Re: Kenneth Miller's Interview with NOVA

On Dec 3, 2:05 pm, Seanpit <seanpitnos...@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:
On Dec 3, 6:58 am, hersheyh <hershe...@xxxxxxxxx> wrote:





*********Sean's appendix
Well, first we have to calculate the likely gap size. Using an
average between the calculations of Yockey and Sauer, the ratio of
potential beneficial vs. non-beneficial for 100aa systems is about
1e-40.
*********end

As I pointed out below this number is not what Sean says it is.
Rather it is the ratio of the number of sequences that have a specific
teleologically determined or specified function (as estimated by how
far one can mutate an existing molecule that has that function and
still have detectable function) divided by the *total* number of
possible sequences of that size. The reciprocal of that last number,
of course, would be the calculation of the probability that a
particular *single* sequence could arise by the "747 in a tornado"
strawman argument. So, by recognizing that more than one sequence (so
long as it is closely related to the one 'chosen' sequence that the
"747 in a tornado" strawmen makers use) can also have the same
'function'.

Finding the likely ratio for a single type of function that has a
particular minimum structural threshold requirement gives a pretty
good idea as to the overall ratio of potentially beneficial targets in
sequence/structure space.

Sean, I said that the number you gave is NOT the one you claimed to
give. You claimed to be presenting the ratio of beneficial to non-
beneficial sequences. But what you *really* presented is the ratio of
a particular functional sequence (estimated by those variants of an
already existing molecule with that function that still retain the
activity) divided by hypothetical total sequence space. Nothing you
have said below denies that.

Now, in this case the ratio would not differ significantly because the
difference between total sequence space and all sequences that are not
the sequences included in that particular function would not be
significantly different. But that is not the point. The point is
that you cannot even *copy* a number from another 'creationist' source
without botching the description in ways that say one thing and mean
another. In other cases, of course, the discrepancy would be
obvious. The ratio of heads in an honest coin flip divided by the non-
heads (the equivalent of your ratio of the teleologically specified
sequences to all sequences not so designated -- notice I did not say
'beneficial' and 'nonbeneficial', because you made no honest estimate
of that) would be approximately 1.0. Yet the number that is relevant
(and that your sources provided) is the equivalent of the probability
of heads (0.5, the ratio of heads to all trials), not the ratio of
heads to non-heads.

A simple ignorant mistake is the most likely explanation. But such a
mistake (mislabelling a specific function as if it were *all*
'beneficial sequences' and mislabelling total sequence space by
calling it 'non-beneficial sequences') means that you do not even
think you have to engage in the hand-waving b.s. you do below.

For example, lets be generous

How do you know when a creationist is going to present absurd
numerology and b.s. hand-waving? Ans: When they say "lets be
generous".

and say that
there are 100,000 uniquely different potentially beneficial systems
from the perspective of a particular life form at the 100aa threshold
level.

And where did you get this estimate from? Isn't the number you
*really* need NOT limited to any particular life form? You need to
know the frequency of all possible useful sequences or partial
sequences, both in the past and in the future, in all of sequence
space if you are going to present an estimate of the ratio of
sequences that potentially could be useful to total sequence size. I
see no way to get such an estimate, especially since life has not
explored all of sequence space and most *functional* sequences are
transmitted by inheritance rather than synthesized from scratch or by
a random walk. That means that *all* cytochrome c proteins have been
transmitted by descent as cytochrome c. Because evolution quickly
optimizes a function, that typically excludes the same function
arising a second time independently. All the niches where that
function is useful are filled by descent, not by 're-inventing the
wheel'. That doesn't preclude a second 'invention of the wheel'.
Indeed, there are examples of just that (including the rotary
flagella) in nature. And we *know* that functional protein sequences
completely unrelated to those that currently perform a function that
can arise that bind to and catalyse reactions that are performed by
other proteins or by none at all. Nylonase and the lactase derived
from antibodies are examples. I know of no way to estimate the
numbers you claim to have presented (but only by mislabelling one
specific function as "all beneficial sequences" and mislabelling total
sequence space as "non-beneficial sequences".

Let's say that each of these has a ratio of 1e-40.

Let's not. Using the data for cytochrome c is intentionally biasing
the results. Cytochrome c is not an average protein. It is a highly
conserved one.

What is
the overall ratio? It is only 1e-35.

And how much of "total sequence space" can be searched by known
mechanisms? How much *has been* searched? A slightly more relevant
number would be the number of "beneficial sequences" of all types (but
I see no way to estimate that, since what sequences are "beneficial"
is a conditional state and not an absolute or innate one) over the
number of *trials* or "actual searched sequence space", not "total
sequence space". Whatever the number of "trials" is, it is far, far,
far, far, far smaller than "total hypothetical sequence space".

You see, you notion that the existence of many different types of
potentially beneficial systems doesn't really change the ratio of
beneficial vs. non-beneficial to any significant degree.


The fact
remains that non-beneficial vastly outnumber beneficial structures/
sequences in sequence/structure space. Beyond this, this ratio of
potential beneficial vs. non-beneficial

Are you brain-damaged? The ratio you have presented is NOT the ratio
of potential beneficial vs. non-beneficial. It isn't even close to
that number, which I don't think can even be estimated (because
'benefit' is conditional and not innate). The ratio you presented is
the probability of a specific teleologic function (so long as it is in
a sequence close to that of the currently existing protein that fills
that function) divided by total hypothetical sequence space.

decreases exponentially with
each increase in the minimum structural threshold requirements.
That's a fact that anyone with any glimmer of a rational mind should
be able to grasp right away.

That is your assertion, of course. But I have only seen hand-waving
assertion and GIGO numerology used to support this claim.

Your notion that the minimum likely gap distance has nothing at all to
do with minimum structural threshold requirements is simply and
obviously ludicrous.

No it is not. *If*, as is likely, cytochrome c was initially
generated as a truncated duplication of some ancestral cytochrome c1,
the number of mutational steps needed to produce it is not correlated
with its size (nor that of the ancestral sequence). Similarly, the
number of mutational steps required to produce a lactase (or even
whether it can be formed or not) is not predictable or correlated with
the size of the protein. It is correlated with the presence or
absence of an appropriate precursor, with the number of steps a
function of the nature of that precursor. Again, the ONLY model of
evolution for which your numerology would have relevance is the
strawman model that I call the "747 in a tornado" model. But I notice
that you still are not discussing the assumptions of the model of
evolution you are refuting with your numerology.

In this same line, your wild idea that the
minimum likely gap distance is therefore always the minimum possible
distance (i.e., one mutation) is equally inane.

Quite the opposite. What I say is that, for those functions that
*did* evolve, they did so *because* there was a short pathway (in
number of mutational steps) available. If you remove the short
pathway, the function will not evolve. That is, those functions that
*have* evolved represent a biased sample.

The minimum likely
gap distance is directly related to the minimum sequence/structural
threshold requirements - - in an exponential manner. Very quickly this
exponential decline in the ratio produces a gap great enough so that
the odds of crossing it in less than trillions of years of time are
indeed equivalent to your 747 being created by an explosion or
toronado in a junk yard. Again, there is no real argument here. This
is a fact if there ever was one.

Only if you think evolution works by the "747 in a tornado" strawman
mechanism.

The fact that you can't recognize what is otherwise overwhelmingly
obvious

How do you know when a creationist is blowing smoke? Ans: When they
claim that their assertions are "overwhelmingly obvious".

is simply the result of the fact that you can't get your mind
around the concept that everything you've believed and argued for is
dust in the wind - absolutely worthless. This is why you continually
go around misrepresenting my position and what I've actually said to
you over and over and over again. It doesn't matter how many times I
point out to you that the minimum structural threshold requirements
are NOT the gap distance; that the gap distance is ALWAYS smaller than
the minimum structural threshold requirements.

And I have repeatedly pointed out that you only admit that when I
point out that the number you keep giving is not the number you need.
If you would only point out that your working assumption is that
evolution works, at least mathematically, by a mechanism
indistinguishable from random assembly of amino acids into a protein
of given length, allowing only that more than one sequence (so long as
they are all similar to one another) can serve that function. And
that, therefore, the only model of evolution you can refute by this
argument is the strawman "747 in a tornado" argument. BTW, I have
clearly pointed out that the numerology you use either is *directly*
the same as the 20^n argument (modified to account for the
embarrassing fact that many more than one sequence can have a
particular function) or is using a probability that can only be used
in such an argument (albeit mislabelled as the ratio of beneficial vs
non-beneficial sequences).

Regardless of how many
times I explain this to you, you still go around saying that my
suggested gap distance is the maximum possible distance.

That is what you imply when the only number you give is total sequence
size. You NEVER start out saying that gap distance is smaller. In
fact, what I have repeatedly pointed out is that you are NOT
presenting the number you need (gap size) *and* I have been the one to
have to point out to others that gap size number is much smaller than
the number you do present.

That, Howard, is a deliberate lie on your part. It is a complete
fabrication and you know it. Why keep up this desperate
misrepresentation over and over again when you know it simply isn't a
true representation of my actual position and never has been? You're
just a liar.

No. I am merely forcing YOU to tell the whole truth and, in fact, to
focus on gap size rather than total size. Why do you think that I
think that gap size would be large? I don't. You, in fact, have just
above accused me of claiming that gap size is very small. I certainly
do think that *when* new function arises it does so *because* gap size
(mutational steps) is small. But when you talk about a random walk to
a "function" in total sequence space and the only number you give is
total sequence size, that is falsely implying that that is the number
of steps needed.

Sean may think he is being 'fair' and not engaging in a
"747 in a tornado" argument. But the fact remains that
the only 'evolutionary' method in which this
number would have any meaning is
one where evolution worked by chance
and chance alone. This number is *still* assumes that evolution works
by a method that is only a slightly more sophisticated "747 in a
tornado" strawman.

The proposed evolutionary mechanism is random mutation and function-
based selection.

That is not what the math you are using says.

That is your proposed mechanism. As it turns out,
this mechanism does work by chance and chance alone until a new
beneficial target is actually discovered by pure chance. Until this
point, your selection mechanism is blind as far as any sort of biasing
agent.

Which is why, *when* a new function arises, it does so by modification
of pre-existing sequences by *short* pathways. [IOW, all your
numerology is irrelevant because new functions do not represent change
in average sequences. New functions are a biased subset.] After you
have *some* useful function, selection rapidly leads to the
optimization of that function within the constraints of the original
sequence. How much "function" do you need to lead to rapid
optimization? Take Q-beta phage. The w.t. phage requires a host
factor (HF-1) for replication. Knock out the host HF-1 gene, and the
Q-beta replication rate is reduced 10,000 fold. But continue growing
the phage in these cells and mutation and selection will eventually
produce a phage that no longer needs HF-1 to replicate at rates
similar to the w.t. Did Yockey include sequences that had 10^-5 as
much activity as w.t. cytochrome c's?

That is not a strawman version of your mechanism. That is
truly how your mechanism is supposed to work. If the gap distance
that must be crossed by pure random walk/selection grows, the time
needed to cross the gap also grows - - exponentially.

Evolution of new function only occurs in the biased subset of cases
where the gap distance is not large. There are predicted consequences
of such a model, including the existence of sequence families. Since
most evolution involves changes in regulation rather than the
invention of new functions, the size of most gaps are quite small.
After all, we have only the same number of genes as other mammals and
even nematodes and Arapidopsis.

I certainly agree that *if* the gap distance (number of mutational
steps) grows too large, that particular change will probably to almost
certainly not occur. But your assumption that gap distance is
correlated with size is wrong.

The only way
your pet theory could work is if the gap distance never grew relative
to any increased structural/sequence threshold requirements. That is
the only way your theory would have a chance. The problem is that the
gap distance does grow in relationship to increasing minimum threshold
requirements. It doesn't stay at the minimum possible distance of
"one" as you have argued many times.

But that has nothing to do with the size of the sequence. It has to
do with pre-existing functional sequences that can either be modified
in a few steps or their absence. In the former case, you will get a
new function. In the latter you won't. Most functions, of course,
are inherited from ancestors, not invented from scratch in each
species.

Again, don't keep up your strawman lie by saying that my suggested gap
distance is the maximum possible gap distance. That suggestion is a
deliberate lie on your part.

I don't. I only accuse you of, until I enter the conversation,
presenting only the total size number and implying, by not presenting
any other number like gap size, that that total size number is the
number of mutational steps. I am merely correcting your sloppy
language and, indeed, I do keep pointing out that even you recognize
that the real number you need is the gap size.

But using the average gap distance is just as bad if you don't supply
the variance from actual data. And assuming a relationship that
simply and dumbly extrapolates larger gap sizes from smaller to larger
in the absence of supporting evidence is numerology.

I have no idea why you feel the need to
keep up this deliberate misrepresentation. You must be worried that
if my true position were actually recognized for what it is, that
people might be convinced by its obvious veracity.

< snip rest of desperate obfuscation and outright lies >

Sean Pitmanwww.DetectingDesign.com

.

Relevant Pages

  • Re: Sean Pitmans Misleading Statements Regarding Bias and
    ... only listing some clues as to the likely average gap distances to make ... what makes a sequence "potentially beneficial". ... And "minimum structural threshold" (which you have defined as ... if you can come up with a significantly smaller gap distance that my ...
    (talk.origins)
  • Re: Most valuable poster
    ... nylonase or lactase evolution examples. ... residues), you have the ability, so you say, to tell us *exactly* what ... the average gap size is based on the size of the end product. ... recognizable sequence homologs or recognizable intermediate functions. ...
    (talk.origins)
  • Re: The last ancestor of all life
    ... binds, in an beneficial way, to a particular protein antigen sequence. ... antibody binding to the antigen will be rewarded with improved ... differences that produce the gap problem for evolution. ...
    (talk.origins)
  • Re: Richard says Howard Hershey is wrong
    ... cytochrome c, the only sequence you have presented actual data for, is ... It is neither an "average" nor is it "gap ... isn't "average gap size" at all. ... "sequence specificity" as cytochrome c. ...
    (talk.origins)
  • Re: The Relationship of Gaps to Thresholds
    ... other than a fair degree of sequence similarity. ... Your MATH explicitly says that the size of the gap needed to be ... is roughly 30% of the total size of the end protein. ... start at some average distance away (that distance being a function of ...
    (talk.origins)
Loading