Re: Experimental basis for the Non-Beneficial Gap Problem

On Jul 10, 7:48 pm, hersheyh <hershe...@xxxxxxxxx> wrote:

You don't have it quite right. All the systems that have been
observed to evolve have had a structural threshold minimum of no more
than a few hundred fairly specified residues. The vast majority of
these real examples of evolution in action have crossed gaps of only
one or two residue differences. This true regardless of if the
minimum structural requirements were 40aa or 400aa. So, according to
your argument you would say that there is no correlation between the
size of the minimum gap distance and the structural threshold
requirement. That's where you're quite mistaken.

Why is that a mistake? It is, rather, direct evidence that there is
no correlation between "minimum structural requirement" and "gap size"
for SYSTEMS THAT HAVE EVOLVED. My hypothesis is that, for systems
that DO evolve, there is no correlation between the number of
mutational steps required and the structural threshold requirement
(size) of the end result. Rather, my claim is that proteins evolve
when there is a short pathway wrt number of mutational steps and that
is a function of what specific precursor or ancestral sequences exist
in the particular organism. And that is what the data shows.

According to your hypothesis, there should be, on average, a ten-fold
difference in "gap size" for the 40 aa and the 400 aa system. That
is, for systems that have evolved, there is no evidence to support a
model of evolution in which your hypothetical "gap size" plays a
role.

Not true. You are forgetting that you are starting with a very large
population of sequences as starting points in your genome. If you
were starting with just one 400aa sequence, there would be a lot of
difference between the likely gap sizes even at these low levels.
And, if you could keep increases the number of different starting
points as the levels of functional complexity increased, to the same
degree, the minimum possible gap distance of one would continue to be
likely.

The problem, of course, is that population sizes are limited. It is
because of this population size limitation for starting point
sequences that evolution stalls out in an exponential manner as the
one moves beyond the statistical limitations that can be overcome by
the size of a particular population.

Consider a novel system of function that only requires 40aa.

Such as the "maximum gap size" of a protein slightly larger than
cytochrome c (which was about 30 aa, according to your math, although
you mislabeled it as "average gap size" rather than what it really is:
"maximum" gap size).

The maximum possible gap size for a function with the minimum
threshold requirements of CytoC would be 80aa changes, not 30. The
maximum gap size for a 40aa threshold would likewise be 40aa changes.
The average gap size is always significantly less than the maximum
possible. And, the likely minimum gap size is related to the size of
population of starting points.

Most
bacterial genomes would have a large number of pre-existing starting
point sequences that would be within a handful of mutations of this
potentially beneficial target.

Bull***. *Some* bacteria somewhere in space/time would, by chance,
have genomes that have sequence and structure that *could* reach
*some* other structure/sequence that has a modified or different
function within a handful of mutations. Evolution does not require
that *all* bacteria have these sequences. If some bacteria only has
proteins that are, say, 40 mutational steps away from cytochrome c,
that bacteria will not evolve cytochrome c.

All bacterial colonies of, say, 100 billion individuals would have
many genetic sequences that would be within a handful of mutational
changes from a fairly specified 40aa functional threshold. You forget
that such a 40aa system is not fully specified. This reduces the
ratio of such a system to around 1 in 1e8 sequences.

Compare this with a 400aa target
system. There will be exponentially fewer starting points within the
genome that will be within the same handful of mutations from this
higher level system.

That depends on what starting points a specific organism has. No
organism has some reservoir of 400 aa random sequences that are used
as the source of all new functions. Organisms have currently
functioning 400 (and smaller and larger) aa sequences, which have
various functions and subfunctions, some of which may be useful in
some new function. For example, one does not need to change *all* the
sites involved in binding cortisol and transducing evidence of that
binding across the membrane in the evolution of an aldosterone
receptor. In fact, one does not need to change *any* of the
transmembrane subfunctions. Nor most of the aa's involved in binding
generic steroid. Duplication and divergence producing a *gene
family* of steroid receptors, all roughly the same size and with much
of the same structure, is not a process where "average gap size" (if
you could actually calculate it and not the maximal gap size) has any
relevance.

If every new function that ever arose had to arise from scratch or
from some random sequence, you could talk about "average gap size" and
actually make sense. If, for example, some organism has no proteins
that bind to any steroids and no proteins can cross membranes and
transduce a change from one side of a membrane to another, you would
have found a cell in which one would have to generate the protein from
scratch or from some random sequence. That would be a cell that
would not evolve an aldosterone receptor.

Similarly, if there were bacterial cells that completely lacked any
pores that contained extruded whips and/or completely lacked any
transmembrane motor that converted energy into mechanical motion, then
such a cell would be unable to evolve motorized rotation of a pore/
whip. And even most cells with those features may not be able to link
them easily in a few steps. But all that is required is that a cell
have those features *and* be able to link them in a few steps.
Regardless of the size of the end product. IOW, the evolution of a
function is not a function of size. It is a function of specific
structure/sequences that pre-exist in particular arrangements in
particular organisms.

See the correlation? That is why 1000aa systems don't evolve in real
time - - because, as you extrapolate this pattern while moving up the
ladder of functional complexity, the odds of any pre-existing starting
point being with striking distance of any potentially beneficial
1000aa drop exponentially.

That is based on a hypothetical relationship between "minimum
threshold size" and "gap size" that you *still* have not supported.
In fact, the actual material evidence, even you agree, says that there
is no such correlation. And you keep pretending that evolution works
by starting from some random sequence or maximally distant sequence
rather than, *when* (note the *when*) evolution happens, it is because
there are systems available in a particular organism at a particular
time that can be converted to one with a new added subfunction in
addition to all the subfunctions that already existed in the proteins.



< snip >

The minimum number is related to the average number along a Poisson
distribution. The odds that the minimum number will stay at the
minimum possible number of 1, as those like Howard Hershey suggest,
drop dramatically along a Poisson distribution as the average number
increases exponentially.

As the "average" gap size increases, assuming you could actually
calculate "average gap size* form total size, the probability that the
minimum *available* pathway at any one time involves a single step
would decrease.

That's right. That's the only important factor affecting predictive
value here.

Except that "gap size" is not a function of "minimum threshold size"
unless you think evolution starts from some random sequence maximally
distant from the teleologic endpoint.

Even you now admit, along with John Harshman, that there is a
relationship between the size of systems and the absolute number of
differences between systems. That's a fact. The likely minimum gap
distance is always smaller than the minimum structural threshold
requirement - always. It is never at the maximum possible distance -
never. I really don't know why I have to repeatedly correct you on
this idea.

But the minimum *possible* pathway will still be one,
regardless of the size of the end result.

That fact, while true, is completely irrelevant as far as predicting
the outcome of evolutionary potential is concerned. Its an irrelevant
red herring.

No. It is very relevant. It says that *when* evolution has happened,
it has done so in an organism in which there was some pre-existing
structures and functions that could be modified in a few steps.

That's true, but completely irrelevent when to comes to predicting
when this is likely to actually be the case. You have absolutely no
idea. Saying that it happens when it happens is not a scientific
statement or hypothesis. It's nothing at all. I carries with it no
predictive value whatsoever.

*When* the gap is large or larger (which is NOT a function of the size
of the end product), then evolution of the novel feature will not
happen in that organism.

You just admitted, along with John Harshman, that there is obvsiously
a relationship between the size of a protein-based system and the
absolute number of differences between this system and the next
closest beneficial system that we know of. That, my Howard, is a size-
gap relationship that is demonstrable. It's a fact. You yourself
already admitted it.

In fact, that is exactly what your
discussion of ebg and lactase shows. *When* an organism has the pre-
existing ebg, evolution of lactase of 400 aa in length is easy.

The ebg enzyme is a lactase. The bacterial colony had a pre-existing
sequence that was close to the ebg sequence - a single character
change away actually. The question is, what are the odds that such a
small gap distance would actually be found in any bacterial colony?
Well, as it turns out, the odds are pretty good for a large colony of
any bacterial type to have at least one sequence that is within a
handful of residue differences from at least one of the potential
lactase sequences in 400aa sequence space. That is why the evolution
of a lactase function is relatively "easy" - - because the odds of
finding one in fairly short order are relatively good.

*When* an organism lacks all the sequences that have the potential to
easily evolve lactase, the evolution of lactase, still 400 aa in
length, is much more difficult and may be extremely rare in most real
time experiments. That is, the evolution of lactase function is NOT
determined by the end size, NOT correlated to the end size *at all*.
It is a function of what pre-existing functional sequences exist in
particular organisms.

The evolution of lactase is extremely rare from the perspective of
most genomes in real time, not because of a very large gap, but
because even small gaps of 4 or 5 required residue changes are
uncommonly crossed in observable time. The fact of the matter is that
lots of different types of bacterial colonies are within a half-dozen
or so mutations of a lactase enzyme. Given the Poisson distribution
at this level, odds are pretty good that some bacterial colonies would
be within only one or two mutations. And, that is exactly what we see
at this level of functional complexity. This is why we see lactases,
nylonases and other such enzymatic functions evolving relatively
commonly at the level of a few hundred fairly specified residue
positions.

But what those odds are
depends on what the starting system contains and not on the size of
the end product.

The starting points are known. The target locations are not. Unknown
target locations are what introduce the random variable into the
equation and necessitate the use of odds to determine the likelihood
that any target will be within striking range of any one of the known
starting points.

Evolution works by searching the nearby sequence and structure space
of the pre-existing genome of that organism. *When* such a search
happens upon a change that is useful, it will be selected for.

Obviously true. Can you predict the "when" though? You cannot. Your
position is therefore not scientific beyond the mundane non-predictive
statement that when it happens it happens. Well duh!

*When*
such a search does not happen upon a useful change, the organism will
remain unchanged except for neutral drift.

Brilliant! Simpy brilliant! And they haven't offered you the nobel
prize for that theory yet? ; )

Hello Hershey Collective! That's not a theory. That's not even a
hypothesis. That predicts nothing.

If there is some target,
in a particular organism, that requires crossing a large functionless
gap, that target will not be reached and the organism will be
unchanged.

Why state the obvious like this? This says absolutely nothing about
when gaps are and are not likely to be crossable . . . you have
nothing.

You don't use any calculations nor do you even try to estimate the
odds that any target will be within striking distance of any one of
your known starting points. That's what makes your position
completely devoid of predictive value and therefore non-scientific.

I certainly can predict consequences of my model of evolution. I
already have. I predicted that one consequence will be that "new"
functions will have evidence of having come from older functions.

Like what? Sequence similarities? Sequence similarities say nothing
about the odds of your proposed mechanism doing the job. At best they
suggest common ancestry of some sort. They don't, however, support one
particular mechanism over another. This particular discussion
concerns the mechanism of evolution. Does the proposed mechanism have
scientific evidence to support it. So far, the best you have
presented is that it works when it can work. That's a hilarious
argument Mr. Hershey Collective. A real knee slapper!

And
if the new function has arisen in a short time frame, it will show
extensive homology with that ancestor rather than show no correlation
to any other protein.

Back to homology are we? Where is the homology extensive enough for
your mechanism of RM/NS to work beyond the 1000aa threshold?
Hmmmmmm? A high percentage homology isn't enough. What you need is a
very high absolute homology that is within one or two required residue
differences. What do you have that sort of homology beyond the 1000aa
threshold?

If the 'new' function is actually very old,
then neutral drift and subsequent losses and gains may hide the
ancestral relationship. But often, even then there will be sequence
(and even more structural) evidence of homology with proteins that
perform related but non-identical function.

Must be that all 1000aa + systems are quite old indeed then. Your
notion that practically all of the differences between systems at such
levels of complexity are neutral is also false. The differences
between different functional systems with novel types of functions are
not neutral, but function differences that are maintained by NS over
time.

Naturally, because you do not have a scientific explanation, your
explanation (magical poofing by an invisible fairy) is consistent with
anything one observes. However, what one does NOT observe is any
protein having crossed large functionless gaps. The simplest
explanation for that failure is that there are no such gaps that
evolution has ever had to cross.

That's a nonsense argument since you can't show that the gaps that
evolution supposedly crossed where actually small. You simply assume
that RM/NS did the job without any actual evidence or statistical
calculations whatsoever. That's not science Howard.

< snip >

If anyone considered that the likelihood of 1000aa protein systems
arising without precursors was an argument in favor of evolution, you
may have a point. But no-one does, so you are simply attacking a straw man.

I never said that one had to start without precursors.

That is what your numerology proclaims while you deny that it does so.

How is that Howard? Do you not understand basic mathematics? My
mathematics assume that the likely minimum gap distance is always
smaller than the maximum possible distance.

Yes. You handwave those smaller numbers into existence. You do NOT
mathematically determine them.

They are supported by the estimates of those like Yockey, Sauer, Olsen
and others - statistically and experimentally. At the very least,
your continued claim that I'm using maximally distant numbers is
simply not true. It's an outright lie at this piont.

And the only gap size you have ever
calculated has been the *maximum gap distance*, the total number of
invariant sites in a model that mathematically assumes that the
protein only has invariant and completely variant sites.

Where do you get that idea? I've never said this nor is this idea
reflected in my calculations. If it were, the gap distance for CytoC
would by 80 residue changes. That isn't the likely minimum gap
distance at all. As I've always said, both the average and likely
minimum gap distances are always less, much less, than the maximum
possible gap distance.

And you
assume that the starting point is a sequence that differs from the end
protein at every invariant site (the completely variant sites don't
matter). You mislabeled that *maximum* gap size as the "average" gap
size.

I do nothing of the sort.

< snip >

Start with
whatever precursor you want in an organism that never had the novel
function in question to begin with.

Like Lenski did? Like nylonase?

That's right . . .

There are always starting points
that are closer to a potentially beneficial functional system than the
maximum possible distance. However, the gap that remains is still too
large to cross when it comes to systems with minimum part requirements
beyond 1000aa.

What evidence do you have to support this claim?

It's not been observed to happen . . .

That is not evidence.

It most certainly is. The lack of an observation over time is actual
evidence.

It is assertion based on your lie that systems
that *have* evolved have done so by crossing some hypothetical large
neutral gap *all at once*.

I have never made that claim. That's your strawman
mischaracterization. My position is that systems that have evolved
have crossed small gap from pre-existing functional systems in the
gene pool. The observation, however, is that systems that are
actually observed to evolve in this manner are always low-level
requiring no more than a few hundred fairly specified residues at
minimum - as a minimum structural threshold. That's a real
observation. You simply cannot explain this observation. All you
say, over and over again, is that evolution happens when it happens.
How lame is that? You call that science?

Sean Pitman
www.DetectingDesign.com

.

Quantcast