Probability and proteins

Brian Miller (bjmiller@phy.duke.edu)
Mon, 4 Sep 95 22:46:07 -0400

Terry Grey wrote:

>now and say that even if there is 10^55 possible sequences that's
>still only one in 10^65 of the possible 10^110 sequences. I reject
>this argument as being a problem because it assumes that functional
>protein sequences are assembled by random processes which they are
>not and it denies the efficacy of selection. Other work from Sauer's
>lab (Davidson and Sauer) estimates that 1 in 20 random amino acid
>sequences produces a folded protein (no function here). I claim that
>fold precedes function in evolution and then with minor (random)
>tweaking, function and increased stability can be selected for.

I understand that evolutionists propose that some directed process
might form proteins in origin scenarios, but once cells are formed,
are not mutations fairly stochastic? Also, the amino acid sequences,
studied by Sauer, which did not fold properly were quickly destroyed
when produced inside bacteria. Why would one amino acid sequence which
is quickly destroyed inside a bacteria have a selective advantage over
another which is quickly destroyed. If for most proteins, only one out
of 10^65 sequences has any selective advantage for some particular
function and if mutations are not directed, Terry's objections seem
somewhat questionable. Perhaps he and Mike Behe could help clarify
these points.

Brian Miller
Dept of Physics
Duke University
bjmiller@phy.duke.edu