Re: Online archy courses
From: Philip Deitiker (Donevenask_at_worlnet.att.net)
Date: 12/23/04
- Previous message: Jason Eshleman: "Re: MV is the most published author of SAP"
- In reply to: G Horvat: "Re: Online archy courses"
- Next in thread: G Horvat: "Re: Online archy courses"
- Reply: G Horvat: "Re: Online archy courses"
- Messages sorted by: [ date ] [ thread ]
Date: Thu, 23 Dec 2004 03:15:12 GMT
G Horvat <g-horvat@shaw.ca> says in
news:smfjs05pk67tdhlbbe2asijs4agds51ver@4ax.com:
> On 22 Dec 2004 15:39:29 GMT, Philip Deitiker
> <Nopdeitik@att.net.Spam> wrote:
>>>... how is it known that exo-African homo erectus did not
>>> have mtDNA sequences which were closer to those of
>>> exo-African homo sapiens than to sub-saharan homo sapiens?...
>
>>Same reason for Neandertals. Erectoids migrated out of africa
>>[or evolved out of africa] 1.6 million years ago. While you
>>may disagree that Paabo and companies estimates of mtDNA
>>coalescence are late, you cannot propose that mine are because
>>I included also in those estimates the issues of recurrent
>>mutations and its effect on estimating the MRCA. When this was
>>done the earliest possible MRCA for humans is around 320 years
>>for mtDNA. ...
>
> If mutations accumulated in a clock-like manner, then it should
> be possible to determine fairly accurate times of coalescence
> but I don't think that they did and so both your's and Paabo's
> estimates mean nothing to me (to be perfectly frank).
>
> With regards to this subject, Ingman et al. (2000) wrote:
>
> "The molecular clock hypothesis postulates that DNA sequence
> evolution is roughly constant over time in all evolutionary
> lineages. We used a test that compares the log likelihoods of
> trees reconstructed with and without the molecular clock
> assumption to examine the supposition that the mtDNA lineages
> evolve at `clock-like' rates. The human mtDNA sequences,
> excluding the D-loop, have evolved at roughly constant rates (P
> = 0:094) ..."
Evolution did not speed up because for neandertals. Do you see any
reason to suggest evolution of european mtDNA is faster than african
mtDNA by a 10 fold rate? If not why would you propose that Neandertal
a. went backwards in the direction of chimpanzee, b. sped up by ten
fold.
> So, as you mentioned in an earlier posting, they considered the
> regions in which the Neandertal sequences were obtained
> unsuitable for dating but thought the coding region portions
> were better. In a later study, Herrnstadt et al. (2002), were
> less optimistic:
>
> "The results presented here indicate that further investigation
> of these key evolutionary processes is necessary and that
> advocacy of such haplogroup-dating schemes should be tempered
> with caution. In an important analysis, Torroni et al. (2001b)
> have obtained evidence that there is not an mtDNA "clock" (i.e.,
> a constant, single rate of DNA evolution) and have provided
> preliminary evidence that different subclusters within the same
> haplogroup have evolved at different rates. Those results will
> undoubtedly catalyze further investigations."
But I explained why that is and I provided a confidence interval for
those reasons. The reason is more or less of a technical nature.
There is too much time to use transitions as a proper means of
estimation because of the issue of homoplasies, none the less
estimates using transitions and adjusting place the branch prior to
the radiation of humans, prior to the human MRCA. The level of
transversions are too few to statistically predict that date, however
nonetheless the 3 transitions and 1 deletion between humans and
neandertals place that distance as being >1 million years.
So basically for your argument to be true.
The HVR transition rate would have to had spead up for unknown
reason and no homoplasies.
The HVR transition rate would have to had spead up for unknown
reason.
The average distance between the deepest branches of humans in HVR
is 13 mutations at lHVR1 which can be rooted at the center of this
divergence with chimpanzee sequence. The average number of
transversions between the deepest branches is 0.6
The basic problem is a relative one. Neandertal human temporal
distance is such it falls in a grey area between two estimation
techniques. The only logical reason that it can fall into that range
is simple, it is too old for one technique and too young for another,
therefore it has to be outside the effective range for clocking that
we apply to humans, and too young to be effectively clocked as we did
for C/H divergence. Meaning that the split between Neandertals and
Human logically has to be between human MRCA and C/H MRCA. In fact
the human clocking parsimony places it certainly well before our
MRCA, but uncertainly relative to 2 or 5 fold times that MRCA.
Anybody can make any excuse they want and confuscate the issue. But
the reality is that there exist a human tree of more or less
parsimonious nature with regard to mtDNA. The deepest branches are in
africa, something you will agree with. When Neandertal sequence is
placed in that framework it taps the basal branch and when chimpanzee
is added as an outgroup that 'tap' joins the chimpanzee outgroup
ancestral to human and to neandertal creating a relatively balanced
split. Everyone author admits this, even though some confuscate the
issue.
You can ignore this evidence if you like, but it will not change
the nature of the evidence. I should point out that Ingman/Paabo
style averaging which takes a really good method, pairwise comparison
and then extends it beyond its use has resulted in misstatments.
Notably his own colleage found that the Andaman's likely left africa
prior to his suggested 53 kya. The same mistake is made by other
authors with regard to Neandertal.
Averaging only works real well if the population has expanded in a
more or less circular symmetrical manner. The branching between
neandertals and humans shows neither a human based expansion
similarity or symmetry. The branch legs in Neandertal are
considerably longer than the human average, about 3 times longer, and
they show no branching as is seen with all human haplotypes, except
at the end points. Therefore these average comparisons are not useful
and one MUST look for a parsimonious explanation. That explanation is
clear. To be frank if you don't understand that then you need to read
more about how evolution works.
With regard to Paabo and Ingman, There assumption is that if the
various branches are approximately the same age, then they must
represent a single migration. The flaw in that logic is that the
variation of parameters means that the relative exit times of a
branch are +/-15% relative to the age, So that exit times of 30%
difference 'really' show up as being insignificantly different in
their analysis. Their analysis fails to consider the complexity of
human migration and they do so with 50 exoafrican sequences. This
does not improve on the information gained at HVR1, however it offers
the potential with good even sample to gain better information. Have
we seen that yet though, not in my opinion.
There analysis of the coding region was unable to establish
correctly when, where humans left africa. In the end there
conclusions are not better than HVR1 comparisons. And, they do not
have genomic comparisons of any neandertals and less a handful of
chimpanzees.
And BTW, where would we most likely to see branch length
differences that could be used as a precedence? In chimpanzee and
gorilla. Who looked at this carefully, me, who else? No one, not
Paabo. What did I find, I found that branches of smaller populations,
like bonobos had shorter branches. By how much? about 30% shorter.
Have you looked at or compared any branching patterns in chimpazee?
Have you looked at any branching patterns in gorilla? Has anyone
looked at the branching patterns of genomic sequences in chimpanzee
or gorilla before they wrote a paper establishing that genomics were
far superior to use versus HVR1? No-one has done so, and especially
not Paabo. So pretty much don't tell I don't know why I am setting a
limit where I set it, until you have examined these issues yourself
as thoroughly as I have done. And don't tell what Paabo thinks is
exceptional, when he hasn't even laid the proper foundation for the
basis of his work.
How could a person possibly know what rate variation is if they
have never looked beyond a population that radiated less than 160kya.
Any variation you see for that population could simply be
statistical. It requires the examination of deeply branching
populations to see that clearly. And none of your 'heros' have done
that, neither have you.
> So, where are we right now?
We are in a position to suggest that rate varies within a range, and
that range is not 10 fold greater than our estimate. When we have
uncovered sequences that have appeared to have varied greatly in
humans, most are now considered to be errors. Therefore I stand
behind my prediction of an outside limit of 320,000 for the human
MRCA, unless someone can show me convincing data are reasoning to
suggest that mutations occurred much slower than the very generous
boundaries I have set, I see no reason to change that estimate. AND,
no reasonable person would change that estimate to 1.6 million years
without having a very, very good source of data that says so.
What proof do you have specifically that the rate of mtDNA
variation in any long lineage of humans L1, L2, L3 has on average
varied by a 10 fold rate, 5 fold rate, 3 fold rate, 2 fold rate.
You have no evidence that there is variance greater than about 1.5
fold which could be explained by statistics. Even your vuanted author
who claims a p=0.09 (Acceptable p for rejecting the null
hypothesis is P<0.05) that rate is the variable by 10 fold.
> Has anyone demonstrated that Ingman was wrong about the HV regions?
Not wrong about HV, just not right about coding. HV does what it can
do within the bounds of interpretation. HV did not get the major
african lineages wrong, did it?
> Has anyone determined that
> Torroni was wrong about the absence of "a constant, single rate
> of DNA evolution" in complete sequences?
Duh, particularly in the coding because the coding region is under
purifying selection, something I warned you about. The Coding is
there to _Add_ information to HVR, the variable rate of genomic
evolution is expected and has to be conditioned. When you get all
those coding sequences done you going to run smack into Arnason all
over again, 10 papers are going to flood the literature claiming that
2/3rds of coding mutations are useless because they are to close to
genes where purifying or region specific selection has taken place.
mtDNA clocking is not wrong, its just imperfect, for statistical
reasons, for precision reasons, for selective reasons. I have sided
with you on the genomic issue for a technical point, as mtDNA
branches approach the present there are too few mutations and
branches, too much relative error for it to be useful. The assumption
is that coding mutations will occur 4 to 5 times more frequent (per
genome) than HVR. However I am fully aware that not all coding
mutations will occur at the same rate, just as in HVR1.
-- Philip - - - - - - - - - - - - - - - - - - - - - - - - - - - ____Groups_____ Mol Anthro http://groups.yahoo.com/group/DNAanthro/ Pal Anthro http://groups.yahoo.com/group/Paleoanthro/ Arch. Aux http://groups.yahoo.com/group/sciarchauxilliary/ Gliadin Sci http://health.groups.yahoo.com/group/GliadinScience/ ____Sites_____ Mol. Evol. Hominids http://home.att.net/~DNAPaleoAnth/ Evol. of Xchrom. http://home.att.net/~DNAPaleoAnth/xlinked.htm
- Previous message: Jason Eshleman: "Re: MV is the most published author of SAP"
- In reply to: G Horvat: "Re: Online archy courses"
- Next in thread: G Horvat: "Re: Online archy courses"
- Reply: G Horvat: "Re: Online archy courses"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
