[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
RE: up!up!and away!
DNA that does not code for proteins is often called "junk DNA".
Proposed purposes for non-coding DNA include better attachment points
for replicases, structural uses (DNA folding), protective uses
(against mutagens), and a host of others. Because of the
difference in bond strengths between base pairs some
constraints exist which help stabilize the DNA molecule
as a whole. Obviously a great deal of DNA that is
non-coding is not really junk - it has an influence on
fitness at the local level. Real junk DNA is represented
by the stuff in which a mutation is neutral.
This is really tough stuff to deconvolve and if huge
books by real geniuses do not fully do it justice,
we are not going to do so in emails.
It suffices to say that a DNA molecule has a mathematical
representation in a Hamming space with dimension
equal to 2X the number of base pairs. A single strand of DNA
is represented by a point in this space. A mutation at one
base pair moves the point two spaces. A population is
a cloud in Sequence space. (See Eigen's Scientific American
article on quasi-species.) Gould's "model" replaces this
huge space by a one-dimensional line.
But the fit of Gould's model to reality gets worse.
When genes are duplicated during replication, the
Hamming space becomes larger- its dimensionality increases.
If the new genes are just copies of the old ones, the
information encoded is about the same. But mutations in
the new genetic material may occur over time which cause
it to "discover" a new enzyme. Each active mutation may be
tested for fitness (on a random basis). Over time the filtering
of this "noise" generates new information. The filled part
of the space is bigger. But, because phenotype space is
also involved (and fitness is most tested there) the fraction
of the huge space which contains "fit" species is small.
In addition, using only natural selection, the only
parts of space that can be explored by the "natural selection"
algorithm are those connected to the existing species.
So, Gould's model of a drunkard coming out of a bar into
an alley can be replaced by a series of drunkards coming
of the bar into an alley, which opens into a large square,
at the edge of which is an ocean,... The "force" that drives
some drunkards to stagger to the ocean is just that the space
there is bigger - it has more dimensionality. The ones that
get to the ocean may drown, but they don't stagger back
to the bar, because they can't randomly get back to the
square. The ones in the square can't find the entrance to
the alley by random walk. The fact is that many drunkards
end up in the ocean, far from the bar- in the more complex
part of town, but there are still lots of drunkards in the alley.
> From: John Bois[SMTP:email@example.com]
> Sent: Wednesday, March 24, 1999 12:57 PM
> To: Derkits, Gustav E, JR (Gus)
> Cc: 'Bill Adlam'; firstname.lastname@example.org
> Subject: RE: up!up!and away!
> On Wed, 24 Mar 1999, Derkits, Gustav E, JR (Gus) wrote:
> > I AM WRITING SHORT EMAILS. OBVIOUSLY EVOLUTION
> > HAS FEEDBACK LOOPS THAT STRONGLY FILTER THE OUTCOME.
> > HOWEVER, NOTE THAT "JUNK DNA" OR NON-CODING DNA
> > DOES EVOLVE BY RANDOM WALK.
> I just read an article talking about the hunt for a "purpose" for junk
> DNA. This search is partly due to the finding that "junk" DNA turns out
> to be fairly highly conserved.