That isn’t what the graph shows, though. He may be interested in it, but the graph just shows that (with an exception or two), the longer ago that a taxon split from humans, the less genetic similarity it has. That’s not about conservation. And the big lesson supposedly in that graph, of a sudden leap in genetic information, is spurious.
But that’s not what he presents in the graph. He shows BLAST results. Neither the target space nor the search space is the same thing Szostak is talking about.
Gpuccio uses BLAST results as a proxy for FI as defined by Szostak.
It’s a laughably bad proxy, so claiming that he’s using Szostak’s definition is objectively false. Why do you insist on repeating this falsehood?
You’re correct. This is why Dembsky wrote a book titled « the design inference » rather than « design detection ».
That’s like equating longitudinal minutes with the minutes on a clock.
As pointed out already by many, it’s a ridiculous and invalid proxy. All your little one-line non sequiturs here put you in danger of going down @colewd’s road. Are you sure you want that kind of image?
The fact that someone use a spurious proxy for some entity doesn’t mean that he uses an incorrect definition of that entity.
The spurious proxy is an integral part of gpuccio’s definition. Thanks for admitting that you know it is worthless, though.
It means exactly that, for any reasonable meaning of “uses”.
Surely using a spurious proxy invalidates any claim to have detected a jump in functional information.
Notice that it isn’t actually a jump in anything, since the x axis is meaningless. That the y axis is also meaningless is just gravy.
This may be seen as a non sequitur!
The appeal to « many » doesn’t guaranty you are right. It would be better if you could explain why you think Gpuccio’s proxy of FI as developed at 39 of the thread we are discussing is spurious.
True, but I am in fact right, and several people, including me, have explained why this is so. Why, you even seem to have admitted it yourself. Feel free to read those explanations, already available above.
It seems to be a simple fact.
Perhaps you should just reread that thread instead of trying to rehash points already made.
No, it’s the one you references, the one in post 39. The one you show there has a meaningful x axis, though the locations of points on it may not be. Perhaps the legend explains what we’re looking at there? How was it determined that the deuterostome common ancestor (if that’s what “deu” means) lived only around 440ma? I assure you, that’s nonsense.
By whom? Everyone posting here agrees. You even admitted that it is spurious as a proxy.
I have already done so in great detail in a thread in which YOU claimed that it was part of the definition.
I’ll repeat my generous offer to you from 2 years ago. Why did you not even acknowledge it?
So, Gil, shall we go through these papers together?
Yes, but what on Earth could have persuaded him that is even within the realm of being a valid proxy?
Ironically, even Douglas Axe has proven Gpuccio’s method so exceptionally incapable of that, that it is effectively incomprehensible how wrong it is, in degree.
Remember Axe’s number from his 2004 paper? 10-77 for a protein 155 amino acids long. Let’s forget all about all the things that make Axe’s number wrong and just, for the sake of argument here take Axe’s number at face value. The total sequence space for L=155 proteins is 20155 = ~4.6 × 10201
So if only 1 in every 1077 sequences is functional, but there’s 20155 = ~4.6 × 10201 total possible sequences of equal length, then there must be
~4.6 × 10201 / 1077 = ~4.6 × 10124 total functional sequences.
Isn’t the common saying among ID-creationists that evolution can at most have sampled something in the range of ~1045 sequences in the entire history of life? If that is true, does that not immediately imply there is an unfathomable diversity of unsampled functional sequences out there still? How do those show up in blast searches again?
Nothing of what Gpuccio is doing makes sense. NONE of it.
Now comes the key question I want you to try your best to answer, with math:
How on Earth is evolution supposed to have sampled the totality of functional sequences in a way where you can derive that totality from a blast bitscore?
Edit: Suddenly I recall we’ve been over this exact thing before.
But the BLAST scores he’s using are worse: they sample only two sequences and tell you how similar they are, i.e. the probability that they have that degree of matching by chance. That has nothing at all to do with functional information by anyone’s definition, and fits no feature at all of Szostak’s definition.
Here is the legend: figure 1 shows a plot of the mean bits-per-amino acid score in the various groups of organisms, according to the mentioned approximate times of split.
Yes, « deu » means deuterostome.
The point « deu » in the graph denotes the approximate time of split between pre-vertebrate deuterostomes (including chordates like cephalocordates and tunicates) and vertebrates (cartilaginous fish). According to Gpuccio, this split roughly occurred 440 My ago. But he acknowledges that he is not an expert on this issue and so would welcome any suggestions to improve his judgement.
