a typical ID troll. Teleology and ID might not merit much respect, but I think @Krauze does.

(Haywood Clark) #32

It tells us that where we have only a single pathway to accomplish something that is essential for life, evolution is constrained to a local optimum, while a designer is free to replace components to provide vast improvements over the status quo. There are at least two spectacularly blatant cases of that in the translation machinery, one with a beautiful internal control.

A helpful, albeit limited, metaphor might be replacing your car’s engine. You can’t do that if it the car needs to keep running constantly (living), but you can replace spark plugs or an air filter without killing the engine.

(system) #33

This topic was automatically closed 6 days after the last reply. New replies are no longer allowed.