Re: Directed evolution: evidence for teleology?

From: Chris Barden <>
Date: Mon Oct 17 2005 - 10:25:39 EDT

Hi Glenn,

> Do you have a reference for this single residue paper? I would be very
> interested in seeing it.
> If so, that would demolish the anti-evolutionary probability argument.
> Almost any permutation would work. Thus I am a wee bit skeptical of that
> but would love to see it.

The paper I am thinking of is "Studies of the Molecular Mechanism of
Discrimination between cGMP and cAMP in the Allosteric Sites of the
cGMP-binding cGMP-specific Phosphodiesterase (PDE5)*"
Illarion V. Turko, Sharron H. Francis, and Jackie D. Corbin, J Biol
Chem, Vol. 274, Issue 41, 29038-29041, October 8, 1999

In that paper the researchers use site-directed mutagenesis to turn
one specific Asp residue into Asn, which causes their PDE5 mutant to
lose its ability to discern cGMP/cAMP. They provide several examples
of other related enzymes with a similar structure. When I made the
comment about the "single residue paper" I thought I remembered they
had destroyed all function; I apparently misremembered, but this paper
shows at least that a single residue can be found to play a pivotal
role in the activity of an enzyme. It's likely that if they had
turned that Asp into, say, Val, function would have been completely

> Humans are very very good at rationalization. But, if every single enzyme
> in a body was globally optimal for that body, it would argue very strongly
> against evolution. (note in your response, you reversed the logic
> (erroneously) to use lack of optimality to argue against God). That wasn't
> what I was saying at all. Indeed, lack of optimality doesn't disprove god
> but total optimality of all systems would disprove evolution because as
> currently understood, there is absolutely no way the entire sequence space
> could be naturalistically searched to ensure GLOBAL optimality. The logic
> is a one-way street, not a two way street. One can disprove evolution, one
> can't disprove God

Okay, Glenn, I see that I should have been more careful in my
response, in that it didn't directly answer your original comment. I
had asked if there could be a watermark in directed evolution. You
suggested that if all enzymes were "GLOBALLY the most efficient
enzymes possible," then that would disprove directionless evolution.
I took your statement about finding "the most efficient enzyme on a
local level of the sequence space" as meaning that you took globally
to mean the sum of all enzymes, and not a holisitic statement about
system optimality or some such. In other words, that in order to
prove that evolution was guided, one would have to demonstrate, for
each enzyme in the body, that it performed its function as efficiently
as possible. If I got that wrong, please set me straight.

You then proceeded to provide a series of probabilities for "a 150
amino-acid long protein." The mathematics, if they are correct, show
two things:

1) that unguided evolution as a process would not be expected to
produce maximum optimality, even for one protein;
2) we can't show that even one protein is maximally optimal due to the
time constraints.

I hope you can see how I might have gotten off-track given that your
argument was for ALL enzymes but you gave numbers for only ONE.
Nevertheless, I don't think there's anything wrong with that if you
believe each enzyme is a separate problem. Evolution, guided or
unguided, produced the proteins in parallel, and our attempts to
discern design for all proteins would have to be parallel as well. In
this regime, each enzyme that could be found to be optimal would be
mounting proof of design. I doubt it's necessary to show all of them,
just enough to demonstrate that unguided evolution could not have such
a high batting average.

The first part of my response was aimed at the 10^195 being awfully
high. Evolution doesn't reach its heights of optimality, peaks or
otherwise, in a random manner so neither must we. In my opinion, the
strongest statement in favor of evolution of structures versus direct,
special creation is that it could have been otherwise: if only 20
residues of a 150 amino acid protein are essential (must be that amino
acid, 1 in 20) and the rest have no effect, then something like 10^175
of the 10^195 would have more or less the same activity -- yet enzymes
that exist in multiple species have nearly identical non-functional
regions. The math would also have to include those residues which are
weakly important (must be aliphatic, for example, or 1 in 4), but it
seems that the problem you put forward is not as intractable as you
may think.

The second part of my response was geared toward the philosophical
problems of such a hypothetical approach. I apologize for rather
sloppily labeling a false result as "evidence against God's action" as
if that's what you intended. Obviously neither of us intends to make
such an argument! But it might still rear its head. Even assuming we
could unambiguously define what an enzyme's function is and what it
means to be maximally efficient -- two big if's -- we still have a
problem due to the probabilistic nature of the argument. Finding a
single enzyme which was not maximally optimal does not disprove God,
clearly, but in the sense that we have now tied God's primary action
of "directing" to his secondary action of "evolution" and subjected it
to scrutiny, some might argue that we have indeed collected evidence
against His action. We might even find that there are enough
maximally optimal enzymes to disprove unguided evolution but enough
nonoptimal ones to give all critics of natural theology new fodder for
the problem of evil. Such might argue, "evolution has been shown to
be guided, true, but why did God choose to keep our (insert beneficial
life-sustaining or disease-fighting enzyme) from being optimal? Isn't
He just exposing us to needless pain?" At which point it would become
necessary to fall back on the design possibilities I mentioned and
we've solved nothing.

Fortunately, as Terry noted, we can't actually make those if
statements about function and optimality, which is the primary reason
we can't try this scheme.

Received on Mon Oct 17 10:28:13 2005

This archive was generated by hypermail 2.1.8 : Mon Oct 17 2005 - 10:28:13 EDT