Talk:Mitochondrial Eve/Archive 2

Mt-Eve's Haplogroup (mtDNA) classification
Has the ME sequence been reconstructed, so that we can specify her Haplogroup classification? Groups L0-L6 are not extinct, according to Haplogroup L0 (mtDNA). Is Eve simply L0, or could she be L1? I suspect we can't be sure, but can only assign probabilities to her classification. --Michael C. Price talk 09:23, 4 May 2009 (UTC)
 * This article has a suggested sequence for ME. Wapondaponda (talk) 21:33, 7 May 2009 (UTC)


 * A great link! Thanks.  Is it just my browser, or does Figure 1 display okay? --Michael C. Price talk 22:24, 7 May 2009 (UTC)
 * It works in explorer if you accept to run the scripts. Wapondaponda (talk) 22:28, 7 May 2009 (UTC)
 * Seems off line for the present. I'll try later.  Plenty of meat to chew over meanwhile.  Thanks again. --Michael C. Price talk 22:41, 7 May 2009 (UTC)
 * PDF version also available.--Michael C. Price talk 00:11, 8 May 2009 (UTC)

mtDNA recombination and heteroplasmy invalidate assumptions used to come up with mitochondrial Eve
In reality different pieces of the mtDNA come from different individuals so there was no mitochondrial eve.
 * Inheritance of Mitochondrial DNA Recombinants in Double-Heteroplasmic Families: Potential Implications for Phylogenetic Analysis
 * http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B8JDD-4RDBM6P-B&_user=10&_coverDate=02%2F28%2F2007&_alid=928554744&_rdoc=4&_fmt=high&_orig=mlkt&_cdi=43612&_sort=v&_st=17&_docanchor=&view=c&_ct=194&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=c8d243e036e3c6179df34a2299518405
 * Mitochondrial DNA clonality in the dock: can surveillance swing the case?
 * http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6TCY-4KXDW93-3&_user=10&_origUdi=B6VJ1-48H2VNN-3&_fmt=high&_coverDate=11%2F30%2F2006&_rdoc=1&_orig=article&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=241c04ec37381228ba4bf0ce409a75b4
 * Eve ‘n’ Steve: recombination of human mitochondrial DNA
 * http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6VJ1-4D98J1S-1&_user=10&_origUdi=B6TCY-4KXDW93-3&_fmt=high&_coverDate=11%2F01%2F2004&_rdoc=1&_orig=article&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=def3fd0267652dc2979019e4be4b0615
 * Williams, R. Sanders (2002), “Another Surprise from the Mitochondrial Genome,” New England Journal of Medicine, 347:609-611, August 22.
 * http://www.bioforensics.com/conference/mtDNA/NEJMpatmtdna.pdf

Sergiacid (talk) 04:08, 5 June 2009 (UTC)
 * Only the 2nd link has any evidence of (a single example of) paternal inheritance. The other links all relate to recombination within the maternal line. --Michael C. Price talk 13:39, 5 June 2009 (UTC)

There is probably not one unique Mitochondrial Eve
This article indicates that there is one person who is the most recent maternal ancestor of all living humans. This is doubtful since there is no total order that you can put on your ancestors. Who is more recent between your maternal and paternal grandmother? You might say the younger, but then what if one of your great grandparents was actually younger than one of your grandparents? The only kind of ordering that would make sense for ancestors is a partial order, which would allow for a tie. 138.88.139.93 (talk) 02:48, 3 July 2009 (UTC)
 * Your paternal grandmother is not a matrilineal ancestor. --Michael C. Price talk 04:40, 3 July 2009 (UTC)
 * Exactly as Price said. We are talking about a matrilineal tree, not a DAG, of descendants here for any given candidate mtEve. Even though each generation potentially spans a drastically different time frame from other generations, at the end of the day, the candidate proven to be the most recent, matrilineal ancstor of all humans today cannot possibly tie with another contemporary of hers. Fred Hsu (talk) 20:27, 3 July 2009 (UTC)
 * Check out this section. Fred Hsu (talk) 20:29, 3 July 2009 (UTC)

Reverted good faith edit by 76.16.176.166
I reverted this good faith edit by 76.16.176.166: diff. New half-baked sentences about limitations of mtDNA and multiregional evolution are valuable information, but should not be simply inserted randomly into the lead section without supporting changes in the main sections. Fred Hsu (talk) 03:59, 1 July 2009 (UTC)

Check out similar changes by the same user to Most recent common ancestor in similarly reckless manner. It was reverted swiftly as well. Fred Hsu (talk) 04:10, 1 July 2009 (UTC)

I agree. It is important we keep this article from fraying. It can be added to, but only by way of editorially sound, clean additions, not a sprinkling with factoids here and there.--dab (𒁳) 06:46, 1 July 2009 (UTC)

Did you read the source? Fred Hsu did not but he somehow get it. 76.16.176.166 (talk) 16:53, 1 July 2009 (UTC)
 * Did you read what Fred Hsu said? Perhaps you should start listening. --Michael C. Price talk 22:10, 1 July 2009 (UTC)

Illustration
I created this diagram to illustrate the Mitochondrial Eve article in Tamil Wikipedia. I still need to get different icons for the various generations showing pre-civilisation people differently. Perhaps labeling the nodes in the population and explaining the paths would help. What else can improve the understanding of the concept? If someone has better knowledge of Inkscape, please try improving it yourself. -- Sundar \talk \contribs 09:59, 24 July 2009 (UTC)
 * A diagram is a good idea. What does the black strip indicate, or is it an error?--Michael C. Price talk 10:52, 24 July 2009 (UTC)
 * The black stripe was meant to be the label for Mitochondrial Eve, but that went wrong. Will fix it in the next iteration. -- Sundar \talk \contribs 12:42, 24 July 2009 (UTC)

Removed the following section
== Date == "In 2007 Gonder et al. date the most ancient mtDNA lineage L0d to 106,000 BP. This is 20,000 years more recent than Ingman et al. (2000), who date Mitochondrial Eve to 171,500 ± 50,000 BP. Ingman et al. sampled 53 persons, 32 of whom were Africans from different regions of sub-Saharan Africa. They sequenced the complete mtDNA but excluded the evolutionarily unpredictable D-loop in the analysis. The 1987 results by Gonder et al. and Ingman et al. confirm the less precise result found originally by Cann et al. (1987), whose estimate ran to 215,000 ± 75,000 BP."

2004 Dawkins popular book, place her around 140,000 BP. Generally current dating point to more recent dates that earlier estimates. The maximum time of genetic history recorded mtDNA is estimated to 100,000 years.

This section was written by someone who profoundly misunderstands molecular genetics and paleoanthropology. First Dawkins is not an expert on molecular anthropology and his opinion is roughly meaningless. Gonder is the most valid reference to date but whomever wrote this section did not read Gonder's paper, apparently.

There are 4 references 2 early and 2 recent that are pertinent. Paabo's work (Ingman et al can be) essentially discounted since his group underclocks just about everything they do based on 'certain assumptions' "But Dr. Paabo's new report pushes back the language-related changes in FOXP2 to at least 350,000 years ago, a date that no longer supports Dr. Klein's thesis. .... Pushed by the referees of his new report to say why the old one was so wrong, Dr. Paabo told the editors of Current Biology that the calculations underlying the more recent date were ''not flawed, but rely on assumptions that are necessary but also universally known to be oversimplifications of the reality." Paabos group specific assumptions hold true for just about every paper his group publishes.

Ergo here is what is relevant.
 * 1) Cann RL, Stoneking M, Wilson AC.    "Mitochondrial DNA and human evolution.'' Nature. 1987 Jan 1-7;325(6099):31-6
 * 2) Vigilant L, Stoneking M, Harpending H, Hawkes K, Wilson AC. African populations and the evolution of human mitochondrial DNA.  Science. 1991 Sep 27;253(5027):1503-7.
 * 3) Gonder MK, Mortensen HM, Reed FA, de Sousa A, Tishkoff SA.  Whole-mtDNA genome sequence analysis of ancient African lineages. Mol Biol Evol. 2007 Mar;24(3):757-68.
 * 4) Behar DM, Villems R, Soodyall H, .... Genographic Consortium.   The dawn of human matrilineal diversity. Am J Hum Genet. 2008 May;82(5):1130-40.

TMRCAs (Time to most recent common ancestor) Our TMRCA estimate for the global mtDNA genome tree is 194.3 +- 32.55 kya page 761 second column Comparing the estimates, the 95% CI for Gonder is 129.2 to 259.4 kya (while I present this as striat interpretation, it is in fact a skewed right distribution which has a less recent low estimate and a much less recent high estimate) Canns estimate is: 140 to 290 kya (range is 150 ky versus Gonders 128+ ky
 * 1) 140,000 to 290,000 years ago. see first colum page 34.  Uses physical paleoanthropological guides
 * 2) 166,ooo to 249,000 years ago. see page 1506 second column. Calibration based on 4 to 6 million year TMRCA (from V. Sarich). Outside range stated 393,000 years using a C/H LCA of 9 my.
 * 3) We observed the origin of L0 (146.4 +- 25.1 kya) and L1 (140.4 +/- 33 kya, slightly more recent than the appears of modern humans based on the paleontological record.
 * 1) page 5. Thus the age of estimate of the L0d [sub]clade, defined by the available sequences is 101,589 +- 10,318, but it started to diverge from its sister clade L0abfk 143,654 +/- 11,111 ybp.. Page 8 First our results highlight the L0abfk split about 133,000-155,000 ybp (Figure 1) as marking a key point in Homo sapiens matrilinearl population structuring.'' As Figure 1 places the onset of 'Khosian' branch at ~164,000 years ago and the L(>0) as 185,000 years ago.

IOW other words, statistically speaking there is no significant different in either the range or the TMRCA between Gonder et al. 2007 and Cann 1987. And then dating of substructuring with Africa (i.e. the human population in no longer capable of random interbreeding within a constrict context) is older than 150,000 years in age.PB666 yap 04:29, 23 August 2009 (UTC)

Finally, Gonder et al 2007 may be less precise than Cann et al 1987. Gonder used mostly african lineages within Tanzania and adjacent regions proclaiming that: Because the D-loop is prone to problems with homoplasy and has been shown to produce unreliable gene trees..., we excluded the d-loop from all analysis shown in this study .... We included sites corresponding to basepairs 577-16023 of the CRS in our analysis. This means they used coding sequence. however later in the paper they state :''The amount of mtDNA sequence diversity ([pi]) among Africans [.00392] and Tanzanians [.00380] was more than twice that amoung non-Africasn [0.00181]. However the level of variation in Africa may be artificaily elevated to some extent by the selection of genomes for sequencing that would maximize haplogroup representation. later they state '''We detected significant departures from neutrality expectations (table 1), as measured by Takima's D, in the global data set and the pooled African and non-African data sets, but not in the Tanzanianss. D* and F* statistics revealed significant departure from neutrality in all populations.''' Point of fact I examined mtDNA sites based on the evolution of 32 other therians, and found relative to other mammals that all human mtDNA lineages are under significant selection (very high relative frequency of sites rarely/never changed in other species), with the exception 3 seqeunces in the L5 lineage (Tanzanian). Despite the departure from neutrality they indeed used the non-neutral mutations to estimate the TMRCAs and therefore mtDNA selection has altered their estimate of the TMRCA. The highest proportion of the mutations occurred in subclades and the basal LO clades and in certain clades (notably South American). Pairwise differences between L0d lineages and certain south American haplogroup A lineages had temporal differences of >2 million years unless a substantial number of coding sequence mutations were masked from the analysis. The best analysis that included HVR and some coding sequence mutations had TMRCA of 235,000 years using gonder's C/H LCA of 6 my + 0.5my. IOW the interpretation of coding sequence TMRCA is a rough estimate and the 130ky 95.4% CI stated by might fairly represent most sources of error and should not be removed from the presentation of the TMRCAs on the main page. This observation falls in line with what Behar et al. states as LO being developing a substructure, possibly southward and more arid adaptations with the Khosian branch. There is a rather large debate at present on the role of selection in mtDNA evolution, particularly the excess of coding region mutations. One camp argues this is an artifact of the way humans have migrated (however this does not explain the non-neutrality within the tropical zone of Africa), and the other is that cultural evolution (changes in foraging, adoption of farming) and regional migration have stabilized coding sequence variants that normally have been eliminated. Consequently this is not 'settled science' and once the non-neutrality problem has been addressed thoroughly then I suspect there will be new attempts to re-estimate the TMRCA. I draw in my 'original research' not to present it as a substitute but simply to add to the weight of evidence suggesting that hypervariable and some non-hypervariable regions may give as good if not better TMRCA relative to unmasked use of coding sequence. The paragraph that was removed is attacking AC Wilson's work is more or less POV and does not take WP:NPOV. Therefore it should not be returned to the main page. I believe that Gonder's work is the best published work to date, however it should not be taken without some consideration, molecular clocking based on sites known to be under significant selection, unless the selection coefficient is known, cannot but give one a biased result if one is looking for a TMRCA based on a nuetral (or quasi-neutral/variable selection in no net direction) expectation, there is uncertainty even within Gonder et. al.'s presentation of their error. Even if one can account for selection by applying selection coefficients, the application will, none the less, decrease confidence in any given date and increase the size of the confidence interval (i.e variance).PB666 yap 04:29, 23 August 2009 (UTC)

Please read the articles carefully to avoid placing fabrications and misleading information on the main-page.PB666 yap 04:29, 23 August 2009 (UTC)


 * Please sign your posts. Have restored section.  See your talk page. --Michael C. Price talk 04:14, 23 August 2009 (UTC)


 * Post was not completed, then but now it is. Also please do not return misleading or misrepresenting information to the main page, if you repeat this I will bring it to the attention of the Admins. The section removed violated NPOV and was, in essence, OR. The section can be edited on this page, and once it is sufficiently corrected we can agree as to whether it should be returned to the main page, this is as per the talk page guidelines.PB666 yap 04:29, 23 August 2009 (UTC)


 * BTW if you are so concerned about the state of the article, the entire lead of this article is inadequate, the dates give are wrong (outdated as in from 1991-1993), more importantly no error range is given for that date, the exit times from Africa are based on mtDNA (this is not stated) and archaeological dates (Skhul and Qafhez) 90 to 125 kya, Liujiang China 67+ kya, East India 75+kya, LM3 (Australia) 50+ kya (other new dates in Western Australia support this). IOW the mtDNA exit times (theoretically) are not the actual exit times and when one states these times one must also give a reasonable error, the last time I looked it was +- 30 ky. Someone who actually wrote this page should try reading the research instead of getting information from BBC News or Richard Dawkins.PB666 yap 04:42, 23 August 2009 (UTC)


 * You may well know what you're talking about. If so then improve the section.  Do not delete it.  Please observe the guidelines I posted to your talk page. --Michael C. Price talk 04:45, 23 August 2009 (UTC)
 * Yes, we need some error ranges. So insert them.  Improve don't delete.  --Michael C. Price talk 04:50, 23 August 2009 (UTC)
 * I have summarised Gonder in the article. --Michael C. Price talk 05:47, 23 August 2009 (UTC)

Date
Gonder et al. (2007) date the most ancient mtDNA lineage L0/L1 split in modern humans to 194,300 ± 32,500 years BP, and place it in East Africa, probably Tanzania. This is consistent with the earlier work of Ingman et al. (2000), who dated Mitochondrial Eve to 171,500 ± 50,000 years BP, which in turn was an update to the original work of Cann et al. (1987) , whose estimate ran to 215,000 ± 75,000 years BP.

Dawkins (2004), placed her around 140,000 BP, in the lower range of the error margin given by the genetic studies mentioned above.

Dawkins is not a credible source, the importance of the TMRCA is not described. Tishkoff does not state the TMRCA was likely Tanzania. While the section no longer violated WP:CITE it presents the informartion as WP:TRIVIA.PB666 yap

please sign your comments. I agree that it would be better to avoid quoting Dawkins here, but the other references are fine. If "Tanzania" is not in the source, just remove the "Tanzania" part, not the entire section. I do not understand what about this is supposed to fall under WP:TRIVIA. --dab (𒁳) 08:32, 24 August 2009 (UTC)
 * Tanzania is in the source. --Michael C. Price talk 10:46, 24 August 2009 (UTC)


 * My apologies on this issue, I was mentally recollecting the supplimentary materials in table S2 and I did not have a hard copy of this information. Looking that the supplimentary materials it would appear that the Behar paper supercedes this information (south Africa) and it would be uncontentious to use Tanzania (although Behar does not specifically mention Tanzania as a source for the L0d and LOk migrations). I just want to make the point here however about using national boundaries to define a point of expansion 100+ kya, it may not be such a good idea to create that misconception given the low level of typing in the following countries Angola,Democratic Republic of the Congo,Zimbabwe,Mozambique,Botswana,Malawi, and Zambia (the first two are extremely poorly typed). While after reviewing the paper and the supplimentary material still let's not emphasize the point.PB666 yap

As for the other references, the new section Coalescence time that replaced date, references to all these sources and several more (except Dawkins) were readded. I am still looking for two more references, one that refers to an effective population size of 4400 and the other that refers to an effective population size of 5000 females. Some of these articles I read more than a decade ago, so I am rely on pure memory as to these dates and they were connected to mtDNA. I have quoted Takahatas early population size. The issues regarding, Ingman I have dealt with.PB666 yap 21:38, 26 August 2009 (UTC)

Is this an Error?
From the article: "All complex animals can also trace their ancestry back to a mitochondrial MRCA, and chimps and humans share a mitochondrial MRCA, humans and Gorillas share a different mitochondrial MRCA, however the chimps-gorilla MRCA is the same ancestor as the chimp-human MRCA." this would seem to indicate that the chimps split from human/gorillas, which later separated, and that is opposite to what is on the Evolution Timeline in wikipedia, and in other sources. Skates61 (talk) 02:40, 25 August 2009 (UTC)
 * Looks like they meant to say: "is the same ancestor as the gorilla-human MRCA." --Michael C. Price talk 06:44, 25 August 2009 (UTC)


 * Statements removed, it was an error, but in reading the sentence it was found superfluous and removed.PB666 yap 21:38, 26 August 2009 (UTC)

Troubled clock?
I was in the process of fixing a typo in this section when I noticed that this section does not use an original source, but a editorial piece by Anne Gibbons, Anne frequently reports on results from meetings that often are not published. We need a primary literature source on this claim, IMHO.

"In the 1990s, scientists made a startling discovery. Mitochondrial DNA appeared to mutate faster than expected, which raised troubling questions concerning the dating of evolutionary events. The discovery was made upon comparing the noncoding mitochondrial DNA sequence of the remains of Emperor Nicholas II of Russia to that of his relatives. Using this new calibrated clock puts Mitochondrial Eve at a mere 6,000 years ago."

I have nature genetic correspondence and it would appear the Science piece contains a number of factual errors

A little background here, prior to 2003 when the issue regarding mtDNA sequencing exploded in a number of feild critical articles sequencing mtDNA had a high error rate, this was noted by Bandelt who showed star diagrams with excessive reticulation were commonly the result of sequencing errors. An associate and I did a survey of the literature looking the possible presence of gene conversion, and we found several potential examples, we posted these examples on line, and within months of the posting letters criticizing the sequences were published in the journals in which the sequences were published, at least some of these authors reported errors and corrected these. During the period from 2002 to 2004 many article corrections needed to be made on the account of detected errors.

There is a different issue, one that most people working with mtDNA were well aware of. However this issue that I worked around was in the CRS region of 16183 to 16192, its called the poly C region. If there was a T at position 16187 or 16189 then this region was considered stable, however if there were no Cs in either position, not only was it more difficult to sequence, but some studies found it shrank and expanded in individuals. Most modern studies of mtDNA ignore these sites, they might also ignore 16129, 16229, 16311 as it is known that within the hypervariable region these regions are supervariable. I should point out that there are about 10 sites and 2 poly C sites for a total of 12 sites within the 16,500 nt mitogenome that behave as such. As a matter of fact Excoffier and Wang (referenced in the main) discusses this very problem.

, concerning the authors and who said what: ''We analyzed positions 1-370 of the mtDNA CR [Control region] in 33 large maternal lineages corresponding to a total of 288 meitotic (or generational) events. No homoplasmic point mutations were found, yeilding an estimate of the mutation rate (<0.46 (99% CI = 0.0-1.52))''
 * 1) Jazin .... Stonking and Gyllensten wrote a corrospondence making the following claim:


 * 1) Parsons and Holland . . . By restricting their table of pooled data to HVII [Hypervariable region II] jazin et al omit much data on CR substitution rate. . . . .This omits mutations that occur in the hypermutable C-stretch region in HVII (did they mean HVRI?) and substitutions observed in pedigree cell lines. We can pool this data . . . 

Actually I wrote a program called the sniffer which compared sequences from 16124 to 16392 however the issue of the poly C track at 16183 was generally so bad that it had to be approached from both sides 16182 upward and 16192 downward to make sure the alignement of 16187 or 16189 was missed, even so it appears in the sequence database I used (2000 mtDNAs) that there were errors in placement of T in some samples. In some sequence samples there were as many as 6 inserted Cs and these frequently differed between the two most closely related sequences. Consequently there is no way to reconcile these sequences.

Jazin et al claim these sites are hotspots, these have been noted in different sequencings of the CRS (see Genbank versions). Ergo, no-one really involved in DNA work really waste much time with these areas. Ingman et al and also Gonder et al exclude the HVR altogether from their analysis, which I think is a bit draconian, however minimally certain sites identified by Excoffier and Wang should be masked for the best results.

My opinion here is that we should not be creating or selling mythology. The error rates they observed when divided by the entire mitogenome, if we want to discuss the results of this paper then we need a section on the hypervariable region and its early sequencing in humans.PB666 yap 23:17, 26 August 2009 (UTC)

misconceptions section causes misconceptions
The misconceptions section implies that both mitochondrial eve and the MRCA are the same number of generations removed from every human. It's not just "do any of my 8 great grandparents overlap with any of your 8 great grandparents?" You need to compare all ancestors to find an MRCA. Things like, "do any of my 8 great grandparents overlap with any of your 32 great great great grandparents?" Similarly for ME, but you only need to compare the maternal line.

Sstrazdu (talk) 17:57, 8 April 2009 (UTC)

Parenthetical statement under misconceptions
I'm putting this here because I'm unsure of the meaning behind the above comment, but perhaps it pertains to this statement, in parentheses: "(albeit between any two individuals it may not be the same ancestor)." I do not think that is true. MRCA means the most recent discrete individual where, mathematically, it becomes certain that this individual is in every family tree at some point in time. I've heard different estimates--2,000 years, 5,000 years. But that's not important. The important thing is, it would, in fact, be the same individual--i.e., we could say with nth percent certainty that one individual was alive by this time who will show up in every family tree. I would move to omit the parenthetical qualifier, then, unless I misunderstand it, but I'm not an editor on this page, so I'll leave it alone, and instead raise the concern.65.117.234.99 (talk) 15:08, 22 September 2009 (UTC)

Technical detail
The level of technical detail in the current version appears to much too much for the general audience per Make technical articles accessible. Pdeitiker, I believe has doubts about Mitochondrial Eve, but I don't see significant doubt among recent mainstream publications. Should this article therefore have significant portions devoted to doubts about the existence of an mtDNA MRCA. My impression was that many of these doubts were dealt with in the 1990s. Wapondaponda (talk) 06:44, 13 September 2009 (UTC)
 * Pdeitiker's general writing style far too obtuse and obscure. Article needs a serious cleanup. --Michael C. Price talk 18:29, 13 September 2009 (UTC)
 * I agree. I think it is great that this article is being enhanced. And I've watched changes over these past few weeks. But I think it is becoming a dissertation that is mostly inaccessible and unappealing to the general pubilc. Fred Hsu (talk) 21:25, 13 September 2009 (UTC). Look at the current version of the article.
 * Why are following sections even in this article at all? I am not saying they don't describe useful information. I am saying I can't see why this type of detail should be in this article.
 * Common approaches to interpreting Coalescence
 * Dependency of rate determinations of interspecific LCA anchoring
 * What is the second paragraph of the Establishing mutation rates trying to convey? It is completely unclear to me. Perhaps I am slow. Can someone explain to me what this paragraph is saying please? I see some facts, but what are the implications of these?
 * Does Both methods have major pitfalls mean mtDNA mutation analysis is useless?
 * Is On the other hand a code word for the second method? Or am I not searching hard enough for the description of second method's major pitfalls?
 * Certain sites on MtDNA... are not used ... by many researchers, and this implies what?
 * There are three approaches that have been used with mtDNA to calculated the TMRCA. What does this have to do with this paragraph?

Many questions here, I will address them one by one. First I am glad too see that there is interest in fixing this article, yes it is too technical, but I was going over some of the old literature there are a number of outstanding issues and I think I want to get this right.


 * 1) . It is unfortunate that I undertook upgrading this article at the current time. White has weighed in offering that the chimp human LCA is probably greater than 7 mya, between 7 and 10 mya. There has been no contest of his logic, or the feasibility of pushing the orangutan/human LCA by at least 2 million years. The orangutan/Human LCA was one of the primary reasons the C/H LCA of 6 mya was favored (Gonder et al), as it was set at 14 million years for genomic DNA this result in a C/H LCA of about 6 million years. White claims it may be much older. 6 million years in an of itself is the outside range for Cann et al. and Vigilante et al. So that the movement of the C/H LCA from ~5 to >7 Ma is a significant movement. I probably should have left the large errors on this page alone until the dust is settled on two major outstanding issues.
 * 2) . Presgrave and Ye have weighed into the arguement of Nuclear DNA evolution, they claim that nuclear DNA is evolving faster in Chimpanzee because of the rate of cell division in the chimpanzee testes places the evolution about 33% faster for Autosomes and ~90% faster for Y chromosome. I have gone over all the literature to examine to see if this finding can be supported, it can, which places the molecular clocking in Y chromosomal (times) used throughtout Wikipedia in trouble if the rate of Y is based on a Chimpanzee human anchor, this has to be compounded by the fact the LCA may be million(s) of years older than the LCA used to determine the mutation rate.
 * 3) . If this finding proves to be true then the TMRCAs for X-linked will be slightly older than presented. The population size stipulated by Takahate and Shaffner will be slightly larger. Since mtDNA lineages spend all their time in females, they are uneffected by sperm cell divisions and as a consequence the mtDNA DNA TMRCA does not move. This places the mtDNA at the margin of significant different with regard to X-linked predicted effective female population size. This can be repaired if female:male ratio is dropped from 2:1 to 1.5:1 however this places Y chromosomal TMRCA at about 3 to 4 SD range unless it is reclocked.
 * 4) . Certain sites are masked. Sites like 16129, 16183. This issue was brought up a decade ago when someone used these sites to suggest all humans evolved 6000 years ago, even though at the time many researchers were already masking these sites. Both Ingman and Gonder masked all the HVR sites, because of their tendency toward reticulations (star diagrams in which the peripheral branches merge back into each other). In addition some sites should be masked, these sites have shown no mutation in any sequenced genome within mammalian history (which branch lengths, cumulatively, is more than 1 billion years) IOW the rate at these sites is effectively zero, so that mutations that occur at these sites generally are retained for selective reasons or strong genetic drift as a result of isolation events.
 * 5) . Usefulness is not a concern. If usefulness is the major issue then Y chromosomal clocking does not exist. The idea that is being bounced back and forth is how biasing various forces are in the determination of branch times. Let me refocuse the issue. The age of mitochondrial eve has no value in and of itself, it is stated correctly that mtDNA eve lived in a popultion with 1000s of other individuals. In addition it is not at the point that eve lived that is of interest, the point of interest is when the population went from one state (static) to a second state (expanding). If this is not stated clearly in the article the article has virtually no value. This occurred sometime after mtDNA even and is an issue of great debate. As the article states, Ingman looked at selection in 2000 and concluded that selection was not a predominant feature. Gonder looks at selection with a large sample size and sees evidence of selection everywhere. There are studies that argue that there has been no expansion within Africa (despite the fact Africa size is 6 magnitudes greater) and 4 papers of recent have been published that attempt to explain and deal with the issue of selection, the latest of which basically masks sites that exhibit in humans changes that have been extremely rare over evolutionary history. IOW, it through recognizing the errors in various methods, methods can be corrected, HVR alone gave a TMRCA of 161,000 (5my CH/LCA) (210 ka at 6.5 ma CH LCA) and compared to Gonders 194 ka only slightly older, again that variance is small compared to uncertainty about the C/H LCA, which has moved from 4 to 6 Ma to 6 to 10 Ma, and random factors that affect coalescence (namely the 50 or so most basal mutations in the mtDNA tree). However Ingman masked all HVR sites, some HVR sites do not show patterns of reticulation, Both Gonder and Ingman included sites that could not be effectively clocked because they have mutated 0,1, or 2 times in all of animal evolution. IOW there may be a compromize strategy for dealing with the problem areas of the two techniques. . . . . . Let me reiterate the point the TMRCA is not that important, what does it really matter if the population size was 4400 individuals or 2500 individuals for a period of 50,000 years, 150,000 years ago? The important point is when population size shifted, and within the selection data for various regions of Africa is buried what happened and when. The basic problem is if you can't clock the sites that define selection, and there is a reoccurrance of such similar sites, then how does one determine accurately branch times. That we see a difference between Ingman and Gonder indicates that our measures of selection and thus population growth are very sensitive to sampling parameters.
 * 6) . The mtDNA evidence and these refined approaches are placing pressure on physical anthropologist to define events in Africa that explain the molecular phyletics, as we can see in the recent issue of PNAS on 'Out of Africa' however given the problem above in determing when people were limited to certain areas, it promises a period of contentious debate about the peopling of africa and the classification of fossil remains.PB666 yap 06:31, 18 October 2009 (UTC)
 * 7) . I came across this article about 2 months ago while trying to find a references and it was top-loaded with inaccurate information, which for some reason at that time did not seem to bother anyone, but now that the article has a rather full coverage of the literature and the various points of view (balanced) a. The TMRCA stated by gonder was incorrectly placed here, someone wrote there was no bottleneck which is not clear at yet if there was or was not a constriction, someone offered up that the TMRCA was 6000 years ago, misrepresenting the article basically arguing that some sites are unclockable (and should be masked), and all kinds of wives tales, it is no wonder the article was removed from the GA list. I did ask that the article be fixed, and yet no-one wanted to fix it. So now are we going to be serious about fixing the article?PB666 yap 06:31, 18 October 2009 (UTC)
 * 8) . I will go over the article tommorrow and try to condense and reword.

Couple of notes: The whole human genetic history project is a shamble, I have redone the Banner box for HGH which now includes a place where comments are to be made, if you are interested in making this page a GA or FA for wikipedia, please join the HGH and enter comments about what needs to be done to the article to bring it back to GA status. In addition I have created the Portal:Molecular Anthropology, the feature article is this page, we need to sit down and discus which articles we need to push to GA status so that the HGH project will at least have some representation in the Featured article list. A number of members of the HGH project are inactive and have been so for more than 1 year, and we have some new members who contribute to articles, but what we really need right now is contributers that can improve articles making allowing them to progress up the wiki totem-pole. Mitochondrial Eve is a core HGH article, it is our (plural) responsibility to get it back in shape.Talk:Mitochondrial_Eve/Comments

Removing the following
"Homo sapiens is assumed to have speciated from Homo heidelbergensis in the period of 200–160 kya. The fact that Mitochondrial Eve happens to be dated to precisely this period has been taken as evidence of a population bottleneck (e.g. Toba catastrophe theory) giving rise to the human species. There are, however, many ways such family trees can be constructed. A tree can be constructed based on any gene, not just the mitochondrial DNA.  When different such trees including the mtDNA tree are compared, no population bottleneck is found because different trees show different coalescent points.  The inconsistencies between coalescent points indicate that there had been numerous gene interchanges between population groups around the world, even after the first exodus out of Africa.  This idea forms the basis of Alan Templeton's 'Out of Africa Again and Again' theory."

"The Mitochondrial DNA provides another support for the Out of Africa hypothesis in the form of gene diversity. One finding not subject to interpretation is that the greatest diversity of mitochondrial DNA sequences exists among Africans. This diversity is believed to have accumulated because humans have been living longer in Africa than anywhere. Family trees (or 'phylogenies') constructed on the basis of mitochondrial DNA comparisons show that the living humans whose mitochondrial lineages branched earliest from the tree (L0) are prevalent among the  San and the Mbuti people. The subsequent branches of   L1, L2 and L3  are also largely confined to Africa, while only the macrogroups M and N, descended from L3,  participated in the ancient migration(s) out of Africa."

I am removing because most is redundant in the article as currently rewritten. What is not redundant is obsolete. The TMRCAs that are published with X-linked and Autosomal are consistent with a population size of 11,000 individual over the last 700,000 years, however they are tolerant of oscillations. In addition Templeton argued from the perspective of MREH, which is all but a dead hypothesis. His theory is detailed and critiqued by Takahata in 1999, at which time Takahata concluded that less than 12% of human DNA could have been from early are repeated migrations from Africa. More recently the Neandertal genome sequencing project revealed that very little if any DNA was exchanged between neandertals and humans, two of the more closely related late hominid species.PB666 yap 21:58, 5 November 2009 (UTC)

World Map
There was a map added to this page, however it does not show L0.PB666 yap 06:03, 1 December 2009 (UTC)

Off-track
I haven't paid attention to what's been going on in this article, but it appears to have undergone a total transformation. The previous version was not necessarily great, but the current version also has numerous issues. Firstly it is laden with technical detail that I believe make the article inaccessible to the average reader. Secondly, a lot of the information contained in the article is not specifically about Mitochondrial Eve, but instead is related to the genomics of Mitochondrial DNA and is better placed in articles such as Mitochondrial DNA and Human mitochondrial DNA haplogroups rather than in the article about Mitochondrial Eve. Wapondaponda (talk) 04:35, 27 November 2009 (UTC)
 * Yes, this is a concern of a lot of editors. --Michael C. Price talk 05:12, 27 November 2009 (UTC)


 * Several areas are going to be moved off page. Starting tonight I am moving the CHLCA section to its page.PB666 yap 05:40, 1 December 2009 (UTC)
 * That is probably a good idea, though I have only heard of the term CHLCA on wikipedia. Maybe something like "the evolution of human mitochondrial DNA" would be a more appropriate article for much of the content in this article. Wapondaponda (talk) 19:21, 2 December 2009 (UTC)
 * Reverted to August version. I think Pdeitiker should seek consensus for his edits. They are either not directly related to mtEve or just plain original research. Wapondaponda (talk) 20:31, 5 December 2009 (UTC)
 * You reverted to a version with many false and inaccurate statements. If you want to propose a change of this nature WM you should at least make an effort to remove the dozen or so false or highly inaccurate statements. According to the GA review, this article was demoted because it contained too much population science and not enough facts. I completely agree with the reviewer, how do you feel about the review MW?PB666 yap 20:52, 5 December 2009 (UTC)


 * Muntawandi, for your point of view the article is off track now, but somehow it was not off-track when it was demoted. Here are some possible reasonse it was demoted. Not only did it loose FA status, but it also lost GA status in one fell swoop.


 * "Mitochondrial Eve is believed to have lived about 170,000 years ago, or roughly 8,000 generations ago.
 * Misleading, age estimates of papers within the last 3 years range from 108,000 years ago to 198,000 years ago. for Gonder et al that would be 194,000 years + 2S.D. or 258,000 years.
 * "This places her shortly after the speciation of Homo sapiens sapiens and the emergence of the first anatomically modern humans.
 * Speculation. It is unknown when Homo sapiens sapiens evolved.
 * "Mitochondrial Eve lived significantly earlier than the out of Africa migration which occurred some 60,000 years ago.
 * Speculation. Some argue the out-africa migration occurred 45,000 years ago and some argue it occurred 95,000 years ago.
 * “Gonder et al. (2007)[1] date the most ancient mtDNA lineage L0d to 106,000 BP. This is consistent with Ingman et al. (2000)[2], who date Mitochondrial Eve to 171,500 ± 50,000 BP. Ingman et al. sampled 53 persons, 32 of whom were Africans from different regions of sub-Saharan Africa. They sequenced the complete mtDNA but excluded the evolutionarily unpredictable D-loop in the analysis.”
 * False and irrelevant, The L0d branch point is significantly older and it is not the oldest lineage, the oldest lineages are L0 and 'L1', the basal lineages. Gonder specifically stated the age at 194.3 +/-32 ky.
 * “Probabilistic studies[4], place her around 140,000 BP, in the lower range of the error margin given by both genetic studies mentioned above.”
 * False, it is not clearly when mtEve lived, the best and most comprehensive studies have a wide margin of error. For gonder et al. the 95%CI is 128,000 years in width. If we assume all possible ranges the TMRCA is from about 70,000 to 260,000 years.
 * "Mitochondrial Eve is estimated to have lived around 120,000 years ago at the latest."
 * False. In the Review “evalaution the mitochondrial timescale of human evolution" Endicott et al placed favored MRCA for humans at 108,000 years. I don't believe this date but for a NPOV this is minimally a low date, not the lowest given variance.
 * "The theoretical MRCA could have lived as recently as 3,000 years ago.[9]"
 * Theoretically pigs could grow wings and fly. No current publication is arguing that, that statement is based on the sole use of 5 or so sites on mtDNA that undergo hypermuation. This statement from the popular media (information we were supposed not to rely on) is based on the fact that these sites undergo saturation quickly.
 * "The most recent common ancestor (MRCA) is the most recent person whom all of humanity can count as one of their ancestors."
 * False. That ancestor is unknown, but likely much more recent. Speculation on the last person for whom all humans have an ancestor is beyong the scope of Wikipedia.
 * "'Adam' lived only 60,000 years ago.[4]"
 * False, there is much conflict over when Y-Adam lived, much more so than Eve. Early reported estimates were from 20,000 years ago, and several more recent attempts to correct for molecular clocking variation have estimated dates over 150,000 year. More than likely, if the new calibration techniques for Y based on Autosomes and X-linked comparions hold, on can expect this to be much, much older.
 * "Using this new calibrated clock puts Mitochondrial Eve at a mere 6,000 years ago. [17]"
 * False and obsolete, it was a technique for examination of errors. Most current studies mask those sites because of the high rate of saturation. In Soares et al. (2009) 16182C, 16183C and 16194C were excluded in addition for site 16519 they were unable, with 2000 mitogenomes to estimate the rate. Gonder and Ingman excluded all HVR. Because of the high rate of saturation at these sites estimation of rates is difficult with out a large number of sequences. And it is only possible with a very large number of sequences to estimate TMRCA via parsimony analysis, would would be useless for analysis inside of Africa.
 * "The fact that Mitochondrial Eve happens to be dated to precisely this period has been taken as evidence of a population bottleneck (e.g. Toba catastrophe theory) giving rise to the human species. There are, however, many ways such family trees can be constructed."
 * False. In fact several recent papers, including Endicott et al. and Atkinson are arguing for just that, there was a bottleneck. In fact Atkinson goes on to define how constrictive that it was. The modern idea of Bottleneck, as recently presented in PBS Nova series is that Africa underwent a drying period after 160,000 years ago, causing humans to become trapped around the few supportable pockets in South Africa. Toba volcano is generally considered too late to have caused a bottleneck. In addition Atkinson recently defines the size of the population, it was generally smaller than the X-linked and autosomal prediction.
 * Muntawandi, in your desire to correct this page how are you going to fix the hideous number of errors and remain within the guidelines of 'less popular science'? This page got to this point because it reflected the very worst in the Popular science presented, and intermingled within that some out an out false statements. Do you support keeping false statements on this page? Show me what your version is to correct these problems?
 * When I came across this page several months ago I was shocked that a FA on wikipedia could have so many errors, I brought this to the attention of Micheal, most of these errors that were obvious, some which I have since confirmed to mis represent the current literature, the statement was fix it yourself. Muntawandi you have had 4 months to fix these errors or come up with you own solution, are you saying now that you want to work on and improve the article. You have chosen to criticize what I have written, which means you either like the previous (flunked) version, or you simply don't mind pages with one contradition after another, and many false statements.?PB666 yap 21:42, 5 December 2009 (UTC)

Original research noticeboard
I have posted a thread that involves this article here. Wapondaponda (talk) 08:18, 6 December 2009 (UTC)


 * Muntawandi has criticized the range of 70,000 to 270000 years, I replaced this with a sentence, more difficult to read but is a dry rendition of two papers "The time of Mitochondrial Eve (TMRCA)is currently disputed, the early estimate by Gonder et al. (2007) places the at 194,000 +/- 32,000 years while Endicott & Ho (2008) place the estimate age between 82,000 and 134,000 years using a different method of calibration."PB666 yap 21:48, 6 December 2009 (UTC)
 * That sentence bears no resemblance to 70-270,000 years. Where did you get those numbers from? Fences  &amp;  Windows  02:57, 7 December 2009 (UTC)
 * What is frequently quoted in the literature are point estimates that range from 150-200kya. Most of the recent dates have clustered around 200,000 years ago. These dates have remained remarkably consistent since the story was first published 22 years ago. Quoting a range of 270-70kya, gives a misimpression that there is a lot of uncertainty in the dating. Rather they are a few outliers. Endicott 2009 quote 108,000 years, but this is only mentioned in the images, and not in the text of the article. The aim of the Endicott article was not to determine the mtDNA MRCA, but rather to highlight the challenges in in dating mtDNA splits. As a result, I wouldn't give too much credence to Endicott's estimate, because it doesn't appear to have been a very detailed analysis( initial impression). Wapondaponda (talk) 07:01, 7 December 2009 (UTC)
 * OK, folks, this means that you have not been following the literature lately. Endicott and Ho has two sources of error, first, it is unlikely that the first modern humans to have left Africa left before 55,000 years ago, the problem is that even people like Tishkoff, the leading expert on African studies does not believe they migrated earlier than 93,000 (portraying this migration into the Levant as something that eventually retracted into Africa). The earliest evidence outside of Africa, it is contested at 68,000 years of age. Point estimates arguable have no meaning only confidence intervals have any meaning. The confidence interval range for all published studies from 2000 to Present excluding E&H is the exact same as the 96CI of Ingman 71,500 to 271,500 years (i.e. 175000 +/- 50000) this covers all the known confidence intervals of the last 10 years. I thought this was the best compromise but Muntawandi complained.

Before describing the second source of error, let me just state that the problem Endicott and Ho state is exactly in tune with my personal observation, that neither Soares et al. or Mishmar et al. have completely dealt with the issues of selection and adaptive evolution. As a consequence both are slighlty overestimating the temporal length of branches that are under the age ot 40,000 years.

Since Endicott and Ho Anchor fix the intermediate branch points at 50 ky they force the extended outer branches to contract, and this corrects the problem above, however by doing this the overcompensate for the because as Soares points out, the number of mutations that persist beyond 50,000 years persists.

How bad is the problem, I cannot give the details but lets just say that some branch times using an entirely external rate calibration are specifically lopsided in humans and present TMRCAs in the millions of years. The problem is more severe than what Soares et al and Mishmar et al. bring to the table, the is the reason that these authors sent me a copy of their paper, in one aspect it is rather easy to prove that Endicott and Ho are correct. The problem is that for both Soares and Mishmar we don't know what the CHLCA is. That is the key problem, if it is one million years older than what Soares used, and given that he has compensate for many of these mutations, and if we exclude some of the most problematic trees, then the Soares 192,000 is probably correct.

We can also look at this from the Endicott and Ho, if 108/50 * 80 = 172,500 years and considering they overcompensated the again its going to set right on 190,000 years. But again we can't do this. So here is the basic problem
 * Endicott and Ho have identified a serious problem in the estimates that no other authors (except Cann et al 1987) have dealt with, however their method has its own problems.
 * Soares et al have partially compensated for this serious problem, but they don't know what the CHLCA is and their method has not compensated for adaptive selection that is significantly evident in some parts of the world.
 * Mishmar et al. have partially compenstate for the problem Endicott identified but they do not perform the type of compenstations that Soares have applied. They are using an even younger CHLCA, and their confidence interval is based on a fixed CHLCA.

Again we cannot perform WP:CRYSTAL. If the CHLCA is 7 or 6.5 million years in age then
 * TMRCA is lower than Soares, since Soares gives no confidence interval, any level of difference is significantly lower.
 * Mishmar gives a confidence range of 76,000 years (4 * 19 ka) Therefore if the TMCRA drops below 160,000 years then there is a significant error in their calculation
 * White et al basically argues that CHLCA is open ended because of errors made in physical anthropology. IOW CHLACA is greater than 7 million years, greater than the CHLCA used by Ingman, Soares, Gonder, Mishmar, Kivilsid, etc.
 * IOW all of the above appears to have serious errors in their premises. In this instance the best choice by WP standards is to apply the ranges that cover the most possible points of errors and reference them. So sorry folks but that is what Dweller stated and thats what its going to have to be for now.

There are only two current studies that produce TMRCA ranges that are credible, despite numerous internal errors. Ingman et al 2000 171+/- 50 ka= 71 to 271 ka at 96%CI (What was on the page) Gonder et al. 2007 194.3 +/-32.5 ka = 130 to 259 ka At the moment I would choose if I had to choose one Ingman et al 2000 as the range of highest confidence. Read the original sentence it spoke about confidence, not about point estimates, point estimates are meaningless.

Again there is an intense debate currently between different factions regarding the nature and extent of selection at the moment. I can tell you where I stand - by Nature of problem I stand with Endicott and Ho, by method of repair I stand between Soares and Mishmar, by time of CHLCA I stand between Soares and White.PB666 yap 15:29, 7 December 2009 (UTC)
 * tl;dr. However, "OK, folks, this means that you have not been following the literature lately." I shouldn't need to read the literature to read this article. That you want to put in statements you can't source directly to published articles shows the problem with your approach. Stop it. Leave synthesis of the literature to review articles. Fences  &amp;  Windows  17:44, 7 December 2009 (UTC)


 * OK then the review article is Endicott, Ho, Metspale, and Stringer 2008. And is exactly the reason why I altered the lede estimate. There are four pertinent reviews at the moment, 2 by Tishkoff and one stated above, in additoin Atkinson et al on population size is a major consideration. EHM&S intercompare different estimates and they come to the conclusion that the archaeologically based estimates are the best because of the problem in the peripheral branches has dsitored CHLCA anchored estimates.


 * The problem was that this page started with Popular science which drifted into misrepresentations. WP:TLDR POV in editores of the page a problem on the previous page, literature was too long and technical and people did not read it and consequently this article was kicked from a FA to a C-class page. Now the page is too long and technical, so that issue needs now to be readability. I agree with most of the Wiki-guides, however with some aspects pertaining to the literature WP:OR is problematic because some papers either understate confidence or state no confidence (Soares et al. 2009) as a result following WP:OR creates problems of itself. The range I previously gave is a standard interprestation of the 95%CI for 171.5 +/-50. Since this is contested, then I think its better to present the full range of most recent dates.

PB666 yap 18:20, 7 December 2009 (UTC)

Problems with TMRCA and meaning of MtDNA Eve
The primary derivative result of the TMRCA is the population size estimates. We started out with HVR and population size estimates were given because the number of small HVR sequences propogated rapidly, by 1997 I had my hands on 2000 of these sequences. With these sequences there was the popular notion of a big-bang, an explosion of sequences about 90,000 years ago. With this there was a growing body of evidence for small population bottleneck somewhere in SSA dating to the period between 150 and 90 kya.

Constrasting evidence for big-bang'
When the big switch occurred in 2000 to genomic DNA the number of sequences people used dropped to 50 and then rose steadily to 100. Ingman et al observed no evidence of selection, which is frequently observed when populations expand rapidly. The problem was their sample size was too small to observe these things, and he did not detect the inadequacy. As a consequence the evidence for the big bang retracted and disappeared, and we began to see evidence of more recent exodus from Africa, before the exodus there was no expansion evident. The problem is basic to the rest of the discussion, evidence of selection and saturation is very difficult to detect, it is an ignorance is bliss situation. Saturation becomes more visible as the number of sequences increases. Selection in some lineages may be evident, but really becomes evident in contrast with older clades and branch points, particularly in comparisons between other species. As the number of lineages increases the number of positions that can be intercompared increases at different depths and the significance of the difference become more apparent.

Recent studies into populatin structure - Bottleneck
Since 2001 there is now about 2000 sequences and there has been a proliferation of evidence for an expansion of the L3 lineages about 80 to 100,000 years ago. (see Atkinson, Drummand, ...) Consequently after waning support for population bottleneck idea the support has returned. The loss of support for the population bottleneck was largely due to the failed recognition of inadequate sampling. As sampling increased so has the evidence for significant levels of selection acting all throughout the mtDNA tree, only one lineage L5 does not show signficant evidence of selection, and Tishkoff places the bearers, the Sandawe, very close to the epicenter of human evolution. In some of the most extreme examples shown in a recent critique of migrations to the new world (as discussed in Endicott, ....Stinger, we (i.e. many scientist) see profound regional selection for peoples who moved from the tropics to the Arctic and back to the american Artic within the last 50,000 years. For this reason some authors were stating that humans migrated to the Americas 40 and 60kya. However with close inspection the sites they were clocking were very rarely mutated in human, primate or therian evolution. This is the major point of Endicott and Ho and I absolutely agree with them on the extent of the problem at this level or 'depth' within the canopy of the mtDNA family tree.

Problems in the estimate of population size - current
So what is the problem - The problem is that when CHLCA increases and more selection is detected and compensated for in the peripheral branches then that length between the big-bang and the TMRCA increases. As the TMRCA increases the population size increases. This is important and pertains to theories regarding speciation of humans and the environment of sub-Saharan Africa, within the time frame between 70,000 and 190,000 years. In these theories MtMRCA provides a key understanding (its population size, its place(s) and its length).

In addition the time of the so-called 'Khoisan' branch ('L0k/L0d'/'Else' fissure detected by Tishkoff and Behar) will drop from 144,000 years. Remember that this fissure occurred before the big bang, and split two groups by geography and prevented the fixation of L1. So that we can expect in the very near future some discussion about the climate of Southeastern Africa that resulted in a fissure forming in the human population, these are going to be directed at the mtDNA and more than probably at the origin of Y-DNA.

Unfortunatley, these two aspects are the most important aspects regarding the calculation of the mtDNA eve. Negating issues of selection and calibration, the timing of mtDNA TMRCA is largely determined by the size of the population, the length of the population at that its early size (both the time estimate and in variance) within the bottleneck + the time since the bottleneck to the presence (provided that selection is compensated for and calibration is correct). The time since the bottleneck is more or less a spacer, a block of time that has little or no bearing on the fixed calender date in which Eve existed, and branching within that block should suffice for a near exact estimate of the block size. As a result, the variance of the TMRCA of Eve under these ideal circumstances should follow the evidence of branching within L2, L1, L0, L4, L5 trees back (See book intitled "The Coalescent") to the mtDNA MRCA should be a reflection of the variance in those subsets of estimates.

The Khoisan fissure of a rare component (L0) within the East Africa group effectively prevented fixation of L1 as population size was at a minimum, given that Tiskoff states that L0d and L0k were removed from the East African population, this leaves only the L0abf detected at very low levels in africa, L0 might have been excluded had it not been for this split and was on the brink of being excluded. Almost all the L0a and f lineages are extremely long and lack early ancient branch points, they have survived as a result of population expansion in Africa.

The tables I have added to the Main attempt to compensate for the broader aspects, however to go any further, to deal with selection within the context of a changing molecular clock, as indicated by Endicott,et al. and Soares et al. (I will provide quotes from these papers later) would be original research.

The major point however here is that the previous Main had made assumptions about things that were due to poor techniques, or poor usage of the popular or news literature. Recent work, although not perfect is now at the threshold of dealing with these various problems more effectively (Endicott an Ho are one of the groups trying to tackle this issue), it also demonstrates clearly the critique from the GA review, this page should not be using popular science as its backbone. Endicott and Ho are not so important in their estimates of CHLCA, but are important because they bring to the table a perspective that shines light on the trouble that is apparent in the canopy and younger branches of the mtDNA tree, this trouble indicate shifts in branch positions will occur in the future and so this page should stand by the current breadth of estimates so that we are not pandering to a 'popular' point of view, or views that are subject to rapid fluctuations in the mtDNA tree. In addition it is Endicott, Ho, Metspala and Stringer is a review, as with Tishkoffs two reviews should take priority on this page.

Conclusion
The most important aspect of the TMRCA of mtDNA Eve is the population size and geographic placement that facilitated 'fixation' of her descendant lineages (This issue was discussed in the lastest series of Human origins on PBS-Nova). More simply, structure that gave rise to our observed coalescent are important.

The TMRCA is not likely to change that much, however the population size growth, as a function of branching of certain L2 subclades (namely L3) may change. Nodes of branches are subject to considerable shift as positions of nodes approach the leaves of the mtDNA family tree canopy. If this shift does occur, based on Atkinson et al. the early population size estimates may change, probably increase, as the T L3-node pulls away from the TMRCA in the direction. The shift in the tree is a balance between two factors, errors in compensating for selection and errors in the estimate of the CHLCA or errors in entry times into Eurasia. Important to this page and core to the MtEve TMRCA is the period between the event, and the expansion (big-bang). This represents the core concept of the page, ergo information that reflects uncertainty in the population size estimate is a core aspect of the page. Things that within this time window are also important. Less important or tangential to the page are issues dealing with M, N expansion or problems clocking migrations. Therefore Endicott and Ho should take a lessor importance since this paper is primarily designed with canopy topology issues. However E&H reflect on a core and currently unresolved problem, and this critique is important to the page. Again, I did not want to mention E&H or E,H,M&S review in the lede, I think a general time constraint is important, but MW feels that it is original research to display the full level of current uncertainty, in which case we abide by the strict rules, and unfortunately it is a representative. Becareful what you wish for MW because you may get it.PB666 yap 18:20, 7 December 2009 (UTC)

Issues brought up in recent reviews
I am going to go through these because they are considered 'premier' and some may not have access to reviews, since I have been accussed of creating walls of words, and because of fair-use limitations I am cutting these down to bare essentials.

CHLCA
From Endicott, Ho, Metspalu and Stringer, 2009 Chimpanzee human last common ancestor is the last common ancestor between humans and chimps. "The majority of published rate estimates for human mtDNA [....]have been claibrated with references to the human-chimpanzee divergence. In the simplest case, the genetic difference between humans and chipanzees is divided by the time since their divergence, to obtain and estimate of the rate per year. [See formula within Main]"

Trying to stay within the fair use limitations.PB666 yap

"..there is uncertainty over the exact timing of the human-chimpanzee divergence. A recent review of the molecular estimates suggest a possible age of 4-8 million years [Bradley 2008], whereas the fossil record suggests the split at least 6 million years ago [Benton and Donahue, 2007]"

Note: White has revised this to 7 or more million years ago. Also see revised main page 6 studies rely on CHLCA, 2 on archaeological origins.PB666 yap Corroborating this critique from Soares et al. 2009

On the basis of the age of Sahelanthropus tchadensis, Benton and Donaghue have recommended that 7 mya be taken as the lower bound.(The discoverers of Orrorin tugenensis in fact claim an age for the split of considerably earlier than 7 mya, raising questions regarding estimate of 10.5 to 13.5 mya obtained when non-primate calibration points were used in molecular data but the wieght of moelcualr evidence argue against this.

These early dates are also supported by Pierolapithecus Conclusions three papers, Several papers have drawn into question the CHLCA time used to calibrate most major studies.

From Endicott, Ho, Metspalu and Stringer, 2009 "Thus the first assumption [Those for CHLCA] can lead to the estimates of molecular rates that are inaccurate and artificially precise."

So this is from a review and they are cricizing one of the core problems with confidence intervals, artificially precise. I 100% agree with this statement and most of the literature since 2008 to present reflect this critique.

Muntawandi, what is your response to this?

Inconsistent molecular clock and selection
From Endicott, Ho, Metspalu and Stringer, 2009 "The second assumption is that the molecular evolutionary process along the human and chimpanzee lineages, as well as among humans, has been relatively homogeneous and neutral. In practice this assumption is reflection in the usage of methods based on a strict molecular clock [6,7]. There are two lines of evidence that cast doubt on the legitimacy of these assumptions"

"What is the basis of the edits and the statements. This is derived from this critique. "In summary Pdeitiker has engaged in a pattern of original research in several articles. The original research is masked behind jargon, technical detail and verbosity. Overall it involves putting his own idiosyncratic spin on scientific information that is at odds with mainstream interpretations." Muntawand from WP:OR

Here is the basis of the spin (which I prefer to call neutrality and neutral presnetation) from the reviews:PB666 yap "''First, human and mtDNA appears to be the subject to widespread purifying selection [7, 10 - 14]. Second, studies have shown that older calibrations tend to produce slower and older date estimates [15-18]. IN particular the disparity between the pedigree and phylogeny-based rate estimates is conspicuous [19-21]"

From Soares et al. 2009 "The mtDNA phylogeny seems to show a higher proportion of synonomous mutations in the anceitn thatn in the young branches - i.e. young branches present a higher proportion of non-synonomous mutations [Mutations that are strongest evidence of selection] in the protein coding genes and substitutions in the RNA genes."

From Mishmar et al. 2003 "Our data suggest that regional variation in mtDNA sequence is likely to have been shaped by natural selection. MtDNA variation deviates from neutrality in European, Central Asian, and Siberian plus Native American mtDNA lineages but not African lineages."

From Gonder et al. 2007 Actually it deviates in African lineages, it deviates more in Asian lineages than African lineages as Gonder et al shows. PB666 yap "We detected significant departures from expectations (table 1), as measure be Tajima's D, int the global data set and the pooled African and non-African data setes, but not in Tanzanians. D* and F* statistics of Fu and Li revealed significant departure from neutrality in all populations."

From Endicott and Ho. 2008. "In contrast the 4 fold difference of PC1 + 2 [rates based on the first and second coding positions] can be at least partly attributed to the impact of negative selection" I concur with the severity of difference it is quite large. PB666 yap

From Subranmanian 2009 "The average [omega, dn/ds, nonsynonomous divergence relative to synonomous divergence] observed for the young haplogroups <50 ky is 5.8 times higher than that estimated for human-chimpanzee pairs (0.23 vs 0.05). Within the human populations, the average omega of the young haplogroups is 28% higher than that of the groups aged between 50 and 100 ky (0.18) and 64% higher than that of the >100 ky-old African haplogroups. The omega for the human-Neandertal comparison is only slightly higher than that of the African haplogroups"

I can provide more like this if desired. PB666 yap

So that here we have the five most recent papers on the topic and a review, Gonder and Mishmar present the least severe outcome and have the Highest TMRCAs, Soares and Endicott (Soares using the highest calibration) present Lower TMRCAs. The question is not whether selection has morphed the mtDNA TMRCA, the question now is how badly it has distorted the TMRCA. The data is overwhelming here, there is a clear lack of confidence in the rate estimates, that is why the rates change from Ingman's that is why the TMRCAs have tended to fall even though there have been large increases in the CHLCA.PB666 yap

OK, so next critique. Let me start how do we compress this information down and make it understandable.

So who is fabricating data or coming in with idiosyncratic theories, Muntawandi, you simply don't like what you are hearing and be honest about it.PB666 yap 05:37, 8 December 2009 (UTC)

Requesting we consider this deletion twice
I can see different arguments for and against this deletion. Michael obviously sees the argument for. On the other hand, won't many readers coming to this article want to be able to make just this type of link between key concepts? Isn't one of the "big things" that is said about "Eve" concerning the fact that she was African and recent? (Not just in Time, but also amongst scientists.)--Andrew Lancaster (talk) 21:02, 8 December 2009 (UTC)
 * Upon reflection perhaps a one sentence comment might be appropriate. Something like, "The recent date for mtEve was a blow to the multiregional hypothesis and a boost to the recent out of Africa hypothesis."   But we can leave any further details to those articles. --Michael C. Price talk 05:43, 9 December 2009 (UTC)


 * I originally added it in some form; I cannot say whether this is my version or not but I think it can be reduced into more simple wording.PB666 yap 23:04, 8 December 2009 (UTC)


 * I think it belongs, it can be in both Out of africa page and here, just worded appropriately for a lede.PB666 yap 23:45, 8 December 2009 (UTC)

To include it here (especially in the lead) is wrong. The material should not be duplicated in both articles, and it does not clearly belong here. If multiregionalism was correct there would still be a mtEve. It is Out of Africa that demolishes multiregionalism. If it appears here it should also appear in Y-chromosomal Adam, and in every article about mtDNA, which is silly. The whole point of hyperlinking is to reduce duplication.--Michael C. Price talk 23:57, 8 December 2009 (UTC)


 * With all due respect Micheal mtDNA as a single entity was the weightiest evidence against MREH.PB666 yap 02:49, 9 December 2009 (UTC)
 * In which case a comment in mtDNA would be appropriate. --Michael C. Price talk 05:07, 9 December 2009 (UTC)
 * BTW it looks like we are mandated to change the date one more time. Endicott and Ho give to ranges that are HPD (95%CI) and the second is 213 Ka, in which case we do not violate WP:SYNTH and we only need one reference.

Michael, I see no reason to say that a subject can not be included in two article. I guess what you mean to say is that it should not be handled with the same level of detail in two articles? But this sentence you removed was quite a quick summary. There is I think a much bigger discussion somewhere in the body which can come into question, but removing all mention of this aspect of Mitochondrial Eve would not seem right to me?--Andrew Lancaster (talk) 06:48, 9 December 2009 (UTC)
 * Did you see my response above? I think we agree, a short summary is okay. I've suggested a one sentence form. --Michael C. Price talk 06:51, 9 December 2009 (UTC)
 * No problem with that. The words 'blow' and 'boost' might be substituted PB666 yap 10:59, 9 December 2009 (UTC)

Another candidate for deletion from the lead: "'Eve' would have been roughly contemporary with humans whose fossils have been found in Ethiopia near the Omo River and at Hertho." What is the relevance of this? Either it should be explained or deleted. --Michael C. Price talk 07:08, 9 December 2009 (UTC)


 * Don't you think this is the kind of people looking up the subject would want to know?--Andrew Lancaster (talk) 08:36, 9 December 2009 (UTC)
 * Obviously not, or I wouldn't have deleted it... :-) Seriously, I fail to see the relevance of this. We could also say that Eve was contemporaneous with Homo Erectus in Asia, but we don't. I just don't get what the sentence is trying to say (e.g. are they our ancestors or not?), and I don't think the general reader will get it either. --Michael C. Price talk 08:50, 9 December 2009 (UTC)
 * I see what the statement is getting at, but given the current state of the literature one could ask about the relevance of the statement. I would say delete.PB666 yap 11:02, 9 December 2009 (UTC)


 * OK--Andrew Lancaster (talk) 15:27, 9 December 2009 (UTC)

Simplification
I have done one partial version, that is significantly "dumbed down". It is by no means definitive, but I decided to have a go. If others have the time, you can have a whack at it as well. Did it in hurry as I have to disappear for a few days, so its a bit scrappy, but I'll take a look again when I get back. Wapondaponda (talk) 23:39, 10 December 2009 (UTC)


 * Yes I can tell, moved it here for now Talk:Mitochondrial_Eve/MW_Version.PB666 yap 00:50, 11 December 2009 (UTC)

I see no reason to "dumb down" this or any other article, and I have great difficulty in figuring out just what it is Muntuwandi is trying to achieve (not for the first time, either). --dab (𒁳) 22:56, 11 December 2009 (UTC)
 * "Dumbed down" is not literal. The article simply doesn't need excess complexity to explain the concept of mtEve. PB666 has exaggerated the problems of dating, and has made the article about mtEve largely about dating the problems associated. Wapondaponda (talk) 18:21, 18 December 2009 (UTC)
 * dab, I wrote most of the 'technical' version, it does need to be simplified, I am currently going over what MW wrote to see if there are any good ideas.PB666 yap 15:45, 13 December 2009 (UTC)

In addressing several complaints about the state of the Article on December 1.
 * Female and mitochondrial ancestry
 * Estimating time to MRCA
 * Time estimates based on archaeological methods
 * Time estimates based on chimpanzee-human last common ancestor
 * Estimated times of major mtDNA branchpoints
 * Inter-comparing rates and studies
 * Early modern human population structure
 * Solitary female Eve as a misconception
 * Proposals of a flat population structure
 * Proposals for a population bottleneck
 * Implications of dating and placement of Eve
 * Geographic and Temporal constraints of early modern humans
 * Mitochondrial MRCA and the MRCA of all humans
 * In popular science and culture
 * See also
 * Footnotes
 * References
 * External links


 * There were 46 sections, there are now 18.
 * The article was at its largest about 95 kb, it is now about 60 kb. And there are still some reductions to be had in the Female and mitochondrial ancestry section. Although I think 60 Kb for an article of this importance is appropriate and reduction of material is no longer a concern.
 * The reference section has been converted to citation style with full author representation. Removal of redundant referecnes in the text or reference section.PB666 yap 06:19, 15 December 2009 (UTC)

TMRCA section
My general dissatisfaction with the size and verbosity of this section has been greatly reduced, of course with a good night's sleep and a fresh view tomorrow I will find more to trim back. I think the WP:OR issues have been solved in the tables while essentially displaying the same confidence, just a matter of thinking through how to inter-compare different presentation styles. The 'stray' tag has been placed on a couple of more sections, any recommendations?PB666 yap 17:54, 9 December 2009 (UTC)
 * A lot has been made of the dates. I have an opinion on this. The very first date proposed was from Cann et al. which was 140-280kya, mean 210kya. A google books search reveals
 * 
 * 
 * 
 * This is a quick and unscientific poll, but it shows that the most popular date cited in secondary sources is 200,000 years ago. Others cite 100-200kya, 170kya, 150kya and less than 300kya but the majority use approximately 200kya. Unless someone invents time travel, we may never know the exact date when mtDNA eve lived. Its a futile effort to shoot for accuracy. The TMRCA is a statistical quantity and represents probability, not certainty. Of course several dates have been proposed by peer reviewed studies, most cluster around 200kya(Soares, Behar, Gonder etc). There are some outliers (Endicott and Ho) at 108kya. To summarize, my suggestion is to mention the "mainstream view" in the lead, which is either "approximately 200kya" or "approximately 150-200kya" and ignore confidence intervals for simplicity's sake. A slightly more detailed discussion of dates can take place in the subsection devoted to dates and this can include a brief discussion of CIs. Wapondaponda (talk) 19:53, 9 December 2009 (UTC)


 * Are there secondary sources which are recent? Is there any agreement between them?--Andrew Lancaster (talk) 21:47, 9 December 2009 (UTC)


 * Endicott, Ho, Metspalu and Stringer is a secondary source. See above. They use in their review Endicott and Ho (2007), Mishmar et al. (2003) and Kivilsid. If one goes by the review and stick to WP:SYNTH to the exact letter the range is 108,000 to 198,000 (These are TMRCA not confidence ranges) if one includes the confidence ranges is 82,000 to 217,000 years.PB666 yap 23:35, 9 December 2009 (UTC)


 * Didn't we just get finished with the WP:SYNTH, we have to use only those values provided.
 * The question is whether you want to break from popular science or go back to what got this article flunked last year.
 * Cann el. 1987 140 to 290 Ka.
 * Tamura and Nie 1993 is 80 to 480 Ka, mean = 160 Ka
 * Nie also produced a 96%CI for Vigilant 1991 of 110,000 to 504,000 Ka.PB666 yap 23:23, 9 December 2009 (UTC)


 * Whatever numbers we quote, 228,000 just looks silly given the uncertainties. c 200,000 would be preferable. --Michael C. Price talk 23:26, 9 December 2009 (UTC)


 * We need to use 227,000, I just got the OK from WP:OR that we can use standard deviation ranges. ::In which case you can stick with the value.PB666 yap 23:35, 9 December 2009 (UTC)


 * I see no one arguing against taking confidence intervals into account where they exist in a meaningful form. But Michael's suggestion to approximate more also has no "technical" problem (synth etc) as far as I can see?--Andrew Lancaster (talk) 23:41, 9 December 2009 (UTC)
 * Somewhere there is a policy about false precision, which we should abide by. --Michael C. Price talk 23:51, 9 December 2009 (UTC)


 * See above, its quoted, not policy. Endicott, Ho, Metspala, Stringer, the TRMCAs based on CHLCA have a false precision. I really thought that 70 to 270 range covered all possible confidence scenarios based on 2007-2009 studies.PB666 yap 00:00, 10 December 2009 (UTC)
 * The 70-270kya range may cover all CI scenarios, however it misrepresents the dates most commonly cited in secondary sources. Is there a mainstream date or range of dates?. If there is, that is what we should cite, at least in the lead. A date of 70kya does not have a probability equal to a date of 150kya or 200kya. A date of 70kya creates a lot of problems, for example some suggest that the OOA migration had already taken place by 70kya which makes this date for the MRCA impractical with the OOA hypothesis. According to the tree at http://phylotree.org, at least 35 mutations separate haplogroups L3, M and N from Mitochondrial Eve. Haplogroups M and N, which are the signature haplogroups for the OOA migration, have since accumulated 20-30 mutations. With simple arithmetic it means that the time from mitochondrial Eve to the time of the OOA migration is longer than the time from the OOA till today. If OOA migration took place 50-80kya, then we know that Eve lived significantly more than 100-160kya. This simple arithmetic ignores population size, but if it were considered it would push TMRCA further back in time. But most importantly, I haven't seen any publication that cites 70kya. Wapondaponda (talk) 07:13, 10 December 2009 (UTC)
 * Muntawandi, I don't think you realize the depth of the problem. I will try to explain this in as few words as possible. Variation in the rate of certain substitutions at the canopy of the mtDNA tree has been found to be profoundly different from the rates observed at same sites in longer branches. There are 5 references above that confirm this. I did not create this problem. This observations is the consequence of going from 50 mitogenomes (Ingman: no evidence of selection) to 100s of mitogenomes. If you want a basic paper on sampling effects read Tamura and Nei, 1994 and Wakely, 1994. As a consequence the compression of the canopy of the mtDNA tree is something that is currently an educated guess. Or should I state otherwise many people have observed this 'effect' in the last 4 years, however no-one is greatly sure how to compensate for it. E&H do exact a method, but that method itself is based on to controversial assumptions.PB666 yap 14:44, 10 December 2009 (UTC)
 * Personally, I strongly suspect its not as bad as Endicott and Ho present, I think 42 Ka is way too late for exoafrican migration, and even if it is some the canopy variation issue has much less effect on early african variance. Ergo the Soares estimates for differences between MRCA (192.3) and L3 (71.6) delta = 120.7 and given Endicotts et al. minimum ages for archaeology one can be somewhat comfortable with dates above 150 Ka, however some of the studies done in the 1990s are still relevant today, and the conclusion of a couple of papers would argue if the CHLCA is between 7 and 10 million years, then the human MRCA is between 200,000 and 300,000 years, on paper in particular, Tamura and Nei is roughly immune to the Canopy issue, they argue that the TMRCA can fall between 80 and 480Ka. Nie basically said, if the CHCLA is 9 million years in age, then the MtMRCA can be 760,000 years in age. The top side variation is what I have the most concern about, not bottom side. Since it still appears understanding issues remain I am going to go ahead and add T&N to the table. Remember, NPOV.PB666 yap 14:44, 10 December 2009 (UTC)


 * I think it is correct that just counting the mutations does not work Muntuwandi. How many mutations have died out and left no trace? And more importantly, does the rate of dieing out stay constant over long epochs? No it does not.--Andrew Lancaster (talk) 12:13, 19 December 2009 (UTC)
 * Though methods for dating seem complex, it really comes down to counting mutations. For example, the article
 * states
 * When dealing with large amounts of data, it may not be appropriate to call it "counting" but that is basically what it is. The Henn article has some useful information about mutation rates, and it does have some bearing on the dates for mitochondrial eve, but it is a general controversy that is related to human mitochondrial molecular clock and not specifically to the existence of a mitochondrial eve. Wapondaponda (talk) 08:07, 20 December 2009 (UTC)
 * You are headed off into an area you do not want to go into. Interspecies comparisons results in pairwise comparisons that are simply counted or rate classed. Almost all current studies rate class sites. Fast evolving sites saturate, for example in vigilant et al the CH pairwise difference estimated was close to 100%, that is why they counted 15 fold slower transversions. As some sites would have changed several times and some sites did not change. As two lines grow apart a collection of unstable mutations appear.PB666 yap
 * Slightly deleterious mutations
 * Adaptive mutations
 * Rapidly evolving sites
 * Over a period of 500,000 years the slightly deleterious mutations are removed, the adaptive mutations will persist only if the selective force remains constant and more refined mutations do not appear. The rapidly evolving sites will saturate quickly meaning, in essence at any point as the lines diverge if you stop evolution the site will be in either state (all but randomly).
 * If one ignores the third set of sites, comparisons between species of 500,000 to roughly 20,000,000 years produce nearly linear correlation and minor corrections need to be applied past 10,000,000 years for saturation at slowly evolving sites. Therefore the molecular clock has the most trouble in the shorter time frames, under 100,000 years. The way rapidly evolving sites are handled is with parsimonious analysis, it works well when there are lots of branches, over deeper time depths large corrections need to be made and this increases variance. Reversions and homoplasies will be common in deep branches. The interspecies comparisons generate a stable SNP rate for most stable sites. This rate is useful at the base of a species tree, however as one moves toward the outer branches the rate needs to be increased, the so-called soft, of depth dependent mutation rate corrects for selection at some sites and saturation at others. However the method is not entirely objective.PB666 yap
 * Over a period of 500,000 years the slightly deleterious mutations are removed, the adaptive mutations will persist only if the selective force remains constant and more refined mutations do not appear. The rapidly evolving sites will saturate quickly meaning, in essence at any point as the lines diverge if you stop evolution the site will be in either state (all but randomly).
 * If one ignores the third set of sites, comparisons between species of 500,000 to roughly 20,000,000 years produce nearly linear correlation and minor corrections need to be applied past 10,000,000 years for saturation at slowly evolving sites. Therefore the molecular clock has the most trouble in the shorter time frames, under 100,000 years. The way rapidly evolving sites are handled is with parsimonious analysis, it works well when there are lots of branches, over deeper time depths large corrections need to be made and this increases variance. Reversions and homoplasies will be common in deep branches. The interspecies comparisons generate a stable SNP rate for most stable sites. This rate is useful at the base of a species tree, however as one moves toward the outer branches the rate needs to be increased, the so-called soft, of depth dependent mutation rate corrects for selection at some sites and saturation at others. However the method is not entirely objective.PB666 yap


 * Muntuwandi, there are two things, not one: counting mutations, and then also making assumptions about the rates at which lineages (carrying those mutations) have died out or come into being over time. If we were just to count mutations, and not to consider rates dieing out and coming into being for lineages, then I suppose an approach might be to treat every lineage as having undergone a random walk from Eve, as follows:
 * look at every individual tested so far, and count mutations from proposed ancestral haplotype
 * you will have a very wide range of mutation counts I think
 * calculate implied age for all of these, and you will have a very wide range of age estimations I think
 * If you try it let me know what happens BTW. I think you will get an extremely wide range of estimates, but it will also be very strongly effected by what samples you have?--Andrew Lancaster (talk) 13:56, 28 December 2009 (UTC)

(Outdent)I believe the assumption of a constant molecular clock means that even lineages that have died out, would have statistically the same number of mutations with extant lineages, if they hadn't died out. IOW extinct lineages may only increase the variation in the number of mutations, but under the assumption of a constant molecular clock, they will not change the average number of mutations from the MRCA. If the sample size for existing lineages is large enough and diverse enough, then extinct lineages can be ignored, which might be the case for homo sapiens. As for lineages that come into being, every new mutation is potentially a new lineage. I have done some counting from http://phylotree.org. Though there is variation, for a few lineages I have counted so far, range from 39-65 mutations from Mitochondrial Eve. So we are still looking at an average of about 50 mutations on any given lineage from Mitochondrial Eve. The rho statistic is used to determine the average number of mutations from a sample of lineages. This rho is then multiplied by the substitution rate to determine the TMRCA. Wapondaponda (talk) 17:16, 28 December 2009 (UTC)
 * Yes, that is kind of what I was explaining. If you just take every individual line and ignore all others you can get a very wide range of estimations, right? This could at least give a min-max range. But (a) it might not be as big as it should be, if we consider that many consider that databases will tend to be a bit biased against the deepest rooting branches, not only overall but also within every clade, and (b) this still gives us not uncontroversial way of estimating a most likely age WITHIN that min-max range.--Andrew Lancaster (talk) 19:36, 28 December 2009 (UTC)