Intelligent design and "design detection"

I disagree. Similarities conserved throughout long evolutionary periods (more than 400 my) denote positions that have been preserved by purifying selection, which reflects very strong functional constraints. So there is a close connection between the level of conserved similarity through deep time and functional information.

But Gpuccio isn’t even measuring that. All he is doing measuring is similarity, not conservation.

That might be true if you were actually measuring sequence conservation and if the sequences you examined actually were an adequate sample of sequence space. But neither of those things is actually true. Any such measure certainly doesn’t match Szostak’s definition, which requires exhaustive examination of both the sequence and the degree of function of all possible proteins.

Of course, you tend to ignore any point I make. But I guess you tend to ignore any point anyone makes.

Sure, it implies those are the positions that have higher fitness.

That just flat out doesn’t follow. For all the reasons stated previously. To which you had no response.

To brielfly reiterate the reasons why you’re not measuring information, and why the inference to design from FI logically cannot work:

  1. Evolution has only sampled a tiny portion of sequence space, so you can’t use conservation to show there is no function out there in very dissimilar sequences(we know of many examples of entirely dissimilar structure-sequence relationships that perform the same functions, such as being enzymes that catalyze a specific reaction).
  2. Different sequences have different fitnesses, so one being discard in favor of another doesn’t mean those lower-fitness sequences didn’t strictly “work”, just that they had lower fitness. A higher fitness sequence can evolve from a lower fitness sequence.
  3. Many different functions overlap in sequence space(all enzymes overlap with simpler binders, for example.)
  4. Ultimately this also comes back to the Texas sharpshooter fallacy, as you’re trying to calculate the probability of randomly guessing a sequence performing a particular function on a particular hill, instead of randomly guessing any new fitness-improving function from the ancestral state.
1 Like

Shall we also point out that Gpuccio and you suffer from human exceptionalism syndrome? By his measure, any human sequence is the pinnacle of functionality (or has the most functional information) and that of any other species is less functional, increasingly so as the date of divergence increases. Is that in any way rational? Do you really think that tunicate cytochrome c is less functional than trout cytochrome c?

1 Like

You can rest assured, Gpuccio doesn’t suffer from human exceptionalism syndrome with regards to his methodology. To see that, I invite you to have a look at his post no 37 in the thread below.

I’m afraid you don’t understand what he says there. His BLAST bitscore is a measure of similarity between a human protein sequence (entered as the search image) and a presumably homologous other sequence (which BLAST finds by comparing similarities in its database). The more similar to human, the higher the bitscore. Thus if bitscore is considered a measure of functional information, the human sequence has the most functional information of all, being 100% similar to the human sequence. It’s possible that Gpuccio doesn’t understand what he did either, but that’s what he describes. Equating this with Szostak’s measure of functional information is pure fantasy.

Why, just look at the graphs. All those supposed increases of functional information just demonstrate increasing similarity to human proteins. That’s all they show.


Agreed. We’ve been through this before with Gil. If there is a “close connection,” then myosin has far less FI than actin, which is absurd.

Gil and Bill also never explained why similarity between species counted but similarity within a species (humans, even!) didn’t. That’s even more absurd.

Agreed. Gil, can you at least concede that?

Gil certainly has ignored every point I’ve made regarding this.

Your conclusion would appear to be overstated. It would appear to be more accurate to state:

So there may be a connection between the level of conserved similarity through deep time and functional information.

As far as I can ascertain, neither you nor Gpuccio have compared your proxy measure to FI calculated using Szostak’s definition – so you and Gpuccio are merely ASSUMING that there is a “close connection”.

  1. Even assuming that there is a genuine “connection” there, you have no way of knowing how much noise there is in the correlation.

  2. You have no way of knowing if the relationship between the two would be linear, logarithmic, exponential, etc – so no basis for assuming that the relationship would be one to one.

  3. You haven’t ruled out what else your proxy measure might be “closely connected” to – so have no way of telling if it is picking up spurious variations that are unrelated to variations of FI.

These are the sorts of things that would need to be nailed down, before Gpuccio’s argument has any hope of garnering any credibility.

Gpuccio made clear several times that he doesn’t estimate the totality of FI, but only the human specific FI. Let me remind you what Gpuccio said at 37:
I am not saying that vertebrates are in any way special. I am not saying that humans are in any way special (well, they are, but for different reasons).

2. It should be clear that my methodology is not measuring the absolute FI present in a protein. It is only measuring the FI conserved up to humans, and specific to the vertebrate branch.

So, let’s say that protein A has 800 bits of human conserved sequence similarity (conserved for 400+ million years). My methodology affirms that those 800 bits are a good estimator of specific FI. But let’s say that the same protein A, in bees, has only 400 bits of sequence similarity with the human form. Does it mean that the bee protein has less FI?

Absolutely not. It probably just means that the bee protein has less vertebrate specific FI. But it can well have a lot of Hymenoptera specific FI. That can be verified by measuring the sequence similarity conserved in that branch for a few hundred million years, in that protein.

It would be more accurate to state that Gpuccio’s methodology assumes “that those 800 bits are a good estimator of specific FI” – as we have no evidence that it’s a “good estimator” of anything of the sort.

This is not “design detection” but merely ‘design assumption’ and ‘design assertion’.


It sounds like Gpuccio doesn’t understand what he’s doing. A fine example of the pseudoscience of ID. I can’t imagine why you’re trying to defend it when it dies no credit to you or ID.

“Human specific FI” would - by definition - be in humans alone, so you’re really not making sense here. However Gpuccio’s measure does assume that humans are special because the measure of information is similarity to the human protein. Neutral changes will be counted as information - and if the sequence is highly conserved there will be neutral changes that persist for long periods of time.

It should be clear that Gpuccio’s methodology is not measuring that at all. There’s not even an attempt to measure FI. You ought to know that by now,

Since he doesn’t measure bits of conserved sequence similarity - and they wouldn’t be a good measure of FI if he did his methodology is hopelessly flawed. Again, this is a clear example of pseudoscience.


What does that even mean? How does “human-specific” FI compare to plain vanilla FI? How can he be measuring this quantity of unknown significance and simultaneously claim to be following Szostak’s definition? And now you and Gpuccio admit that he isn’t measuring Szostak’s FI, just the similarity of sequences to human sequences, which is what the rest of us have been saying all along. This does not help your point.


I think you’re right here, and I should have said « Gpuccio made clear several times that he doesn’t estimate the totality of FI, but only vertebrates specific FI. »

Not sure to understand your point here. But no, neutral changes don’t persist for long periods of time.

Then how can he claim it can’t evolve? The closer you get to human, the more human-specific the sequence will become. The canonical human sequence evolved from a not-entirely-human sequence (which was specific to that ancestor), which evolved from one even less so, etc.

Same goes for all the other species-specific variants of the protein. They each evolved from ones less like the present one and ancestral to it. Incrementally over generations.

His method doesn’t work. He’s not really measuring FI, just similarity, and since extant species-specific sequences evolved incrementally from ancestor-specific sequences, he can’t justify any conclusion that the extant sequence couldn’t evolve.

Why don’t you get this? I must reiterate my request to have you explain how measuring sequence-similarity tells you that any extant sequence could not have evolved from some ancestral sequence increasingly dissimilar to it?

Please connect the dots. Explain how the similarity between two sequences becomes proportionally less and less with increasing time since their common ancestor, implies the extant sequence couldn’t have evolved incrementally from that common ancestor. And further, how this implies the last common ancestral squence in turn could not have evolved from some other sequence possibly performing some other function?

If you’re going to say you can infer design by ruling out evolution, and that you can rule out evolution with these similarity measures, you need to do a lot of work you haven’t done. You’ve just assumed what we are asking you to explain.


It still makes zero sense.

Why not? Please show your math.


Espistasis can lock in changes that were initially neutral. In the genetic background, specific mutations are neutral, but then when they have occured, further mutations can “lock in” those neutral mutations such that further changes to those residues are deleterious.

Yes, they can - if the sequence is highly conserved. If only a few changes to a sequence are possible, change will be rare - whether it is neutral or beneficial.

No, you were right the first time. He measures human-specific FI if he measures any sort of FI. The problem with that is that human-specific FI makes no sense, and he doesn’t measure FI anyway, just the degree of similarity of some protein sequence to a human sequence. The great scientific discovery Gpuccio has made, in fact his only discovery, as encapsulated in the two figures posted here, is that the more closely related a taxon is to humans, the more similar its proteins are likely to be to human proteins. This is cargo cult science at its finest.

1 Like

Who is Gpuccio? Any sort of CV or Bio?