[FRIAM] random v stochastic v indeterminate

Wed Aug 23 12:33:27 EDT 2017

Okay, that wasn't quite fair to Andreas Wagner.  He's got aspects of
neutral networks in his redundant gene networks, that is, if you duplicate
all the genes that participate in a control system, then one set can evolve
while the other set maintains the original function of the system.  This is
like Peter Shuster's work on neutral networks of RNA sequences which
preserve the RNA secondary structure while evolving the primary sequence,
or the parallel work on proteins evolving their primary sequence while
retaining the 3d structure and function, both of which got mentioned in the
Phillip Ball article that Robert linked.  Which is hardly surprising given
the time both spent at SFI.

I think this also corresponds to Gerald Edelman's idea of biological
degeneracy.  In fact, the wikipedia article
https://en.wikipedia.org/wiki/Degeneracy_(biology) makes Wagner's point
about degeneracy and robustness without referring to Wagner as a source in
the body of the article.  Hmm, there is an A. Wagner in the footnotes, but
he isn't listed as a known researcher on the topic in the article body.
Wonder who's editing this article?

This all reminds me of the underlying basis for the ensemble dynamics used
by Folding at Home to combine multiple independent short random walks.  I
remember once tracking that to a paper from LANL witten in the 50's, but I
can't find any trace of it.  The gist was that starting multiple short
random walks from a given starting point and appropriately combining
results would allow you to bias the starting point for the next bundle of
short random walks, and the net result would be a super-linear speedup over
spending the same resources on one long random walk.

So the equivalent evolutionary ratchet might be to duplicate the genetic
basis for something, creating multiple starting points for random walks;
 allow each duplicate to evolve independently for some period, creating
multiple random walks;  select against multiple copies and harvest the
survivors;  rinse;  and repeat.  But rather than starting from a single
genotype, which would be game over, you start from an existing diverse
population.  Might be done by cycles of resource abundance and starvation.

-- rec --

On Wed, Aug 23, 2017 at 11:00 AM, Roger Critchlow <rec at elf.org> wrote:

> Gracious, I'm having fits trying to construct complete arguments this
> morning.
>
> The original Darwinian mechanics of natural selection was formulated in
> complete agnosticism about where the underlying variety came from.  The
> variety existed, under the proposed mechanics it would be winnowed into
> more or less fit progeny, and somehow after the winnowing there would still
> be more variety for future rounds of winnowing.  The Neo-Darwinian
> synthesis identified Mendelian genetics as the mechanism of inheritance
> without any idea of how it worked, either, and was limited to recombination
> of alleles for their mechanics of variety.  That is a formidable source of
> variety given enough alleles on enough loci, and it's also a sustainable
> source of variety since unfit allelic combinations could still be preserved
> in the population in other combinations.  So if genotype AB was fatal, A
> and B might still persist in the genotypes AC, BC, AD, BD, and so on.  Then
> with the molecular biology of the gene we finally get a physical mechanism
> for inheritance and an embarrassment of riches for the origins of variety.
> We have, off the top of my head:
>
>    - single base pair replication errors or point mutations;
>    - insertions or deletions where the DNA being copied or the copy gets
>    looped out;
>    - cross overs where two homologous DNA strands swap ends (though we
>    had microscopic evidence for crossing over before molecular biology);
>    - chromosome duplications and losses;
>    - copy error correction mechanismss;
>    - mutators which increase copy errors;
>    - viral sequences integrating into chromososmes
>    - viral sequences disintegrating out of chromosomes
>    - mobile elements jumping from location to location in chromosomes
>    - extra chromosomal genetic materials
>    - incorporation of other organisms as organelles
>    - neutral networks of RNAs or proteins where the underlying function
>    is preserved across nucleotide or peptide substitutions
>    - differential expression under environmental variation
>
> DNA polymerase is a complex of proteins which synthesizes a new strand of
> DNA complementary in sequence to an existing strand.  It takes single
> stranded DNA and makes it into double stranded DNA according to the base
> pairing rules.  It's actually more complex than that, I believe it unwinds
> the existing double strand as goes, cutting a strand to release the
> torsion.  There are thousands active in a human cell that is duplicating
> its DNA in preparation for cell division.  Embryonic cell divisions take
> about 8 hours, DNA replication is happening at 2.08e5 nucleotides/second
> according to a google search.  My source claims that: "DNA replication is
> accomplished with an average of only 1 error per billion (109)
> nucleotides.38 <http://www.ehd.org/dev_article_unit1.php#fb38> 190?39
> <http://www.ehd.org/dev_article_unit1.php#fb39>"  (Amusing to me, that
> 190? refers to a text book of Molecular Cell Biology by an author who was a
> scientific advisor to a company where I was a founding employee.)  That
> error rate, whatever its exact magnitude and significant digits, that
> replication rate, the concentration of DNA polymerase in the embryonic
> cells, the rate of cell division, these are all subject to natural
> selection, as are the biosynthetic pathways that supply the raw materials
> for DNA replication.
>
> Right, construct complete arguments.
>
> The theory of natural selection does not explain the origin of variety,
> however it does depend on variety existing and continuing to exist.  If
> natural selection led to runaway fixation of genotypes in a population, it
> would be game over.  One would expect that the fingerprints of natural
> selection should be found everywhere that living organisms might modulate
> the origin and maintenance of variety in their populations.
>
> As for the origin of innovations, I'd say that's a value system which has
> nothing to do with natural selection, it's a moral aesthetics that has to
> do with measuring progress, curating shiny baubles from evolutionary
> history for the purposes of arguing with other curators of shiny baubles.
>  (Oops, may have tarred more than necessary with that brush.)
>
> I would guess that most arguments against evolution having had time enough
> to get anywhere depend on an assumption of one ancestral genome
> diversifying by replication with point mutation.  But natural selection
> might have begun curating a diversity of RNAs, peptides, and small molecule
> metabolites before the origin of life, and the original life may have
> incorporated a lot of existing pre-biotic variety.
>
> -- rec --
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20170823/c8338636/attachment.html>