Re: Paper: Adaptive Evolution of Conserved Noncoding Elements in Mammals

"Perplexed in Peoria" <jimmenegay@xxxxxxxxxxxxx> wrote:-

Er..., Peter? Who do you have in mind as a scientist who has called
a CNC "junk-DNA". Larry Moran is the guy posting here who I would
most associate with the "Most non-coding DNA really is junk" ideology.
But he would certainly be of the opinion that most CNCs have a
function.
That first 'C' in CNC stands for conserved. I don't think anyone
believes that CNCs are mutation cold-spots. That hypothesis has always
been included just to cover all the possibilities. It is something of
a straw man.

JE:-
Hi Jim,
I don't think mutation rates (which are just random) have much to do with
it. My understanding is that Moran et al refer to non coding DNA as
"junk"
because this type of DNA, which constitutes the majority of almost any
genome is conserved consistent with a HW distribution (binomial
distribution). Population genetics interprets a gene frequency match to
the
HW distribution providing proof of non selection _incorrectly_ inferring
non
use. Most of the genome only presents as "junk" because Moran et al
refuse
to _correct a critical oversimplification_ assumed within models of
Darwinian evolution by natural selection.

Several points of confusion here.
First, your mention of HW distributions is completely beside the point.
We are talking about genomes here, which are, by definition, haploid.
Twenty-three chromosomes, not forty-six.

JE:-
Jim,
My understanding: we are not _only_ talking about heuristic haploid "in
vitro" genomes but are necessarily including empirical "in vivo" haploid
genomes, i.e. genomes which are normally found and selected in the diploid
state and reproduced over fertile organisms cycles within nature. The non
coding part of these haploid genomes, which constitutes the vast majority of
most genomes, has to be successfully reproduced over fertile organism
generations and not just successfully replicated in a test tube.
Reproduction over these generations of the non coding part provides a
binomial distribution of it which population genetics defines to mean not
selected and therefore, of no use. It remains a _serious_ error to claim
that when something is not selected then it can only be of no use (allowing
the derogatory term "junk").


You may have meant to refer to some other statistical distribution which
would (in theory) be indicative of random mutation with no selection.
Yes, indeed, there are those kinds of statistical models in use to
determine
whether something probably was subject to selection, how recently it was
subject to selection, and how strong the selection was. Feel free to
criticise those models. They may very well not be particularly reliable.

JE:-
I was not referring to a random distribution just to a binomial
distribution. This represents a maximal distribution such that anything less
than it must have been reduced by selection. However if you employ only a
random distribution it can be empirically demonstrated that the non coding
part of a haploid genome reproduced over many fertile organism generations
drifts more than the coding part of it. Population geneticists interpret
this to mean that the non coding part is neutral and again, _incorrectly_
infer from this that it does not do anything.


But Moran's "most non-coding DNA really is junk" ideology is not based on
fancy statistics like this. It is based on something much simpler and
(one hopes) more reliable. If a segment of DNA has changed over time in
a lineage at about the rate one would expect from random mutation and
drift
without selection, then it most likely did change under the exclusive
control
of random selection and drift. But if it changes much more slowly, then
it
probably was constrained by purifying selection.

JE:-
Two points. Firstly, there is no such thing as "random selection". Such a
term only constitutes a misleading contradiction. Secondly, just because
something displays more random sampling error than something else does not
prove that it is of no use. All finite systems undergo some degres of
sampling error. Drift and mutation remain the products of a supposed
heuristic random process. Such processes are defined to only provide random
patterns whereas non random process _emprically_ provide both random and non
random patterns. Therefore, a verification of a random pattern does not
allow the conclusion that only a random process had caused it. If a non
random process produced it then the conclusion that this random pattern does
nothing may indeed be false because that pattern remains joined at the hip
to the same process which produces selectable non random patterns. It is
this _epistatic fitness join_ which has been deleted and remains as a
chronically uncorrected oversimplification within all population genetics
models. Allowing something to vary as a side effect of a selectable non
random process provides a vital function for the Darwinian fertile organism
unit of selection because it can increase/decrease the heritable
temporal/spatial variation on which selection operates. It should be self
evident that the rate of random mutation and genetic drift produced by any
NON random process can be seleted to be reduced or allowed to remain maximal
depending on which level increases fitness at the mono-centric Darwinian
fertile organism level of selection. I hasten to add that the random pattern
is not being selected just the non random process which empirically produced
it.

The CNCs show the signature of purifying selection by this very simple
test.
Extremely strong purifying selection. Hence, I doubt that Moran or anyone
else would call them junk.

JE:-
CNC's drift more than the coding part. This is sufficient to only
mathematically conclude, i.e. conclude on just a 100% RELATIVE BASIS
(conclude without any required constant frame of reference) that CNC's
remain neutral. Mathematics is not a science.

Regards,

John Edser
Independent Researcher

edser@xxxxxxxxxxxxxx


.

Relevant Pages

  • Re: linux on a old laptop
    ... it was on a 5x86/133 with 32MB RAM. ... but if you're careful about what you install and run it ... just puts all of it's limited selection on the drive. ... that I sure would have included in a distribution). ...
    (comp.os.linux.setup)
  • Re: bootstrapping in Perl
    ... Perl can easily select 100 items at random from 10000, ... distribution of the original data does not follow a gaussian ... The probability of an item being selected by randshouldn't normally ... themselves don't have any influence on the selection. ...
    (perl.beginners)
  • Re: Sex for diversity or diversity repression?
    ... distribution of two alleles at one locus within a heuristic infinite ... population without selection or sexual selection) be different within a ... and if there is no selection acting to kill off some genotypes during ... demanding that the question not be evaded, ...
    (sci.bio.evolution)
  • Re: Paper: Adaptive Evolution of Conserved Noncoding Elements in Mammals
    ... most associate with the "Most non-coding DNA really is junk" ideology. ... But he would certainly be of the opinion that most CNCs have a function. ... genome is conserved consistent with a HW distribution (binomial ... Darwinian evolution by natural selection. ...
    (sci.bio.evolution)
  • Re: Steven J introduces a "Rule of Science" which would bring Science
    ... Almost none of the atheists in the ... conceivable possible explanation for a particular event under ... Artificial selection (which is Darwin's process -- the process is the ... And we do know the rate of change in genomes, since, in the best ...
    (talk.origins)
Quantcast