Talk:Y-chromosomal Aaron/Archive 3

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Cohanim J2 - Eleazar - Phinchas - Zadok

Agree. Dr. Karl Skorecki have the final word about Cohanim genetic signatures. He, more than anyone else, is the most credible person that can publish articles and true informations about Cohanim genetic signatures, not Jheald. He was the one who discovered the CMH in 1997. After 10 years, he came out to announce in 2007, that "he and his research team have discovered not one but two Cohen Modal Haplotypes, which he called J1 and J2". Katz, Kaplan, Rappaport, Shapiro, Kovacs,Konh, are all J2. Those surnames and families are the most prestigious Cohanim families today. They are all Cohanim J2. J2 Cohanim is much more common than J1 Cohanim. "Skorecki reported that Pinchas the zealot mentioned in the Bible may be the origin of J2" This is a very important information that confirms matematical, historic, and genealogy facts. Phinchas was the son of Eleazar. Itamar was his brother. Out of the 24 Cohanim houses, 16 did belong to Eleazar-Pinchas, while only 8 came out from Itamar seed. Eleazar-Pinchas (J2) had a much larger population than Itamar. All Cohanim genealogy were divided into those two houses, until the destruction of the second temple. So, if Phinchas is the origin of all Cohanim J2, than it makes sense, since his Cohanim seed produced much more descendants than his brother Itamar. So, it is very normal and logical to understand why genetic cohanim signatures presents two CMHs wich is the caracteristics of the two houses of Cohanim. All genealogic tree for Cohanim since Aaron until today are still divided into these two houses: Eleazar-Pinchas and Itamar. During the 1270 years of the cohanim services (Mishkan and Mikdash) the Cohanim jews were always selected by the 24 houses. 18 from Pinchas, and only 8 from Itamar. That´s why J2 Cohanim is almost twice as common than J1 Cohanim" —Preceding unsigned comment added by 189.27.127.148 (talk) 15:56, 3 February 2009 (UTC)

WHAAAAT

{During the 1270 years of the cohanim services (Mishkan and Mikdash) the Cohanim jews were always selected by the 24 houses. 18 from Pinchas, and only 8 from Itamar. That´s why J2 Cohanim is almost twice as common than J1 Cohanim"}

lol :+D

Please STOP BEING STUPID and be Real

J1 if for real Jews ArabIsraelites decent ...and J2 is for GOYIM jews(false jews)

78.101.90.200 (talk) 17:34, 7 February 2009 (UTC)

Agree. Dr. Karl Skorecki of the Ramban/Technion Medical Center in Haifa, who is renowned for his discovery in 1997 of the Kohen genetic signature, reported in 2007 that he and his research team have discovered not one but two Cohen Modal Haplotypes, which he called J1 and J2. "Pinchas, the zealot mentioned in the Bible, may be the origin of J2" he suggested. The extended CMH-12 is established for J1 Cohanim only. If somebody in J1 also has CMH-12, then he is likely to have Cohen ancestry. J1 Cohanim form a tight network around CMH-12. But this is not the case for J2. The CMH-12 does not apply to J2. If a J2 is in CMH-12, it tells us nothing about one being a Cohen. In J2, the CMH-12 has no special status. The J1 lineage of Cohanim in CMH-12 could be much more recent. Jewish Cohanim in J2 form a much larger network around CMH-6 than Jewish Cohanim in J1 do. Cohanim Jews belong mainly to haplogroup J2 because Cohanim as a group (J1,J2) have a high frequency of J2. The J2 Cohanim have much more ancient common ancestor, according to the *observed* mutation rate (which is three times faster than the standard, "effective" rate). J2 Cohanim serves as the deep background of the priestly Jewish caste (J1,J2). “The finding of J2 among contemporary Cohanim can lend credence that today’s Cohanim (J1,J2) may be descendents of two different founding dynasties – including coalescence to Phinchas as the founder of J2 Cohanim” concluded Dr. Karl Skorecki.

Why the confirmation that Phinchas may be the orign of J2 Cohanim is so important?

“Pinchas, son of Elazar, son of Aharon the Kohen turned back My anger from upon the Children of Yisrael, so I did not destroy the Children of Yisrael. Therefore say - behold, I give to him My covenant of peace” (Numbers 25:10 - 30:1). “God gave to Pinchas a new permanent priesthood, for him and his descendants for all generations.” (Zohar HaKadosh 3:124a). God indicated that Phinchas had acted justly by rewarding him with the most counter-intuitive form of blessing - an all-new form of priesthood. He is given the Divine Covenant of Peace and the gift of eternal priesthood. Zadok, a descendant of Phinchas, was the Kohen Gadol who anointed David King, and every Kohen Gadol after Zadok was in the direct line of Phinchas. In Zadok the line of Phinchas resumed the dignity, and held it till the fall of Jerusalem. THIRD TEMPLE PROPHECY. “But as for the priests, the sons of Zadok, they are to come near me to do my work, they will take their places before me, offering to me the fat and the blood, says the Lord” (Ezekiel 44:15).

THE Y CHROMOSOME OF THE JEWISH HIGH PRIESTS ZADOKITES

THE Y CHROMOSOME OF THE JEWISH HIGH PRIESTS ZADOKITES

By Anatole Klyosov* from "DNA Genealogy, Mutation Rates, and Some Historical Evidences Written in Y-Chromosome. 2008" (Nature,2009. Precedings)

J2 Cohanim

The Cohanim J2 haplotype is noticeably younger (in terms of time spans to their common ancestors) compared to Jews with J1 CMH. The ancient common ancestor for the J2 Cohanim lived 3,100±200 years before present (BP), exactly around the date of the birth of Pinchas, the grandson of Aaron Hakohen. The youngest common ancestor of J2 Cohanim (Sephardim and Ashkenazim) lived about 900 years before present (BP), apparently, during the First Crusade. On the other hand, Jews with J1-CMH shares a common ancestor who lived 4,000±200 years BP, around one thousand years before Pinchas. J2 Cohanim and Jews with J1 CMH are descendants of two different ancestor of which J2 lineage has the accurately coalescence to Aaron Hakohen – Pinchas, and J1 Jews with CMH an ancestor who lived 4,000±200 years BP, around one thousand years before Aaron Hakohen - Pinchas. J1 Jews with CMH do not have the same ancestor in a time frame of Aaron Hakohen (3,100±200 years BP).

J2 Haplotypes

The Jewish haplotypes of haplogroups J1 and J2 do not mix with non-Jewish haplotypes from the same haplogroups on the cluster tree. In other words, the Jewish haplotypes have their own exclusive branches. As a result, they have developed quite distinct lineages in haplogroups J1 and J2. Jewish J2 haplotypes do not mix with non-Jewish J2 haplotypes and J2 Jewish haplotypes do not mix with the J2 Cohanim Jewish haplotype on a haplotype tree. Jews have two modal, or ancestral haplotypes in haplogroup J2. One directly related to regular Jews, and other directly related to Cohanim Jews. This means that J2 Cohanim Jews form their own exclusive cluster in J2 Haplogroup and regular Jews are not part of it.

J2 Cohanim 6- and 12-marker modal (ancestral) haplotype

The J2 Cohanim 6-marker modal is 15-15-23-10-11-12. The extended 12-marker modal for J2 Cohanim is 12-23-15-10-X-Y-11-15-12-13-11-29 with X=14 and Y=17 for J2 Askenazi Cohanim and X=13 and Y=18 for J2 Sephardi Cohanim. With the extended 12 markers for J2 Cohanim, it was clearly revealed that both J2 Askenazi Cohanim and J2 Sephardi Cohanim do share 10 out of 12 markers. Such tight connection clearly indicates that both J2 Cohanim Ashkenazi and J2 Cohanim Sephardi had a common ancestor not a long time before their separation. The only real differences found among them are located on two specific markers: DYS 385a, and DYS 385b. The variation found has just one-step mutation in each of them. Thus, both of the J2 Cohanim lineages have a recent ancestor that lived around 900 year ago. The second Jewish haplotype found in J2 haplogroup is attributed to regular Jews, and is derived from a much older lineage, with a common ancestor who lived over 10 thousand years ago. This lineage continued 6,000 years BP in people, who will became Jews two thousand years later, which is exactly Abraham times, and 1200 years ago a modal haplotype 14 16 23 10 11 12 had appeared in haplogroup J2, which accidentally, as a result of game of statistics, looked exactly as the J1 "Cohen Modal Haplotype". In haplogroup J2 it has nothing to do with either J2 Cohanim or J1 with CHM, and cannot have a common ancestor that could live in a time frame of Aaron Hakohen (3,100±200 years of BP).

The J2 Cohanim Exclusive Cluster

Non-Jewish J2-haplotypes take an opposite side of the J2 haplotype cluster tree from the Jewish J2 haplotype. Furthermore, almost all Jewish haplotypes (J1 and J2 separately) have taken the opposite sides when compared with non-Jewish haplotypes on the tree. It also can be seen that non-Jewish J2 haplotypes are older (more distant from the tree base) than Jewish J2 haplotypes. Indeed the oldest Jewish J2 haplotypes are related to regular Jews and their ancestors from about 6,000±1,000 years ago.

J2 Sephardi Haplotype

There are 43 different J2 haplotypes among 174 samples collected among J2 Sephardim of Iberian peninsula (Adams et al., 2008), and these 43 J2 haplotypes split into 8 genealogy lineages, or branches. This can be compared with just one separate and exclusive branch that was formed by the J2 Sephardim Cohanim. Each of the 8 branches is a genealogical lineage of J2 Sephardim. A common ancestor of 8 branches formed by the J2 regular Sephardim lived about 11,500 years ago. On the other hand, the ancient common ancestor for the single exclusive branch formed by all J2 Cohanim lived 3,100±200 years before present (BP), exactly around the date of the birth of Pinchas, the grandson of Aaron Hakohen, and the lineage of J2 Cohanim Sephardim we can detect today (with a limited number of samples tested) splits from the ancient lineage about 900 years ago. The DNA analysis has shown that the J2 Cohanim Sephardim haplotype is not in the mainstream of J2 Sephardi haplotypes. J2 Cohanim Sephardi Haplotype (X=13 Y=18) has a very unique and directly related haplotype with J2 Cohanim Askenazi Haplotype (X=14 Y=17). Hence, it was formed one exclusive single cluster. This is because both Cohanim Jews, Askenazi and Sephardi, for thousands of years preserved their genealogical lineages since the Temple period. Of course they did not think in terms of the DNA, however, they have faithfully followed Cohanim Jewish tradition until present day. As a result this is exactly what the haplotype tree shows.

J2 Sephardi Cohanim

J2 individuals with the following 10 markers 12-23-15-10-X-Y-11-15-12-13-11-29 belong to J2 Cohanim modal haplotype, and they are part of the J2 Jewish haplotype tree as a whole. However, within this J2 Cohanim signature, there are two distinct clusters, that is Ashkenazi Cohanim (ancestral X-Y = 14-17) and Sephardim Cohanim (ancestral X-Y = 13-18). These clusters form the exclusive J2 Cohanim Jewish Haplotypes. Cohanim haplotypes within the clusters may have some (few) mutations, because mutations in 12-marker haplotypes occur randomly every millennium, and some people have two mutations in their 12-marker haplotype after a thousand of years, some people have one mutation, and some people have no mutations after a thousand years. Hence, there are little variations between haplotypes within a cluster, but not much. So, J2 Cohanim split into two lineages, with 14-17 and 13-18 in their markers 385a,b, which nicely places them into Ashkenazi and Sephardic communities. At the same time, J2 Cohanim haplotypes, being within their Ashkenazi and Sephardi clusters, noticeably differ from J2 haplotypes of "regular" Jews, who do not belong to the Cohanim lineages. To have 25- or 37-marker haplotypes of J2 Cohanim Ashkenazi or Sephardi, it will be priceless for their study in relation to Pinchas-Zadok.

Statistics

According to the tradition, and also taking into account statistics and cemeteries tombstones, the Cohanim represent approximately 5% of the total of Jews. Out of 174 samples of Sephardim of various haplogroups in Iberia, only 9 - on average - might have been the Cohanim. However, merely 81 samples were determined as of haplogroups J1 and J2. On average, there might have been four Cohanim among them, two J1 and two J2. In fact, among all 43 of J2 Sephardi individuals not one belonged to the J2 Cohanim haplotype cluster. By adding 3 new samples of J2 Sephardi Cohanim found in another database, not listed on the first 43 J2 Sephardi "regular" Jews (Adams et all, 2008), the frequency of J2 Sephardi Cohanim among J2 Sephardi Jews reaches a normal rate around 7%. Obviously, a number of tested Sephardim in that study was not statistically very significant, however, it gave reasonable estimates. The two mutations found on J2 Sephardic Cohanim is just right for their common ancestor to live 800-1000 years ago, that is time of the First Crusade. Haplotypes show, that all J2 Cohanim, both Ashkenazi and Sephardi, are directly related to each other, by forming a single exclusive cluster. All J2 Cohanim in the full cluster, both Ashkenazi and Sephardi, share the same ancestor who lived 2175 years ago.

The Haplotype of J2 Cohanim Zadokites

The haplotype 12-23-15-10-X-Y-11-15-12-13-11-29 which was incomplete in (Behar et al, 2004), is in fact the 10-marker modal J2 Cohanim haplotype for both Cohanim Askenazi and Cohanim Sephardi. When J2 Cohanim Askenazi present this 10-marker haplotype along with X=14 and Y=17, and J2 Cohanim Sephardi who also have this 10 markers along with X=13 and Y=18, it became the complete 12 marker for Cohanim Askenazi and Cohanim Sephardi ancestral haplotypes, respectively. It turns out that both J2 Cohanim, Askenazi and Sephardi, have the same ancestor in a time frame around only 900 years ago. It indicates that this haplotype can best represent the J2 Cohanim and therefore it forms a unique special haplotype that unites both J2 Askenazi and J2 Sephardi in one single and very exclusive cluster that is directly related to the main line of Pinchas-Zadok. Therefore, it can be called the haplotype of Cohanim Zadokites.

Families with the extended 12 match Zadok Modal Haplotype

By selecting and analyzing the DNA of the J2 samples it was found that among thousands of surnames only a certain circle of families have a 12 match with the 'Zadok Modal Haplotype'. The families that have the full 12 match with the 'Zadok Modal Haplotype' are directly associated to traditional Cohanim surnames recognized in both Askenazi and Sephardi Jewish communities.

The modal extended 12 'Zadok Modal Haplotype' 12-23-15-10-14-17-11-15-12-13-11-29 was found in Askenazi Cohanim families with the surnames Cohen, Katz, Kaplan, Rapaport, Kahan.

The modal extended 12 ‘Zadok Modal Haplotype' 12-23-15-10-13-18-11-15-12-13-11-29 was found in Sephardim Cohanim families with the surnames Cohen-Rodrigues, Cohen-Pereira, Cohen-Machado.

_____

Anatole Klyosov: Chief Professional Scientist, Patents and publications (2003-2008). Professor of Biochemistry at Harvard University (Cambridge, MA), Harvard Medical School (Boston, MA),Institute of Biochemistry, Russian Academy of Sciences (Moscow), Consulting Chief Scientist and Chief Scientist for Pro-Pharmaceuticals (2000-2008); Professor and Head, Carbohydrate Research Laboratory, World Academy of Science (Member, since 1989),USSR National Prize in Science , Gold Medal, the USSR National Achievement.

THE REPLACEMENT OF THE ARTICLE: Y-CHROMOSSOMAL AARON

Y-chromosomal Aaron is the name given to the hypothesised most recent common ancestor of many of the patrilineal Jewish priestly caste known as Kohanim (singular "Kohen", "Cohen", or Kohane). In the Hebrew Bible this ancestor is identified as Aaron, the brother of Moses.

Background

The Middle East through the eyes of the ancient Israelites

Although membership in the Jewish community has, since at least the second century CE, been passed maternally (see: Who is a Jew?), membership in the group that originally comprised the Jewish priesthood ("Cohen" or "Kohen"; plural: Kohanim), is patrilineal. Modern Kohens claim descent from a biblical person, Aaron, brother of Moses, in the direct lineage from Levi, the patriarch of the Tribe of Levi, greatgrandson of Abraham, according to the tradition codified in the Tanakh (שמות / Sh'mot/Exodus 6). DNA testing is aiding scholars to trace the lineages found among modern Jewish populations, including contemporary Cohen families, to decipher origins of the people groups that were joined to the ancient Israelites and to identify genetic admixture and genetic drift.

For human beings, the normal number of chromosomes is 46, of which 23 are inherited from each parent. Two chromosomes, the X chromosome and Y chromosome, determine sex. Women have two X chromosomes, one inherited from their mother, and one inherited from their father. Men have an X chromosome inherited from their mother, and the Y chromosome inherited from their father.

Males who share a common patrilineal ancestor should also share a Y chromosome, diverging only with respect to accumulated mutations. Since Y-chromosomes are passed from father to son, all Kohanim men should theoretically have almost identical Y chromosomes; this can be tested with a genealogical DNA test. As the rate that mutations accumulate on the Y chromosome is relatively constant, scientists can estimate the elapsed time since two men had a common ancestor. (See molecular clock.) —Preceding unsigned comment added by 201.22.163.182 (talk) 00:40, 28 February 2009 (UTC)

The 'Cohen Modal Haplotype'

We know 6-marker haplotypes, called the "CMH", or the "Cohen Modal Haplotype". In the 6-marker format they fit J1 Jews with surname Cohen (and with many other different surnames) and do not fit J2 Cohanim. This confused Karl Skorecki and his co-authors (Nature,)twelve years ago. They thought that the 6-marker CMH is the signature of Cohanim. Granted, it was easy to be confused that time because they operated (as well as Behar in 2003 ) only with "J" haplogroup and with 6-marker haplotypes. When science moved to 12-marker (and longer) haplotypes and to J1 and J2 haplogroups, it became obvious, that 6-marker haplotypes were giving false impression. It turned out that there are two J1 "CMH" Jewish 12-marker lineages (both started with the respective 12-marker ancestral haplotypes) which are different.

Hg		DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii
J1	CMH-6	12	23	14	10				16			11
J1	CMH-12	12	23	14	10	13	15	11	16	12	13	11	30
J1	CMH-12	12	23	14	10	13	17	11	16	11	13	11	31

Both of them have the same shorter 6-marker J1-CMH, but differ by four mutations in the 12-marker format. Furthermore, it turned out that the both lineages have their common ancestors (obviously, different too) who lived in different times: one just 1350 years ago, in the Diaspora, and another at the beginning of AD, approximately 2000 years ago. Both lineages coalescence around 4000+/-200 years ago. However, even more striking was that J1 Jews that presents the CMH, in fact belong to yet another lineage which in the 12-marker format has the following ancestral haplotype:

Hg	ANCESTRAL HAPLOTYPE	DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii
J1	CMH-12	12	23	14	10	13	15	11	16	11	13	11	30

Again, it fits the shorter, 6-marker J1-CMH, but in the 12-marker format it does not fit either of the J1-CMH. A common ancestor of all J1-CMH Jews with known haplotypes lived about 1400 years ago, in the Diaspora. This is essentially all that we know about J1 Jews with CMH and their connection to the 12-marker "Cohen Modal Haplotypes". The latter haplotypes were determined in 25-, 37-, and 67-marker format, their ancestral haplotypes were identified, and the above conclusion were confirmed.

Hg	MODAL	DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii
J1	CMH-12	12	23	14	10	13	15	11	16	12	13	11	30
J1	CMH-12	12	23	14	10	13	17	11	16	11	13	11	31
J1	CMH-12	12	23	14	10	13	15	11	16	11	13	11	30

These three J1-CMH lineages indeed remain separated, and belong to three different common ancestors who lived 2000, 1350 and about 1400 years ago, and all were split from an ancient lineage coming from 4000+/-200 years ago.

Comments. Anatole, first of all, please note that Wikipedia is a collaborative process; please also note that material here should be sourced and verifiable (see WP policy on reliable sources); and that WP is not a venue for publishing original research, or repeating self-published material (see WP policy on no original research). So please don't be offended if people want to discuss your edits here, ask why you want to remove so much that is already in the article, and clarify some of what you want to add, rather than just accepting a wholesale re-write of the article.

Now, regarding the above. The importance of the 12 23 14 10 13 15 11 16 12 13 11 30 haplotype is that it is the one very widely presented as the extended J1 Cohen haplotype by FTDNA and others. Discussion must start from that point. Now, you want to add another haplotype to the discussion, 12 23 14 10 13 17 11 16 11 13 11 31. This is not the 12-marker haplotype presented by the more obvious authoritative sources. Furthermore, it does not appear in the main FTDNA Cohen project [1], nor the main FTDNA Haplogroup J project [2], nor Ysearch [3] -- whereas there are very many Cohen matches for the other haplotype. To be discussed in the article at all, this new alternative haplotype must have a credible source.

Secondly, the TMRCA dates. These cannot be considered reliably sourced unless you can point to where they have been published by somebody. Furthermore, to assess their credibility it is essential to know the dataset they were based on. As you well know, these estimates are very sensitive to the choice of haplotypes included or not included in the estimation, particularly of potential outliers that may or may not be included, and how the data is grouped. If we are going to include TMRCA estimates in such a contentious area, it is essential that readers can assess and check them.

Additionally, as discussed at length on this page in these discussions which you have since deleted (something which is considered very bad form on WP talk pages), there are simply no grounds for asserting an accuracy of +/- 200 years in these calculations. The most accepted estimates for the uncertainties in TMRCAs are those of Bruce Walsh. Plugging in the 8/12 match between your two assumed recent consensus haplotypes into a calulator [4] based on Walsh's model gives an estimated 95% confidence range of 82 to 525 generations - i.e. about anything from 2000 to 13000 years. Assessing the full range of possibilities accurately is essential before making pronouncements on what may or may not be possible.

These are the kind of issues we need to address before adding discussion of J1 haplotypes different from the commonly accepted widely publicised 12-marker one. Jheald (talk) 11:38, 2 March 2009 (UTC)

Comments. The whoever-is-it confuses the so-called "CMH haplotype" and actual Cohanim haplotypes. "CMH" haplotypes have great many people including Arabs, Basques, Frenchmen, and the list continues. There is the "Arabian peninsula" project, including many people with the "CMH". After studying the "CMH" quite extensively, I came to a conclusion that is it quite spread in the Arabic world, was "picked" by the Jewish Patriarchs, and the "CMH" was carrying in the Jewish community for some four thousand years. It split into several lineages, at least three of them carry the "CMH". To use the "CMH" as a measuring stick for Cohanim would be a mistake. In other words, the "CMH" is a mainstream of Jewish J1 haplotypes. Of course, in many Jews inherited this haplotype it partially mutated, and it is very clear on a haplotype three. Apparently, the "CMH" haplotype is a signature of "12 Tribes of Israel" in J1 haplogroup. However, the "CMH" passed through a bottleneck in the 7th century AD, and after this time it quickly proliferated. Hence, this can be attributed to a powerful and influential person, who lived in the 7th century. This is why I came up with the Bustenai concept. It is described in my paper "Origin of the Jews". That it is a plausible concept. It optimises data available. If someone would come up with an alternative hypothesis, I would appreciate it. However, if people try just to dismiss it, without offering anything constructive, it is just fruitless. Apparenly, Jheald does not know that Walsh considered pair of haplotypes. I have considered series of haplotypes, in 12-, 25-, 37- and 67-markers. Walsh has never considered those haplotypes. Therefore, to quote Walsh in that regard is to show lack of knowledge. What are those cliches about? Let us consider.1. That the CMH haplotype are "Cohens". Wrong. Only a little fraction of the CHM belong to Cohanim. Other "CMH" are regular Jews, but - I believe - are descendants of King David and Bustenai.But not Cohanim. 2 - Jheald don´t know that the J-1"CMH" in 12-marker haplotypes split into three lineages, with common ancestors about 2000 years ago, 1350 years ago. 3 - Jheald dont know who in J2 haplogroups are Sephardi Cohanim and who not. 4 - Jheald don´t know that J1-CMH appeared about 9,000 years ago. 5 - Jheald still apparently do not know that Walsh considered only pairs of 12-marker haplotypes, and I work with dozens and sometimes hundreds of 25-, 37- and 67-marker haplotypes. And they have guts to question my findings and my error margins. Let them calculate data themselves, and I will consider their "calculations". 6 - Jheald, alleles are not carved in stone or in "L24" (?). Mutation can happen in any generation. Those who have DYS 450=9, can in the next generation have 8, and vise versa. That is why we should not pay too much attention on the alleles, which mutate, but we should identify the alleles of the common ancestor. Jheald still apparently do not understand it as well.Jheald, one or two mutations in 67-marker haplotypes does not say anything because it corresponds to about only 3 generations. --189.63.107.69 (talk) 02:45, 3 March 2009 (UTC)

Anatole, to take your points in order

(1) I am quite clear on the difference between CMH and "Cohens", thank you. The existing article itself discusses how having the CMH for a typical Jew only raises their probability of Cohen ancestry to a little over 20%, i.e. still leaving an almost 80% chance that they are not a Cohen. And the article makes clear that the CMH is widespread elsewhere in peoples where there is no suggestion of Cohen ancestry.

(2) No, I don't know this. As far as I know, the claim originates with you. There is no such division made, for example, on the FTDNA Haplogroup J project site. At the very least you need to publish your dataset, and show that other people accept your analysis.

(3) Behar drew his sample from religious Cohanim. So we should take his data seriously.

(4) This date seems quite possible, though it would be good to find something WP would consider a reliable source for it. So what?

(5) Two points here. Walsh gave calculations for 12, 25 and 37 marker haplotypes, and a method which could be extended to others. I used a 12 marker calculation because those were the data you presented. Yes, longer haplotypes would give different numbers. But there would still be quite a high uncertainty.

Secondly, yes, you can use lots of haplotypes. But as you go back in time, these rapidly converge together. Most of the uncertainty comes in the time for the last stages of the convergence. Eventually, you say further than 2000 years back, you are dealing with only two lines. Therefore you would expect uncertainties from this part of the history similar to those in Walsh's calculations.

It is perhaps worth revisiting two quotes cited in the previous conversation on this page, which you deleted.

• " We really don't know how many generations take place between those early occurrences of nodes in the tree for the sample population. There is intrinsic noise uncertainty in any ability to infer them no matter how large the sample population (which is constrained by total population). Those intervals which contribute to the total generations are shared in common by a very large fraction of the sample population haplotype pairs , so there is no way to drive down the statistical uncertainties in their contribution to total G. It requires INDEPENDENT measurements of something to invoke the 1 / square root of N diminishment of noise errors. " (Ken Nordtvedt) [5]
• "Last I heard, Anatole had not developed an error analysis for his estimation technique. The most he ever showed on this list was a comparison of two or more non-independent estimates from the same or similar data sets and an assertion that the spread of the estimates was the uncertainty in the technique." (John Chandler) [6]

(6) L24 is an SNP, not an STR, and defines the subgroup currently designated J2a1k.[7] From tests done, we can say it includes the Ashkenazi J2 14-17 cluster, and others. As you say, other haplotypes might belong to this cluster, but they might not. The SNP test is available from FTDNA for about $25. A Bayes classifier has been developed which can predict probabilities for membership of J2a1k for an STR haplotype, which appears to be reasonably reliable; but the test itself is the best way to achieve maximum certainty.

People can also test the slow-moving STR markers DYS 450 and DYS 445, which appear to take very characteristic values in, and only in, L24. Taken together these can give almost certainty as to the result of the SNP test.

The group of three twelve-marker samples you identify as so-called "Sephardi Zadokite Cohens" show a 70-30 chance of belonging to L24. But their haplotypes are also close to known haplotypes in J1a*. Jheald (talk) 10:10, 3 March 2009 (UTC)

Disagree. Unfortunately, Jheald is not qualified in DNA-genealogy of the Jews. Let me repeat some of those statements. Frankly, I do not care if someone “questioning” my studies and publications. As I told before here, I recognize (and appreciate) only one form of scientific (and any) debate – if the opponent explains data from different positions, provides a justified alternative. If the opponent just “questions”, he is not an opponent. Just an unqualified talker. I have seen many of them. Published papers on haplotypes of Jews (Behar, Thomas, Skorecki et al) were very useful from two points of view: (a) they contained haplotypes as a “database”, and (b) they have attracted attention to DNA genealogy. Unfortunately, their interpretations were shallow and/or largely incorrect. This is related to two papers in NATURE in 1997-1998, paper by Behar et al (2003, 2004), paper about “Black Jews of Africa” (Lemba). The paper by Levi-Coffman (2005) was very poor and said nothing new. However all those papers were, I repeat, useful, and reflected a level of knowledge those times. We cannot blame the authors. They indeed knew too little (though, the Behar paper-2003 was good in identifying R1a1-Jews). However, to make those old papers the Bible and to repeat and insist those obsolete and largely incorrect conclusions (added in Wikipedia and in the J Haplogroup project in FTDNA) as mantra – it is shame nowadays, for sure. The currently main article is full of these incorrect conclusions, thus need professional editing.

Anatole, we need to review what these papers said, and didn't say, so that people have the references as to where to find them, and know how to assess what they may find written about them elsewhere. I think the coverage we have at the moment is appropriate. It explains what the papers said, and what was so striking about it that it caught people's attention. But it also I hope explains the limitations of them. I don't think our coverage derives more from them than is still true. Jheald (talk) 23:17, 4 March 2009 (UTC)

Having said this – back to Jheald comments for the last time:

“- Behar drew his sample from religious Cohanim. So we should take his data seriously.” - Jheald

No, he did not. In his 2003 paper he described the analysis of 988 haplotypes of Ashkenazi Jews, Sephardi Jews, and several non-Jewish populations. As a main criterion was “self-designation by the participants providing the sample”. And then – “The Jewish samples comprised 236 Ashkenazi Jews and 163 Sephardi Jews, who were further divided into 100 Ashkenazi Israelites, 76 Ashkenazi Cohanim, 60 Ashkenazi Levites, 63 Sephardi Israelites, 69 Sephardi Cohanim, and 31 Sephardi Levites”. And again Behar writes – “identified themselves”. And then – “Information regarding paternal lineage was collected from each of the participants”. As you see, they just called themselves so. No other control was made. Aren’t you surprised, that among the 399 Jews 145 happened to be “Cohanim”?? 36%? More than a third? Do you want me to take it seriously? We cannot consider them “Cohanim”. Behar himself writes in the paper that it supposed to be 4% of Cohanim in the Jewish population. To my surprise, he has never mentioned this discrepancy, that “Cohanim” happened to be 9 times more, than it supposed to be. 900%. Now, among those “Cohanim” were haplogroups BR(xDE,JR), E, K, P(xR1a), R1a1. The rest was J, without separation into J1 and J2. This is how the haplotype tree looks like (my data, 6-marker haplotypes of J haplogroup from the Behar’s paper):

Haplotype Tree – 6 markers - CMH

The “comb” around the trunk – the “CMH” haplotypes. These are all 194 haplotypes of haplogroup J, and 91 of them (47%) are the “CMH”. Can you imagine this? 47% of the CMH from all Jewish J haplotypes. Clearly, they are not “Cohanim” at all. They are not “Cohen” haplotypes that Jheald refer to. He or She apparently does not understand that they cannot possible be “Cohen” haplotypes. They are mainstream Jewish haplotypes of haplogroups J1 and J2 (mainly J1, but of J2 as well). Some of them are obviously VERY old, some (the “CMH”) are obviously very new. I analyzed all of them, they belong to many families. Who else did it” Who else has analyzed them besides me? They all published in my papers “Origin of the Jews”. The same in kind tree containing 6-marker Jewish haplotypes of J1 haplogroup taken from YSearch database:

Haplotype Tree J1 with CMH

As you see, again plenty of the “CMH” sitting around the trunk. To appreciate how the Behar’s set and the YSearch set are similar, let’s consider amounts of mutations per haplotype: there are 263 mutations in 194 haplotypes in the Behar’s set, and 123 mutations in 90 YSearch haplotypes, that is 1.36 and 1.37 mutations per haplotype, respectively (0.226 and 0.228 per marker, respectively). These average numbers only say that the datasets are very similar. However, they contain a number of “common ancestors”, that is a number of different lineages. Well, the rest is described in detail in my paper. The J1-“CMH” in the 6-marker format splits in two lineages, with 1350 and about 2000 years to their common ancestor. Here how this split shows in 37-marker format. ALL these haplotypes are the “CMH” in their 6-marker format:

Haplotype Tree 37 markers J1 with CMH

As you see, there are clearly two different branches here. The “younger” have a DYS385a,b pair 13-15, the “older” 13-17. There are other rather significant differences in 37- and 67-marker formats. This pair was not determined and, hence, considered in the 2004-Behar paper.

Anatole, you're being silly. Skorecki, and later Behar, deliberately recruited a sample of Cohens, as Cohens, by seeking out Cohens - presumably through synagogues, advertisements, etc. The idea that they just randomly recruited from the population, and just happened to find that a third of their sample were Cohens simply isn't the case. Instead, they wanted to test a hypothesis about Cohens, and recruited a sample deliberately of Cohens for that purpose.

Yes they were self-identified, but that is already a lot better than just going by people's surnames. And if they found people through synagogues, then it would match the community's identification too -- and orthodox communities tend to be quite rigorous in needing proof that a Cohen really is a Cohen, eg before calling them up for an aliyah.

It is possible that the selection wasn't completely reliable, but it is clear that it must have been pretty reliable just from how distint the Cohen population was from the Yisrael in the studies. Jheald (talk) 23:32, 4 March 2009 (UTC)

And yes, not all CMH people are Cohens. And not all religious Cohanim are CMH (nor all of them from the same family). But the article already makes this clear. Jheald (talk) 23:37, 4 March 2009 (UTC)

As to your paper, I haven't seen your data so I can't assess your conclusions. I am surprised if you have found such a substantial cluster of Cohens other than those that match CMH-12 haplotype, when there is so little trace of any such in the databases above; but why not publish your data on a website, then we can all look more closely at it, and see whether you are right. Jheald (talk) 23:46, 4 March 2009 (UTC)

"Anatole, to take your points in order: (1) …The existing article itself discusses how having the CMH for a typical Jew only raises their probability of Cohen ancestry to a little over 20%, i.e. still leaving an almost 80% chance that they are not a Cohen. And the article makes clear that the CMH is widespread elsewhere in peoples where there is no suggestion of Cohen ancestry." Jheald

See above. If I understand you correctly, it is not “a little over 20%”, but 1/9, that is about 11%.

The calculation that gives the 20% probability is clearly set out in the article. How do you get your 11%? Jheald (talk) 23:46, 4 March 2009 (UTC)

(2) No, I don't know this. As far as I know, the claim originates with you. There is no such division made, for example, on the FTDNA Haplogroup J project site. At the very least you need to publish your dataset, and show that other people accept your analysis. 'Jheald'

The division of the “CMH”, shown above, is published in my paper “Origin of the Jews”, and in pending paper in “NaturePreceedings”. The two J1-“CMH” haplotypes in the 12-, and 37-marker formats are shown as #8 and 9 (p. 220) in my paper. The split is shown above.

The two “CMH” J1-12-marker haplotypes are

12 23 14 10 13 17 11 16 11 13 11 31

12 23 14 10 13 15 11 16 12 13 11 30

The last allele (DYS389-2) could be 30 in both cases, it is still unclear to me. They are mixed. However, the first two (DYS385b and DYS439) are quite clear. Now, the J1-Cohanim, according to your data, are

12 23 14 10 13 15(16) 11 16 12(11) 13 11 30

Hence, they belong either to the “younger” lineage, or form their separate lineage. In order to say which way, I need to see their longer haplotypes.

It is, as you have noted above, necessary to distinguish Cohanim from others who also happen to have a 6-marker haplotype close to J1. The claim is that a very large number of Cohanim, close to FTDNA's CMH-12 are closely related and share a comparatively recent common ancestor.

The FTDNA CMH-12 cluster is the one that is most apparent in the public databases. If there is another cluster nearby, specifically of Cohens, please present the data for it.

However, even if such a cluster does exist, there would still be the possibility that both ancestors were original priests; or alternatively that this was an early branching that just happened to experience a run of early mutations.

But the bottom line is that for your claim to be verifiable, you data needs to be verifiable. Jheald (talk) 00:01, 5 March 2009 (UTC)

"(4) Two points here. Walsh gave calculations for 12, 25 and 37 marker haplotypes, and a method which could be extended to others." Jheald

It is incorrect. He cites the Walsh paper related to only two individuals with 12-marker haplotypes. Here is the refernce:

“Time to Most Recent Common Ancestry Calculator, Using Genetic Marker Similarity between Two Individuals” http://nitro.biosci.arizona.edu/ftdna/TMRCA.html

In that reference Walsh clearly indicated that the more haplotypes in the series, the more accurate the obtained timespan to a common ancestor. I have used 25-, 37- and 67-markers haplotypes, and the series of 37-marker ones (where the “CMH” splits into two lineages) was calculated using 85 37-marker haplotypes. It is more that people normally use.

Yes, longer haplotypes would give different numbers. But there would still be quite a high uncertainty. It is a typical “dismisser’s” language. It is a very non-productive discussion when it is conducting in such a manner. Haplotypes give pretty good estimates. 12-, 25-, 37- and 67-markers haplotypes with calibrated average mutation rates give quite consistent results, often within 200-400 years error margin coming back to several thousand years. With such an “uncertainty” as those dismissers proclaim, there would never be such consistent data. As an example – I used 15 Basques R1b1b2 haplotypes, and calculated mutations using several different methods (including ASD). All of them gave me 3600 years to a common ancestor. Then I took 750 haplotypes from the recent paper on Iberian haplotypes (the Adams et al paper), and obtained 3625 years to a common ancestor. When is his “uncertainty?”. Haplotypes with their mutations are stubborn things. They give quite reproducible data. Secondly, yes, you can use lots of haplotypes. But as you go back in time, these rapidly converge together. Most of the uncertainty comes in the time for the last stages of the convergence. Eventually, you say further than 2000 years back, you are dealing with only two lines. Therefore you would expect uncertainties from this part of the history similar to those in Walsh's calculations. It is a typical ignorance in dynamics of processes, whether haplotype or chemical reactions. The dismisser Jheald does not understand that he or she is talking to a professional chemical kineticist. One can tell that the last two molecules in a vessel would not find each other, and the reaction would never complete. That is ignorance. We count a half-life time, not the last two individuals. That is what calculations of mutations are about. If there are 100 12-marker haplotypes in the series, and 20 of them are still intact, not mutated (that is, preserve the ancestor’s haplotype), than a common ancestor lived about 1975 years ago, in the 1st century AD. Here I did not count any mutations. Because I am a specialist in reaction rates. If it happens that in this series there are 160 mutations, then a common ancestor lived 1975 years ago, Now I got the same number counting mutations. When I got identical figures using two quite different approaches, WHO can tell me that “there is a large uncertainty”? I can politely indicate 1975±200 years bp, and nobody has a moral right to blame me. Particularly I get 2000 years to a common ancestor using 25-, 37- and 67-haplotypes for the same series.

Objections?

One should be nuts to object it. The only question could be on a calibration of the average mutation rate and on a correction for back mutations, but I have included all of it already to the calculations. One can ask about reliability to consider mutation rates constant for the last several thousand years, but it is already proven.

Gee, I am getting tired from those dismissers (don´t take me wrong). But I got used to it for my 40 years in science.

Here is a link to Bruce Walsh's full paper, Estimating the Time to the Most Recent Common Ancestor for the Y chromosome or Mitochondrial DNA for a Pair of Individuals, Genetics 158: 897--912. As you can see, he sets out the mathematics quite generally, and it can be adapted to any number of markers.

Anatole, you may be a professional in chemical kinetics, but you are not a professional in bioinformatics, and it seems your training in statistics is shaky. The falseness of your analogy with the "last two molecules" should scream out at you. You can ignore the "last two molecules" in chemical kinetics because you are dealing with an enormous number of molecules all acting independently; besides, you are only expecting the reaction to run to equilibrium. But this system is different. The lines aren't all independent: they are strictly hierarchical. And you can't ignore the last two, because it is their coalescence we are specifically interested in, that all the lines go through. As Ken Nordtvedt pointed out, that means you cannot treat the 100 haplotypes as independent: they are correlated. You will get a fair estimate of the expectation value of the TMRCA distribution, but you will grossly underestimate its uncertainty. No matter how many haplotypes you test, at the very least you will still have an uncertainty a few times that of the uncertainty in a perfect n/n match, if you are testing n markers - because that is the minimum uncertainty in how long it will take the last two lines to merge. And if you know the last two lines have a much larger disparity than an n/n match, the uncertainty will be so much bigger.

The second point is that we are interested in the uncertainty of the distribution, not the uncertainty of its expectation value. You may have produced three (strongly correlated) estimates of the expectation value of the TMRCA, but that tells you nothing about the width of the distribution.

I don't mean to be dismissive. But in science bad claims about statistics stink. Deluding yourself into thinking you know something more accurately than you really do is a road to nowhere. Jheald (talk) 01:07, 5 March 2009 (UTC)

" We really don't know how many generations take place between those early occurrences of nodes in the tree for the sample population. There is intrinsic noise uncertainty in any ability to infer them no matter how large the sample population (which is constrained by total population). Those intervals which contribute to the total generations are shared in common by a very large fraction of the sample population haplotype pairs , so there is no way to drive down the statistical uncertainties in their contribution to total G. It requires INDEPENDENT measurements of something to invoke the 1 / square root of N diminishment of noise errors. " (Ken Nordtvedt) " Jheald

Disagree. Dear Sir or Madam,Since I do not know you name, though you know mine, we are not in exactly equal position. However, it does not matter. On you side I see a person, who has a rather limited knowledge in my field, that is kinetics of chemical and/or biological processes, and mutations exactly fit into this broad category. To call it “bioinformatics” is ludicrous; however, I do not care either. You can call it “astronomy”; it would not change a thing. However, it reflects your level in that field of science. You see, it is easy for me to describe your lack of knowledge, since you do not have a name. Now, I am still responding, only because I think I will publish this epistolary piece as rather educational for some audience. Now, the first question which I missed completely. The question is – why we are doing those calculations? From your comment it follows that you do NOT do those kinds of calculations. However, you dare to “criticize” a person who does it by hundreds, and long ago developed an “educated intuition” what can be done and what cannot. I do those calculations for some 40 years. Having said it, I observe with bitterness how poorly people do calculations in DNA-genealogy. It is insane. And it seems that by doing this in an absolutely wrong way, their major concern is to fool themselves with “error margin calculations”. Let me explain this. In chemical kinetics, a long established field of (mutation) rates of reactions, it is very important to separate time courses of parallel reactions, or consecutive reactions, or parallel-consecutive ones. Otherwise one will obtain a mess, a bunch of (mutation) rates, some phantom parameters. That what exactly most of people do not do in mutation calculations. They grab some series of haplotypes, which represent mixtures of lineages, stemming from different “common ancestors”, and calculate something, which does not have any sense. And then they “carefully” calculate some pseudo-error margin”, being miles away with the principal result. That is what Hammer, Scorecki, Thomas et al did in their papers in Nature. In the “Origins of Lemba” paper they did not even bother to do any calculations at all, ridiculously talking on the “CMH” as “exclusively Judaic origin”. Wrong. The initial claim on the “Cohen modal haplotype” and calculations in the first Nature papers was wrong from the very beginning. It turned out that the authors took several different “CMH” lineages, recent and old, and calculated them as one pool. Wrong. The whole “CMH” premise as a “Cohen” one was wrong. This a typical Arabic lineage as well, 9,000 year-old, at least. Among the Jewish people it is some 4,000 years old, well before Cohens. Do you follow my point? DNA-genealogy is full of wrong things, wrong in principle. And a seemingly major concern is to wrap those wrong things into calculations of “error margin”! I do not deny an importance of calculations of error margins, but ONLY when the principal answer is basically correct. And even in this case one should understand what is the question addressed, and when error margins seems to be important, and when nor necessarily. I means, when sometimes a semi-qualitative result is very important. For example, there is a popular fable that current R1b (R1b1b2) folks in Europe have their common ancestor who lived some 30,000 years ago. One, on the contrary, using different populations – in the Isles, Iberia, Germany, Italy, Flemish haplotypes, etc. gets repeatedly 4200, 3600, 4400, 4500, 3800 years. Repeatedly and relatively reproducibly, using different populations. Do you understand that an error margin here would be useful, granted, but practically irrelevant to the principal question? Clearly it is not 30,000 years.

What it is important here, is to demonstrate, that one dissected those haplotypes into lineages, shown – on a haplotype tree – that it is a “homogeneous” population, having one common ancestor, that these statements are supported with quantitative criteria, that the mutation rates are calibrated, that 12-, 25-, 37, 67-haplotypes give principally the same, or similar result of calculations. That is what DNA-genealogy needs at the present time. However, “critics” do not care. They demand “error margin calculations”, despite the most ridiculous ways of principal calculations and wrong principal results. “Error margin” is what they do care about.

And you know why? Because it is easy. It is formal. Because they do not need to know more complicated things beyond their knowledge.

And, yes. One more thing. They do not want to acknowledge that those “error margin calculations” are themselves based on wrong or arbitrary conditions. There is a lot of assumptions in those calculations.

I prefer to use ACTUAL numbers, without those assumptions. I know that when haplotypes in the series are generated from ONE common ancestor for all haplotypes in the series, a two criteria should be consistent to each other: a number of base (non-mutated, presumably ancestral) haplotypes and a number of mutations. They should give the same or similar numbers of generations to a common ancestor. If both of them are the same, I got the right number. And it is ridiculous to tell me that my number has a huge “uncertainty” such as between 500 and 2500 years. People who say it do not understand principles of mutation rates and principles of working with right series of haplotypes.

And cone more thing. Let me teach you a principal thing in science. First, you line up what is known about your subject. Regarding Jewish haplotypes – write on a sheet of paper KNOWN facts and findings. Some of them might look as conflicting with each other, it is O.K. It is life.

Then, try to come up with an OPTIMISED version which would unite all those findings, or most of them, into a tentative (always tentative!) concept. You can call it a working hypothesis. Try to give an explanation why there are “facts”/findings which conflict with the others: either your concept is too narrow, or you believe that the “facts” are erroneous ones.

Having this concept, try to explain other, new findings. You might need to extend your concept to include new facts and findings.

You concept can be imperfect. In fact, it always is. However, it is your concept, your base, opened for corrections and further upgrading.

That is what I have done with the Origin with the Jews, based on their haplotypes.

If you want to change it, to modify, to upgrade, give me DATA, or a new INTERPRETATION. It is justified, I gladly accept it. It will make my concept better.

However, what are YOU trying to do? You do not give me new data. You do not give me new interpretation of know data. You just try to dismiss my concept. And not even concept! You just toothlessly trying to say that I do all wrong. What specifically? --Scienceprogenetic (talk) 16:57, 5 March 2009 (UTC)

Generally, yes. However, read again that example I just gave above. When one does calculations by different methods, and repeats them with 12-, 25-, 37 and 67-marker haplotypes, and verifies them using my logarithmic method, and constructs a haplotype tree to see how these haplotypes are related to each other, then you know that those general concerns are neutralized by your multi-tier calculations. That is why I consider an “error margin” based not on some imaginary (and self-deceiving) considerations, but on actual variability of actual numbers. "Last I heard, Anatole had not developed an error analysis for his estimation technique. The most he ever showed on this list was a comparison of two or more non-independent estimates from the same or similar data sets and an assertion that the spread of the estimates was the uncertainty in the technique." (John Chandler) I do not care what other people say, because themselves did not develop an error estimations. John took – arbitrarily, I presume - 15% error margin for his mutation rates in 12- and 37-marker haplotypes. The same figure. Why the same 15%? He did not describe. And than he uses those mystical 15% as a base for follow-up calculations. Is it science?

As I said, you're confusing estimates of the expectation value of the TMRCA distribution and estimates of its spread. The uncertainty in the expectation value is not the same as the uncertainty in the whole distribution.

John's mutation rate calculation is not relevant here, but if you look at the paper he wrote on it, you will see he gives quite a detailed error analysis, and the errors don't always come out to be 15%. What he says above is exactly right. You are calculating the spread of your estimates of the expectation value of the distribution. That is not the same as calculating the spread of the distribution itself. Jheald (talk) 01:26, 5 March 2009 (UTC)

_____

I pretend to add this new section in the article, because I know that I can contribute with Wikipedia and my knowledge, studies and publications reflects the most consistent information based only in reliable sources. Dear Wikipedia´s Editors, I think this is enough. You can see yourself a “level” of the questions asked and “reasoning”. I prefer to have a discussion with an opponent from whom I can learn something. Not with a dismisser. Otherwise it is just a waste of time. Let me now work in this new section. Regards for all of you. --Scienceprogenetic (talk) 20:31, 4 March 2009 (UTC)

J2 Haplotypes

Now, to J2 haplotypes. It is a lineage completely different from J1, however, very many Jews belong to this lineage, that´s why it has its own J2-CMH which many Jews descend from:

Hg		DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii
J1	CMH-6	12	23	14	10				16			11
J2	CMH-12	12	23	14	10	13	17	11	16	11	13	11	30

By a game of statistics, the J2-CMH is identical with the shorter, 6-marker J1 "Cohen Modal Haplotype" (which, as we already know, has nothing to do with Jews with J1-CHM), and it is different with their 12-marker haplotypes, both of Jews J1 and Jews with J1-CMH. Furthermore, it has nothing to do with J2 Cohanim haplotypes, which split into J2 Cohanim Ashkenazim and Cohanim Sephardi, with the ancestral haplotypes:

Hg	COHANIM	DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii
J2	ASKENAZI	12	23	15	10	14	17	11	15	12	13	11	29
J2	SEPHARDI	12	23	15	10	13	18	11	15	12	13	11	29

Within the unique cluster J2 Cohanim haplotypes have slight mutations, as it should be after hundreds of years passed from their common ancestors. Both J2 Cohanim Ashkenazim and J2 Cohanim Sephardim have their common ancestors who lived about 900 years ago, in times of the First Crusade.

Comment. As we discussed before, while there are clearly large numbers of J2 Ashkenazi Cohens clustered around the above haplotype (Behar's data would suggest perhaps 20% of all Ashkenazi Cohens), there simply isn't enough data to identify a haploype for Sephardi Cohens. Indeed, if one looks at Behar's data, even though it's only on 6 markers, about 75% of haplogroup J Sephardi Cohens match the CMH, but beyond that there is no clear clustering at all. It is also notable that none of Behar's sample matched this haplotype you are proposing.

Yes, you have now found three Cohens with this signature who presumably have a common ancestor. But that does not show they are reflective of Sephardi Cohens as a whole. Indeed, as the haplogroup breakdown indicates, there are small groups of Cohens with all sorts of ancestries. This is very possibly simply another small group. It is interesting that are genetically not so far from the Ashkenazi group, but I understand the best predictors only suggest a 70-30 chance of that match; and that there is a Cohen with the name Pereira with this haplotype who has been tested who definitely does not match the Ashkenazi group (having different SNPs).

This is an area which I hope will change, but at the moment it seems to me most appropriate to say we simply have very little broad data on the haplotypes of Sephardi Cohens who are not J1 CMH.

One further thing. If we are reviewing Cohen haplotypes in J2, we should include the J2b2e Cohens, who appear to account for about 10% of Ashkenazi Cohens. Jheald (talk) 12:28, 2 March 2009 (UTC)

Disagree:::: “Behar's data would suggest perhaps 20% of all Ashkenazi Cohens” Jheald, your term “would suggest” and "perhaps" is not appropriate. Your point of view is not welcome in here, nor in the article. You cannot use Behars data and suggest things for yourself, because first, who are these Askenazi Cohens that you presumed? Second, you cannot make an interpretative thesis on any data. This is a very common wrong move.

“Indeed, if one looks at Behar's data, even though it's only on 6 markers, about 75% of haplogroup J Sephardi Cohens match the CMH”

Again Jheald, you cannot make an interpretative thesis based on Behar´s data. Does the author come to that same conclusion? It is stated in the article? No, it is not. Second, who are those “J Sephardi Cohens that matches the CMH”? Can you list their surnames, please? Are these surnames recognized as Cohanim in Sephardim Community? Can you provide a citation that says that these “J Sephardi Cohens” surnames are linked to Cohanim house?

“Yes, you have now found three Cohens with this signature who presumably have a common ancestor. But that does not show they are reflective of Sephardi Cohens as a whole”

Jheald, are you familiar with the Cohanim tradition among Sephardi Community? There are only three or four names associated to Cohanim houses among the Sephardim. The (Cohen)Rodrigues and Cohen-Pereiras or Cohen-Parairas, are very recognized among the Sephardi Community. It is not a coincidence that they all present the same 12 markers on their haplotype, and only two step from the Askenazi Cohanim. Sephardim does not use Cohen, or Kohen, or Katz in their surnames. It is not a coincidence that those three families are very recognized in Sephardi community as Cohanim, and they all share the same 12 markers. “who presumably have a common ancestor”. Jheald, you cannot ignore the pedigree of these families, their traditional surnames associated to Cohanim house, and the extensive records that shows their pedigree. Here is only one link to the Sephardi Cohen Rodrigues: http://shum.huji.ac.il/~dutchjew/genealog/cohen_rodrigues/index.htm

There are extensive records among the Sephardi Cohanim families Cohen Pereira or Paraira, Cohen Rodrigues, Cohen-Machado. Just ask any Sephardi rabbi, and they will tell you who are they. Don’t judge for yourself, especially when seems that you do not have any contact with the Sephardim Community in Portugal-Spain. --Chris Cohen (talk) 00:58, 3 March 2009 (UTC)

Chris, if you are so confident, then take the test and put the question beyond doubt.

On the basis of the genetics, on the markers we have at the moment, we can't be so certain. As I wrote above, on twelve markers the best predictors only suggest a 70-30 chance of that match; and there is one Cohen with the name Pereira with this haplotype who has been tested who definitely does not match the Ashkenazi group (having different SNPs).

My understanding is that Behar's Sephardi Cohen samples were primarily recruited from the Sephardi community now in Israel. If you read the sequence of papers, they were careful to only take people who could prove their religious Cohen status. I do not know the surnames because he didn't publish them, but I am confident that they were thorough about checking their credentials. It is possible that their ancestors may have tended to be primarily from Mizrahi countries, rather than Spanish and Portuguese, so I can see that that might be a possibility as to how if there is a distinctive Spanish and Portuguese Cohen genetic signature, it did not appear in Behar's Sephardi Cohen sample. Jheald (talk) 10:30, 3 March 2009 (UTC)

Hi Jheald. How are you doing today? I am trying to understand the main points of your arguments. I am confident with mutations for sure, but not with your own theories. Let me be honest with you. As a matter of fact, I don’t believe in L24 at all, because there is nothing published about it. You guys made important statements about this L24 when the founder of this L24 already said that those who has it `` whose population history remains to be determined”. You did clearly include original research in the article. I don’t believe that because one step mutation on DYS 445 and DYS 450, is not enough to separate 13-18s from 14-17s in a time frame after 3.000 years ago. Can you provide me any official scientific publication that states such a thing? Again, 95% of Sephardi has 450=8, while most of Askenazi has DYS 450=9. So what Jheald? Please, gives us citations from ``official scientific publications only```` that states that this particular allele DYS 450 and or DYS445 is responsible to put a separation line for over 3.000 years between Askenazim and Sephardim. Regarding the samples of Behar`s analyses the author did not appointed Cohen Sephardi or Cohen Askenazi. Your understanding is that the samples came from Sephardi Community in Israel whom ancestral are from Mizrahi countries. As you can see, neither you nor the author explained such thing. Now, you made a good point by saying that the distinctive Spanish and Portugues Cohen genetic signature (13-18) did not appear in Behar`s Sephard samples because of the ancestors of Mizrahi countries. Nice point. :-). This is what happened. ``Out of 174 samples of Sephardim of various haplogroups in Iberia, only 9 - on average - might have been the Cohanim. However, merely 81 samples were determined as of haplogroups J1 and J2. On average, there might have been only four Cohanim among them, two J1 and two J2. In fact, among all 43 of J2 Sephardi individuals not one belonged to the J2 Sephardi Cohanim haplotype cluster. Jheald there is nothing wrong with that. The DNA analysis has shown that the J2 Cohanim Sephardim haplotype is not in the mainstream of J2 Sephardi haplotypes. Obviously, a number of tested Sephardim in that study was not statistically significant. Cohanim haplotypes are not supposed to be in the mainstream of J2 Sephardi haplotypes at all. J2 Cohanim haplotypes 13-18 are very close related to J2 Cohanim 14-17 Askenazim. “J2 Cohanim haplotypes, being within their Ashkenazi and Sephardi clusters, noticeably differ from J2 haplotypes of "regular" Jews, who do not belong to the Cohanim lineages”. This is the point. Cohanim are different from regular Jews. -- Again, withou surnames you will never identify the true Sephardi Cohanim, because they are very exclusive, and only three to four families carry the Cohanim tradition among the Sephardim. These are the (Cohen)Rodrigues, Cohen Pereira or Cohen Paraira, Cohen Machado). Please, give us any citation that says that there is any other Sephardi surnames among Cohanim today in Sephardi Community in Portugal - Spain. Again, the only three families that has the tradition and it is recognized among Sephardi community in Portugal and Spain are these three families mentioned. Not a case of coincidence, that these three families has a full 12 match. Please, send me any citation or any other kind of information that contains three recognized Cohanim surnames among Sephardi Community in Portugal and Spain that has a full 12 macth and also, with one-step mutation from Askenazi J2 Cohanim. [User:Chriscohen|Chris Cohen]] (talk) 19:02, 3 March 2009 (UTC)

Science and Religion

The above is science. Now, how to come in terms with religion? The whole experience of science tells us that to merge science and religion is an unproductive endeavor. Indeed, we know for fact, that J1-CMH Jews and J2 Cohanim are stemming from two practically unrelated common ancestors. The J2 Cohanim lineage is coming from the times of Aaron, Pinchas and then Zadok, at the break of the first to second millennium BCE. There is a historical as well as a Y-chromosomal basis to believe that J1-CMH Jews are descendants of Bustenai, who lived about 1300 years ago, was not a Cohanim, and who himself was a direct descendant of King David, thus J1-CMH Jews can consider themselves as direct descendants of King David, and J2 Cohanim as direct descendant of Aaron Hakohen, Pinchas, Zadok. All it shows that the Jewish heritage is absolutely exciting ones, and there is nothing less than to be very proud of it.

Comment. This kind of paragraph, especially its editorializing tone, is particularly likely to be seen as original research, and not reflecting a neutral point of view -- both key WP policy points.

Broad sweeping comments like "to merge science and religion is an unproductive endeavor", if they are to be included at all, have to be sourced to particular individuals or works.

Specific suggestions like "J1-CMH Jews are descendants of Bustenai" also need to be clearly attributed. Who has made this claim? On what evidence? Jheald (talk) 12:50, 2 March 2009 (UTC)

J2 Cohanim

The Cohanim J2 haplotype is noticeably younger (in terms of time spans to their common ancestors) compared to J1 CMH. The ancient common ancestor for the J2 Cohanim lived 3,100±200 years before present (BP), exactly around the date of the birth of Pinchas, the grandson of Aaron Hakohen. The youngest common ancestor of J2 Cohanim (Sephardim and Ashkenazim) lived about 900 years before present (BP), apparently, during the First Crusade. On the other hand, Jews with J1-CMH shares a common ancestor who lived 4,000±200 years BP, around one thousand years before Pinchas. J2 Cohanim and Jews with J1 CMH are descendants of two different ancestor of which J2 lineage has the accurately coalescence to Aaron Hakohen – Pinchas, and J1 Jews with CMH an ancestor who lived 4,000±200 years BP, around one thousand years before Aaron Hakohen - Pinchas. J1 Jews with CMH do not have the same ancestor in a time frame of Aaron Hakohen (3,100±200 years BP).

J2 Haplotypes

The Jewish haplotypes of haplogroups J1 and J2 do not mix with non-Jewish haplotypes from the same haplogroups on the cluster tree. In other words, the Jewish haplotypes have their own exclusive branches. As a result, they have developed quite distinct lineages in haplogroups J1 and J2. Jewish J2 haplotypes do not mix with non-Jewish J2 haplotypes and J2 Jewish haplotypes do not mix with the J2 Cohanim Jewish haplotype on a haplotype tree. Jews have two modal, or ancestral haplotypes in haplogroup J2. One directly related to regular Jews, and other directly related to Cohanim Jews. This means that J2 Cohanim Jews form their own exclusive cluster in J2 Haplogroup and regular Jews are not part of it.

Comment Again, the TMRCA estimates depend very sensitively on who is included in the group for which the TMRCA is being calculated; and secondly, the small uncertainties quoted are hopelessly unrealistic.

As far as I am aware, both the J1 and the J2 main "Cohen clusters" include both Cohen and non-Cohen Jews. It would be quite surprising if this were not so, as presumably whoever were the original priests came from families that also had other, non-priestly descendents; and since that time, many ancestral Cohen may have lost or forgotten their Cohen status.

The claim the J2 Cohens "do not mix" is particularly odd, especially if you are including the Cohens from the Brazil group in with them, as the Ashkenazi ones form part of the J-L24 group which contains all sorts of people. Jheald (talk) 13:05, 2 March 2009 (UTC)

Disagree::: Jheald what is this thing called L24? Do you have any scientific journal that published any full article that could explain what it is, what exaclty this thing represent? Are you talking about the “rs34126399G” discovered by Dr. Roy King? Jheal this allele is a poorly understood thing. The rs34126399G is “possibly” related to the spread of agriculture in the Mediterranean. You cannot use this L24 to separate Askenazi from Sephardi for 5.000 years? You must be kidding. Look what Dr. Roy King himself wrote about this particular “rs34126399G”

“As I wrote in the blog, rs34126399G captures the circum-Mediterranean migration of the first farmers in Europe and the Near East. Those individuals who are J2a and ancestral at rs34126399, although less frequently present in European and Mid-Eastern populations, do constitute a fascinating subgroup “whose population history remains to be determined”.

Jheald, are you now using “agricultural” thesis to separate Askenazim from Sephardim? Why are you determinating things when the author knows that this is very inconclusive?

“The pattern of spread of agriculture in the Mediterranean is very interesting, although hardly surprising for three reasons. First, after last Ice age human as well as other animals started migrating out of Africa, along the Fertile Crescent into Europe and Asia; and the pattern exists for several undomesticated animals too. Second, the east-west spread of the plain provided favourable climate, and facilitated human to reproduce their newly developed skills of agriculture as they migrated. There are examples in other parts of the world (e.g. south America) where art of agriculture has been reinvented over and over again. Third, rapid deterioration of climate in sub-Saharan Africa and Arabian peninsula, and retreat of glaciation in Europe maintained a steady stream of migrating people, giving rise to the currently observed gradient. “Such events, in combination favoured the spread of agriculture from the Fertile Crescent to the Mediterranean Europe”.

If you look at this project in FTDNA: Ashkenazi Jewish males with Sephardic roots

http://www.familytreedna.com/public/IberianSurnamesofAshkenaz/default.aspx?section=yresults

"There are Ashkenazi families in Eastern, Western, Northern and Central Europe who have Spanish or Portuguese surnames, an oral history of having Sephardic ancestors.”

Jheald, around 95% of those Sephardim has DYS450=8 while Askenazim has the modal of DYS=450=9. This one step mutation, on this specific allele cannot in any thesis, separate two individuals for a time frame of 5.000 years? Do you have a official scientific citation where it says that L24 can separate two individuals for 5.000 years? Jheald, this whole thing is very inconclusive. You cannot use it in the article, nor in your comments. Please, give scientific citation and official references. Don’t jump to conclusion. This is not a right thing to do. Forget about L24 or rs34126399G. This still a very inconclusive thing. --Chris Cohen (talk) 01:34, 3 March 2009 (UTC)

Chris, you can't just wish evidence you don't like away.

"rs34126399G", which Dr Roy King writes about [8], is an SNP called L27, which defines the J2a1 subgroup of J2. [9]

L24 is a different SNP mutation, also called "rs35248080". It defines the more specific subgroup J2a1k. The ancestor of the 14-17 Ashkenazi Cohens had this mutation, and belonged to this group.

Some things are clear about these SNPs, some are not. What is clear is where they are on the haplogroup tree, and what is fairly clear is what haplotypes correspond to them, at least in the Western communities that have taken to genealogical DNA testing. From the differences in those haplotypes we can estimate an age. You can see L24 haplotypes in the Haplogroup J project at FTDNA. [10]. The various J2a1k and J2a1k1 clusters are all part of the wider J2a1k group. As you can see, there is quite a range of STR haplotypes. This points to the kind of age for the L24 mutation that I indicated.

This much is clear from the data we have. What is more tentative, as Dr King rightly indicates, is which peoples and migrations the haplogroups might have been associated with in prehistory. That is a different sort of question, and needs much more thorough geographical sampling to get clues to, and even then can be quite conjectural.

But questions of how L27 and below it L24 divide up haplogroup J2 are much more clear-cut, as is at least a rough age for these mutations.

Note that I'm not using an "agricultural theory" to divide the Ashkenazi and the Sephardi. I am simply going by the values of the 12 markers. Those indicate only a 70-30 chance that the haplotypes belong in the L24 group. Jheald (talk) 11:10, 3 March 2009 (UTC)

Hi Jheald, how are you? We had great moments last year discussing some issues about the article. Of course me and all Sephardi Cohens 13-18 will extend the markers very soon. We are now inviting much more families into the project. This is our goal rigth now. We will extend the markers, but not to satisty your curiosity. Again, since you state so much Walsh, 13-18 has 95% confidential interval to share the same ancestor in a time frame of 245 years to 2095 years. (This is using 35 years per generation, if we use 25 years per generation than the time frame is shorter) Dear Jheald, again, if you cannot list official publication about this L24, please Jheald, dont use original research, and dont jump to conclusions. Again, 13-18 does not have to match the alleles 450 or 445 in order to be related in a time frame after 3.000 years. Do you have "any" official publication that you can list as a reference? Please, don´t considerate your own project J Haplogroup and the conclusions of your team. Here you have to add real references, like a article published by a scientific journal, not your arguments. Dont take me wrong, but your statments about "J2a1k" and "J2a1k1" are not valid here. Do you have any official publication that states and explain theses "J2a1k and J2a1k1". This is creation of your imagination. Never, any scientific oficial journal published such things like that. Can you list "any" official publication with "J2a1k and J2a1k1". Did Behar use this things? No. Did any other professional used this terms "J2a1k and J2a1k1"? No. Please, show the citations, if not, don´t use on the article, nor here as a fundamental argument. This simply doest no work at all. Take care.. --Chris Cohen (talk) 17:19, 3 March 2009 (UTC)

The designation J2a1k is from the ISOGG official current tree [11], which Wikipedia considers a reliable external source.

The same information, showing where L24 fits into the J2 tree, can also be seen on the website of Thomas Krahn, lead researcher at FTDNA [12]. Jheald (talk) 18:06, 3 March 2009 (UTC)

Comments. Hi Again, Jheald. We are finally getting on the main point. :-) You use this L24 and "J2a1k" and "J2a1k1" as the base of your arguments all the time. For you ISOGG has endorsed the very inconclusive discovery about this new thing. However, make no mistakes about it. Read carefully on their website. The Working Draft of the Y Chromosome Tree Below J-M304 “does not represent an endorsement by ISOGG”. Why are you using working drafts that is very inconclusive and poorly understood that has no endorsement by ISOGG? We need to call Wikipedia`s editor to evaluate this information that has no endorsements, except by Jheald, and have not presented official publication that states such inconclusive working draft meaning, and that has no results and endorsement by ISOGG.

Second, as you wrote “The same information, showing where L24 fits and into the J2 tree, can also be seen on the website of Thomas Krahn, lead researcher at FTDNA” Again, the information shows where L24 sits into the J2 tree, but Thomas Krahn, even being a researcher like thousands that never published official articles on scientific journals, he himself has not made any announcements about what this "L24", "J2a1k" and "J2a1k1" mean by the end. Thomas Kranh did not say that "L24" or "J2a1k" and "J2a1k1" can separate two individuals in a time frame after 3.000 years. Neither there is any official publication about it, because again, this is very inconclusive, has no endorsements. Why is you, Jhead, the only one that endorse "L24" "J2a1k" and "J2a1k1" into the article, and use as fundamental arguments. Why are you, Jheald, the only one who jump to conclusions about this L24 and "J2a1k" and "J2a1k1"? Why do you think that this "L24" and "J2a1k" and "J2a1k1" can separate Askenazim 13-18 from Sephardim 14-18 in a time frame of 3.000 years? Can you provide any official publication? Of course you cannot list any official publication, besides this two references websites links that has no endorsements, because neither the founder of L24 Dr. Roy King (who said on his latest article "Those individuals who are J2a and ancestral at rs34126399, although less frequently present in European and Mid-Eastern populations, do constitute a fascinating subgroup whose population history remains to be determined.) , nor ISOGG, nor Thomas Krahn has endorsed anything about it. This is a very inconclusive thing. Jheald, you are not above the facts and cannot jump to conclusions. Also, you cannot insert information like that in the article and even wrost, use your false precipiate conclusions to dismiss the work of credible and very recognize scienfic professionals as Mr Anatole Klyosov, author of a series of articles that was published in oficial scientif journals, along with dozens of pattent registred in the World Academy of Science. Regards. --Chris Cohen (talk) 19:48, 3 March 2009 (UTC)

J2 Cohanim Haplotype Tree

The Cohanim Haplotype Tree

The Cohanim J2 haplotype is noticeably younger (in terms of time spans to their common ancestors) compared to Jews with J1 CMH. The ancient common ancestor for the J2 Cohanim lived 3,100±200 years before present (BP), exactly around the date of the birth of Pinchas, the grandson of Aaron Hakohen. On the other hand, Jews with J1-CMH shares a common ancestor who lived 4,000±200 years BP, around one thousand years before Aaron Hakohen. J2 Cohanim and Jews with J1 CMH are descendants of two different ancestor of which J2 lineage has the accurately coalescence to Aaron Hakohen – Pinchas, and J1 Jews with CMH an ancestor who lived 4,000±200 years BP, around one thousand years before Aaron Hakohen - Pinchas. Below we can see the exclusive J2 Cohanim haplotype tree formed by the traditional Cohanim surnames recognized in both Askenazim and Sephardim communities. This is how the scientific genetic computer software placed them, considering their haplotypes. The computer program composes a tree by analyzing mutations, mutations rates, and how many steps are needed to get from a common ancestor for the entire tree to each haplotype. The more mutations, the more levels, the higher the branch (or a haplotype). Only one different Cohanim surname was used to represent the modal haplotype of each present-day Cohanim family both in Askenazim and Sephardim communities.

007	Cohen	Family	J2	12	23	15	11	14	18	11	15	12	13	11	29
008	Kahan	Family	J2	12	23	15	10	14	17	11	15	12	13	11	29
009	Rapaport	Family	J2	12	23	15	10	14	17	11	15	12	13	11	29
010	Kaplan	Family	J2	12	23	15	10	14	17	11	15	12	13	11	29
011	Katz	Family	J2	12	23	15	10	14	17	11	15	12	13	11	29
012	Cohen-Pereira	Family	J2	12	23	15	10	13	18	11	15	12	13	11	29
013	Cohen-Rodrigues	Family	J2	12	23	15	10	13	18	11	15	12	13	11	29
014	Cohen-Machado	Family	J2	12	23	15	10	13	18	11	15	12	13	11	29
015	Kohn	Family	J2	12	23	15	11	14	19	11	15	12	13	11	29
016	Kovacs	Family	J2	12	23	15	11	14	19	11	15	12	13	11	29
017	Kohen	Family	J2	12	23	15	11	14	18	11	15	12	13	11	29
018	Kunha	Family	J2	12	23	15	11	14	16	11	15	12	13	11	29

The upper branch (12,13,14) is the Sephardim branch. All three Cohanim families are from Portugal. In Ashkenazim branch, they have two low branches. One branch has a common ancestor who lived 275 years a go, in 18th century. There is only 1 mutation in 48 markers. The second branch came from one common ancestor who lived around 925 years ago. They have 4 mutations on 60 markers. In fact, one can see two "younger" sub-branches there. Sephardim Cohanim have five mutations per 36 markers, or 2075 years from a common ancestor. In fact, all three lineages of Cohanim are coming from one common ancestor. The height of the branch shows how early the common ancestor lived compared to other branches. The Cohanim Sephardim is the oldest lineage compared to Cohanim Ashkenazim. The common ancestor for all Cohanim branches, formed by Cohanim Askenazi and Cohanim Sephardi, lived 2175 years ago.

Comment. Best estimates are that the 13-18 haplotypes have only a 70-30 chance of being closely related to the 14-17 Ashkenazi ones - otherwise, if they are not part of the J2-L24 group the lineages are likely to have been separated in excess of 5000 years.

Also, it is simply silly to assign "family haplotypes" on the basis of single samples. If you look at a bigger sample, you find that each of the Ashkenazi Cohen names in this overall cluster shows a range of haplotypes. Note also that the names are largely generic, eg "Kaplan" is Polish for "Priest", "Katz" is a 2-letter Hebrew abbreviation for "Kohen Zedek" (righteous Kohen), etc., so a number of different Kohens at different times are likely to have adopted these names, the same family name probably reflects different family lineages to different Kohens that adopted it. It is simply not the case that all the Kaplans reflect one lineage of Cohens, all the Katzes a different one, etc. Jheald (talk) 13:18, 2 March 2009 (UTC)

Disagree::: Jheald, the graphic above reflects the modal haplotype of each J2 surname family listed. Look at the Katz project. 95% of the Katz there has the 12 markers above. The "Kahan" has only two families, one J2 with the same 12 markers above. The rapaport the same thing just like the Konh, Kovacs, Cunha. Jheald, it is too much of a coincidence that they all presents full 12 matches between them? This is fact. Please, can you provide the “range” of the haplotypes among these particular families? Can you show other results? Again, in order to work with Cohanim, one must look in the surnames above, and clearly see the close relation between all of them. This is a fact. Don’t disqualify this graphic, please. There is nothing wrong there. Second, again, Jheald is using L24 to separate two individuals for 5.000years. Walsh graphic shows that there are 95% interval confidence of 13-18 sharing the same ancestral with 14-17 in a time frame between 245 years to 2095 years. Jheald, where did you get the conclusions from L24 or J1ak1 or J2k1ak? Can you provide us any oficial scientic publication that has endorsed such thing? If not, don't use it again, neither here or in the article.

READ COMMENT ABOVE--Chris Cohen (talk) 01:55, 3 March 2009 (UTC)

J2 Cohanim 6 and 12 marker modal (ancestral) haplotypes

The J2 Cohanim 6-marker modal is 15-15-23-10-11-12. The extended 12-marker modal for J2 Cohanim is 12-23-15-10-X-Y-11-15-12-13-11-29 with X=14 and Y=17 for J2 Askenazi Cohanim and X=13 and Y=18 for J2 Sephardi Cohanim. With the extended 12 markers for J2 Cohanim, it was clearly revealed that both J2 Askenazi Cohanim and J2 Sephardi Cohanim do share 10 out of 12 markers. Such tight connection clearly indicates that both J2 Cohanim Ashkenazi and J2 Cohanim Sephardi had a common ancestor not a long time before their separation. The only real differences found among them are located on two specific markers: DYS 385a, and DYS 385b. The variation found has just one-step mutation in each of them. The second Jewish haplotype found in J2 haplogroup is attributed to regular Jews, and is derived from a much older lineage, with a common ancestor who lived over 10 thousand years ago. This lineage continued 6,000 years BP in people, who will became Jews two thousand years later, which is exactly Abraham times, and 1200 years ago a modal haplotype 14 16 23 10 11 12 had appeared in haplogroup J2, which accidentally, as a result of game of statistics, looked exactly as the J1 "Cohen Modal Haplotype". In haplogroup J2 it has nothing to do with either J2 Cohanim or J1 with CHM, and cannot have a common ancestor that could live in a time frame of Aaron Hakohen (3,100±200 years of BP).

Comment. This is simply repeating what you have already said. There is still only a 70-30 chance of the 13-18 haplotypes being related to the main 14-18 group. There are a lot more than only two clusters in J2, if you are looking at this level of detail. And the claimed, very small, uncertainty margins are still completely WP:OR and completely bogus. Jheald (talk) 13:24, 2 March 2009 (UTC)

Disagree::: "There is still only a 70-30 chance of the 13-18 haplotypes being related to the main 14-18 group". Where did you get that from? Based on the mutation rates used by Wash, the 13-18s and 14-17s have a 95% confidence interval of sharing the same ancestor in a time frame of 245 years to 2900 years only. This is exactly after the times of Aaron Hakohen who lived around 3.300 years ago. --Chris Cohen (talk) 01:59, 3 March 2009 (UTC)

The problem is that these 13-18 Cohen haplotypes are also similarly close to haplotypes which are known to belong to J2a*, not J2a1k. So these could as easily be related to J2a*, as J2a1k. The only way to know is to take the test. Jheald (talk) 11:16, 3 March 2009 (UTC)

Jheald, where did you get the conclusions from L24 or J1ak1 or J2k1ak? Can you provide us any oficial scientic publication that has endorsed such thing? If not, don't use it again, neither here or in the article. And dont use it to disqualify informations and facts.

Comment. See previous comments.--Chris Cohen (talk) 20:59, 3 March 2009 (UTC)

Cohanim Exclusive Cluster

Non-Jewish J2-haplotypes take an opposite side of the J2 haplotype cluster tree from the Jewish J2 haplotype. Furthermore, almost all Jewish haplotypes (J1 and J2 separately) have taken the opposite sides when compared with non-Jewish haplotypes on the tree. It also can be seen that non-Jewish J2 haplotypes are older (more distant from the tree base) than Jewish J2 haplotypes. Indeed the oldest Jewish J2 haplotypes are related to regular Jews and their ancestors from about 6,000±1,000 years ago.

J2 Sephardi Haplotype

There are 43 different J2 haplotypes among 174 samples collected among J2 Sephardim of Iberian peninsula (Adams et al., 2008), and these 43 J2 haplotypes split into 8 genealogy lineages, or branches. This can be compared with just one separate and exclusive branch that was formed by the J2 Sephardim Cohanim. Each of the 8 branches is a genealogical lineage of J2 Sephardim. A common ancestor of 8 branches formed by the J2 regular Sephardim lived about 11,500 years ago. On the other hand, the ancient common ancestor for the single exclusive branch formed by all J2 Cohanim lived 3,100±200 years before present (BP), exactly around the date of the birth of Pinchas, the grandson of Aaron Hakohen. The DNA analysis has shown that the J2 Cohanim Sephardim haplotype (X=13 Y=18) is not in the mainstream of J2 Sephardi haplotypes, but has a very unique and directly related haplotype with J2 Cohanim Askenazi Haplotype (X=14 Y=17). Hence, it was formed one exclusive single cluster. This is because both Cohanim Jews, Askenazi and Sephardi, for thousands of years preserved their genealogical lineages since the Temple period. Of course they did not think in terms of the DNA, however, they have faithfully followed Cohanim Jewish tradition until present day. As a result this is exactly what the haplotype tree shows.

Comment. See previous comments. The date, the uncertainty in the date, and the proposition that this is an "exclusive single cluster" are all bogus. And becoming irritatingly repetitive. Jheald (talk) 13:30, 2 March 2009 (UTC)

J2 Sephardi Cohanim Haplotype

J2 individuals with the following 10 markers 12-23-15-10-X-Y-11-15-12-13-11-29 belong to J2 Cohanim modal haplotype, and they are part of the J2 Jewish haplotype tree as a whole. However, within this J2 Cohanim signature, there are two distinct clusters, that is Ashkenazi Cohanim (ancestral X-Y = 14-17) and Sephardim Cohanim (ancestral X-Y = 13-18). These clusters form the exclusive J2 Cohanim Jewish Haplotypes. Cohanim haplotypes within the clusters may have some (few) mutations, because mutations in 12-marker haplotypes occur randomly every millennium, and some people have two mutations in their 12-marker haplotype after a thousand of years, some people have one mutation, and some people have no mutations after a thousand years. Hence, there are little variations between haplotypes within a cluster, but not much. So, J2 Cohanim split into two lineages, with 14-17 and 13-18 in their markers 385a,b, which nicely places them into Ashkenazi and Sephardic communities. At the same time, J2 Cohanim haplotypes, being within their Ashkenazi and Sephardi clusters, noticeably differ from J2 haplotypes of "regular" Jews, who do not belong to the Cohanim lineages.

Comment. as per comment above. Jheald (talk) 13:30, 2 March 2009 (UTC)

Statistics

According to the tradition, and also taking into account statistics and cemeteries tombstones, the Cohanim represent approximately 5% of the total of Jews. Out of 174 samples of Sephardim of various haplogroups in Iberia, only 9 - on average - might have been the Cohanim. However, merely 81 samples were determined as of haplogroups J1 and J2. On average, there might have been four Cohanim among them, two J1 and two J2. In fact, among all 43 of J2 Sephardi individuals not one belonged to the J2 Cohanim haplotype cluster. By adding 3 new samples of J2 Sephardi Cohanim found in another database, not listed on the first 43 J2 Sephardi "regular" Jews (Adams et all, 2008), the frequency of J2 Sephardi Cohanim among J2 Sephardi Jews reaches a normal rate around 7%. Obviously, a number of tested Sephardim in that study was not statistically very significant, however, it gave reasonable estimates.

Presumably this is meant as an apologia for why none of the samples Behar provided to Adams matched your preferred Sephardi Cohen haplotype. The idea that you can add three samples you found in your armpit one midsummer morning to a balanced sample in a published study, and then say anything statistically useful, is particularly ludicrous. Jheald (talk) 13:35, 2 March 2009 (UTC)

The Haplotype of J2 Cohanim Zadokites

The haplotype 12-23-15-10-X-Y-11-15-12-13-11-29 which was incomplete in (Behar et al, 2004), is in fact the 10-marker modal J2 Cohanim haplotype for both Cohanim Askenazi and Cohanim Sephardi. When J2 Cohanim Askenazi present this 10-marker haplotype along with X=14 and Y=17, and J2 Cohanim Sephardi who also have this 10 markers along with X=13 and Y=18, it became the complete 12 marker for Cohanim Askenazi and Cohanim Sephardi ancestral haplotypes, respectively. It indicates that this haplotype can best represent the Cohanim and therefore it forms a unique special haplotype that unites both Askenazi and Sephardi in one single and very exclusive cluster that is directly related to the main line of Pinchas-Zadok. Therefore, it can be called the haplotype of Cohanim Zadokites.

Comment. We don't coin neologisms here. Use of the term "Cohanim Zadokites" for this context needs a reliable independent external source, or it has no place here. Jheald (talk) 13:40, 2 March 2009 (UTC)

Families with the extended 12 match Zadok Modal Haplotype

By selecting and analyzing the DNA of the J2 samples it was found that among thousands of surnames only a certain circle of families have a 12 match with the 'Zadok Modal Haplotype'. The families that have the full 12 match with the 'Zadok Modal Haplotype' are directly associated to traditional Cohanim surnames recognized in both Askenazi and Sephardi Jewish communities.

Hg	COALESCENCE TIME PINCHAS   -   ZADOK	COMMUNITY	DYS 393	DYS 390	DYS 19	DYS 391	DYS 385a	DYS 385b	DYS 426	DYS 388	DYS 439	DYS 389i	DYS 392	DYS 389ii	FAMILIES W/ 12 MATCH
J2	ZADOK MODAL HAPLOTYPE	ASKENAZI	12	23	15	10	14	17	11	15	11	13	11	29	Cohen, Katz, Kaplan, Rapaport, Kahan
J2	ZADOK MODAL HAPLOYTPE	SEPHARDI	12	23	15	10	13	18	11	15	11	13	11	29	Cohen-Rodrigues, Cohen-Pereira, Cohen-Machado

Comment. You are proposing to write all of this on the basis of three stray samples, and no strong evidence that they match anybody. Jheald (talk) 13:42, 2 March 2009 (UTC)

Disagree Again, the only three families that has the tradition and it is recognized among Sephardi community in Portugal and Spain are these three families mentioned. Not a case of coincidence, that these three families has a full 12 match. Please, send me any citation or any other kind of information that contains three recognized Cohanim surnames among Sephardi Community in Portugal and Spain that has a full 12 macth and also, with one-step mutation from Askenazi J2 Cohanim. What are you expecting? The Cohanim haplotype must not be in the mainstream of J2, after all Cohanim has a unique haplotype, completaly different from regular jews. See comments above. [User:Chriscohen|Chris Cohen]] (talk) 19:02, 3 March 2009 (UTC) —Preceding unsigned comment added by Chriscohen (talk • contribs)

____________

Anatole Klyosov: Chief Professional Scientist, Patents and publications (2003-2008). Professor of Biochemistry at Harvard University (Cambridge, MA), Harvard Medical School (Boston, MA),Institute of Biochemistry, Russian Academy of Sciences (Moscow), Consulting Chief Scientist and Chief Scientist for Pro-Pharmaceuticals (2000-2008); Professor and Head, Carbohydrate Research Laboratory, World Academy of Science (Member, since 1989),USSR National Prize in Science , Gold Medal, the USSR National Achievement. --Scienceprogenetic (talk) 03:41, 2 March 2009 (UTC)

____________

(Above is the correct information and the new main article.)

_____________

Before this becomes an Edit War...

I first came to this article to archive an excessively long talk page and to fix talk page vandalism (deletion of 50,000+ bytes of discussion, twice). 189.63.100.161 (talk · contribs) had deleted a large amount of what 189.63.76.165 (talk · contribs) had written, so I reverted it and dropped a vandalism warning on their talk page ([13]). Then checking the article itself, I noticed that 189.63.100.161 had also made massive changes there as well, so again I reverted and dropped another vandalism warning on the user talk page ([14]). 189.63.100.161 then leaves me a message ([15]) saying it wasn't vandalism and it had been discussed by "the group", which I understood to mean "a group of like minded Wikipedia editors", so I assumed good faith and let it be for a while. 189.63.100.161 has since signed up as Scienceprogenetic (talk · contribs) and has asked for my help ([16])which is why I am here now.
When I checked to see what discussions had taken place recently, there was very little actual discussion to be found. It turns out there had been discussion, but in a display of bad etiquette, it was deleted by 189.63.103.38 (talk · contribs) with this edit. Correct me if I'm wrong here, but I believe that 189.63.100.161, 189.63.76.165 and 189.63.103.38 are the same individual as Scienceprogenetic.
Anyway, I'm now not so sure I was right to let it be for a while. Although I am not an expert in the field, the version of the article from 21 Feb ([17]) seems well referenced with links to peer-reviewed specialist journals which do qualify as reliable souces and serve to verify what has been written. In contrast, the version promoted by Scienceprogenetic seems to be largely based on his new work which is as yet unpublished ([18]) and presumably still awaiting peer-review - something that clearly breaches Wikipedia's policy of "No Original Research".
What is the best way forward: IMHO, if Scienceprogenetic has new material that can be supported by one or more reliable sources, then that material should be added to the article, perhaps in a "latest developments" section. When Scienceprogenetic's paper is finally published, then the "latest developments" section could be expanded further. There is nothing wrong in having an article reflect a progression of understanding and showing the history of scientific progress.
This talk page should be used for civil discussion aimed at reaching a consensus. Judging by the ongoing debate here, there is still a long way to go. Meanwhile, Scienceprogenetic could spend some time learning how Wikipedia works as a collabrative effort - in particular I suggest he reads the etiquette guideline and stops deleting talk page discussions he doesn't agree with.
Astronaut (talk) 03:47, 3 March 2009 (UTC)

Dear Astronaut. I appreciate your help. Scieneprogenetic did not delete anything. I did it, but was a mistake. Anatole is not familiar with Wikepedia edits, so I tried to help. My bad. Now, If you read the article, you will not find any references for the terms used there as L24, or J21ak1, or anything like that. Jheald is adding things like that from his (or her) own project. This is a clear example of "No Original Research". Can you ask for references or not about L24 or J21ak1? Do us a big favor by asking them the references. Thank you. Anatole Assistant. —Preceding unsigned comment added by 201.47.180.125 (talk) 17:14, 3 March 2009 (UTC)

It is up to you experts (Scienceprogenetic, Chriscohen, and Jheald) to reach a consensus as to how the article should be. Like I suggested above, I think there is room in the article for both points of view, so long as both views can be supported by references from reliable sources. At he moment Jheald's view seems to be better supported with references from nearly 2 dozen peer-reviewed papers. If you feel a sentence is in need of additional verification, you can add "{{fact}}" to the end of that sentence.

One other thing: Please do not add thousands of bytes of proposed article contents to my talk page. It will be deleted without being read. The place to discuss changes to this article is here, on this talk page. Astronaut (talk) 04:09, 4 March 2009 (UTC)

Agree with Astronaut. As he well said: “There is nothing wrong in having an article reflect a progression of understanding and showing the history of scientific progress. If Scienceprogenetic has new material that can be supported by one or more reliable sources, then that material should be added to the article, perhaps in a "latest developments" section. When Scienceprogenetic's paper is finally published, then the "latest developments" section could be expanded further. I think there is room in the article for both points of view, so long as both views can be supported by references from reliable sources.” I pretend to help Scientifprogenetic add a new section that can be called “Latest Developments on J2 Cohanim” that will have the section now called “J2 Cohanim Haplotype Tree” (See above). This is a fact and Scienceprogenetic can display the table and graphic with the references, reliable sources, and everything else. If Jheald does not accept this new section, I must call for a dispute and invite more editors to help in this matter. I will try to remove much of the content on the article that has no endorsements at all from the sources and that Jheald is responsible for it. So, if this new section that Scientificprogenetic will improve with references and pretend to add later in the article is not approved and does not reach a consensus with Jheald, we will call for a dispute, and this can became an edit war. Than, I will make sure new editors understands that without endorsements from the reliable sources, much of the information that is currently in this article must be removed Immediately. I am here trying to reach a consensus with Jheald. Scientificprogenetic can add the new section and must add soon. Regards for all of you! --Chris Cohen (talk) 04:58, 4 March 2009 (UTC)

Original Research?

Having carefully read the article and some of the sources, I do have a problem with the More detailed Cohen haplotypes section. It actually states in that section: "... informally released by the private company Family Tree DNA (FTDNA), based on further work by much of the same research team. It has not yet been peer group reviewed by other scientists or published in the open technical literature" (my emphasis added). The references that are provided are either from Family Tree DNA directly, or from data provided by Family Tree DNA - ie. data based on unpublished original research. If this section is based on this original research, then it also calls into question some of the conclusions drawn in the previous Haplogroup placement section. Astronaut (talk) 07:57, 4 March 2009 (UTC)

FTDNA has been using a 12-marker string rather than a 6-marker string to identify "Cohens" since at least November 2004, as this page from that time at archive.org explains.[19] Essentially the same text can still be found on FTDNA's site today. [20]. It is the page at FTDNA that individuals who match the 12-marker string are directed to. (FTDNA is, probably by a factor of 10, the largest genealogical DNA testing comany in the world; and also retains Dr Michael Hammer who is one of the key academics in the area as a lead advisor. So what FTDNA indicates is worth the article noting). Another site quotes FTDNA material: "There is a Cohen Modal Haplotype that we automatically compare everyone against. It is a 12 marker set of results. If you match this haplotype we put a CMH badge on your personal page which is linked to information on what that means." [21]

The values for the 12-marker string were confirmed in 2005 by FTDNA's president Bennett Greenspan to the co-ordinator of FTDNA's Mexico project [22]. The values came as no surprise to those studying Haplogroup J DNA [23], who had already identified them from the public data [24].

That was four years ago. Since then far more individuals' DNA haplotypes have been tested, and posted publicly, in ever greater resolution. The clustering of many many Ashkenazi Cohens' haplotypes around these values is very evident, for example in the haplotypes reported by the Haplotype J project [25], and by the Cohen project [26].

So, that there is a very pronounced cluster around these values is not in doubt.

What Anatole is now suggesting is that there might be other significantly common Cohen haplotypes that were also lumped in together in the original reports based on the original six markers. But if they are out there, they are much much less prominent, and this would constitute new material from Anatole. Jheald (talk) 11:06, 4 March 2009 (UTC)

Agree with Astronaut. The whole ‘More detailed Cohen haplotypes’ section is completely based on inconclusive material which have no endorsements. The references are definitive provided either from Family Tree DNA directly, or from data provided by Family Tree DNA - ie. data based on unpublished original research. This whole section is based on this original research, and the conclusions were completely used in the previous section name ‘Haplogroup placement’. If you look carefully at this section, the readers will see: “J2a1k* (L24,L25)” , “J2a1k* (L24,L25)”, “J2b2e (M241)”, “J2a1k1 ?” . The references for that are not valid at all. Jheald added this material based on false reliable source. For Jheald, ISOGG has endorsed the very inconclusive discovery about this new thing, as he or she said: “The designation J2a1k is from the ISOGG official current tree, which Wikipedia considers a reliable external source” (See comment above). I visited their website. ISOGG has not endorsed such things. Read carefully on their main page. “The Working Draft of the Y Chromosome Tree Below J-M304 “does not represent an endorsement by ISOGG”. So, both sections “More detailed Cohen Haplotypes’ and ‘Haplogroup placement’ must be removed immediately from the main article and every point mentioned there must be transferred to this discussion page. In the meantime, Scientificprogenetic must add a new section that can be named “Latest Developments” and fill with one or two reliable source, and "{fact}" to the end of the sentence. “There is nothing wrong in having an article reflect a progression of understanding and showing the history of scientific progress” said Astronaut and I agree with him too. --Chris Cohen (talk) 14:02, 4 March 2009 (UTC)

Agree with Astronaut and Chris Cohen. We are finally reaching a consensus. Astronaut, Chris Cohen and me, definitely thinks that the main article is full with inconclusive material that has no endorsements. Two sections there must me removed and transferred to the discussion page. The references are definitive provided either from Family Tree DNA directly, or from data provided by Family Tree DNA - ie. data based on unpublished original research. I did answer Jheald new questions. Please read above the section The 'Cohen Modal Haplotype' and search for my answers and graphics. I can work with Chris Cohen and Jheald too in the new section that can be called “Latest Developments”. Until them, we may need to remove two sections in the main article, mentioned above by Astronaut and Chris Cohen. These two currently sections are full of these incorrect conclusions, thus need professional editing. I will work with Mr. Chris Cohen, in order to add this new section in the article. Best Regards.--Scienceprogenetic (talk) 21:35, 4 March 2009 (UTC)

Look, I am very open to ideas and critiques of what is there. But, for better or worse, Wikipedia aims to reflect the mainstream view and cover what the people considered mainstream are saying. FTDNA is considered mainstream, because it is the largest testing company in the world with over 90% of the market, and its internal genealogical database is the largest in the world, with over 500,000 samples. So we should report what it says. ISOGG is considered mainstream, because it has a very good reputation for being careful and up to date, has several times been cited in academic papers, and because keeping track of reported SNPs is more about careful bookkeeping than rocket science. We use the relevant part of its tree on every haplogroup page on this site. And the FTDNA J project is the most visible, partly because for those who elect to have their samples included in it and made public, the data is generated directly from the FTDNA central database. So an article should cover what these people are saying, even if it then goes on to critique or qualify it.

And I am very open to the article including critiques or qualifications or health warnings or alternatives if there are verifiable issues that people have raised or things that have been missed. But we have to understand that Anatole's unpublished (and I fear unpublishable) preprint simply does not have the same authority.

At the moment, the article is intentionally very conservative. It doesn't coin new terms like "Zadokite", or speculate about unverifiable 7th century figures. It largely sticks to the DNA data, and doesn't go much beyond saying "other clusters exist", and where to find them. In my view, that is an appropriate line.

There are things I would change. I think it is a mistake to identify "Y-Aaron" so closely with the J1 cluster; I would prefer a more even handed treatment, for example closer to the corresponding section in this edit from November 2007, not taking it upon ourselves to award "Y-Aaron" to either J1 or J2; and somebody has rather botched editing in the explanation of convergence. But overall I think the section is basically sound, does reflect the position of mainstream sources, and is a fair starting point for any further edits. Jheald (talk) 22:47, 4 March 2009 (UTC)

Disagree. Jheald said: "(01) FTDNA is considered mainstream so we should report what it says. (02) ISOGG is considered mainstream, so we use the relevant part of its tree (not endorsed by ISOGG in any circumstance) on every haplogroup page on this site. And the (03) FTDNA J project is the most visible, the data is generated directly from the FTDNA central database. So an article should cover what these people are saying, even if it then goes on to critique or qualify it".

Here Jheald clearly demonstrate that (01) FTDNA, (02) ISOGG tree (without official endorsements from ISOGG) and (03) J Haplogroup Project in FTDNA website, must necessarily be considerate as reliable sources, therefore “an article should cover what these people are saying, even if it then goes on to critique or qualify it.” I don’t agree with Jheald at all. This is not how science moves, but how a speculation spreads into the world. We do respect FTDNA a lot, but this private company have not endorsed any specific, independent and particular project nor published any official article. Therefore to use pseudo reliable sources not reviewed or endorsed by other scientists or published in the open technical literature is a big mistake that Wikipedia’s editors will notice and judge correctly. Now, Astronaut can resolve this pending situation. Whatever he says should be correct, since he knows about the Wikipedia´s policies more than me. Can we let Jheald insist with this original resource in the main article by using and making false interpretations from data based on unpublished original research? And finally, Jheald just made himself a new editing on the main article, without reaching a consensus. Is this okay? --Chris Cohen (talk) 23:48, 4 March 2009 (UTC)

Questioning Y-chromosomal Aaron sample groups

The Y-chromosomal Aaron genetic studies of 3 Jewish groups, Ashkenazi (Ashkenazi means 'German') Cohen, Levites and Israelites raises several questions regarding Sample groups used and methodology.

• Firstly, the entire study was based upon 988 individuals who were male and had identified themselves as belonging to one of the above groups. The samples group were also from the groups within the Jewish population who descended from the priests within the faith and would therefore be more inclined to exhibit the results found. This would indicate that the study could be seriously flawed as only males who were either known (for instance how were the subjects located? Through adverts in Jewish newspapers or through Synagogues or through word of mouth? How were the requests for samples worded?) to exhibit the desired chromosomal markers may have been used. For example only males exhibiting an Israeli or Arabic appearance within the 3 groups may have been used and therefore the Semitic (Arabic or Middle Eastern origin) content would be exaggerated and would invalidate the study. The study would have to use a control group of independently and randomly selected individuals who would then be DNA tested and then once the results were known use them to identify the group the person later revealed themselves to be from.
• Secondly, Jewish lineage is widely known to be passed down on the maternal side not the paternal side that was used in the study. This raises serious doubts as to what the study could show in historical lineage terms. It would stand to reason that if only males were tested then the sample would not represent the entire Jewish lineage. The results would then only be an acceptable indicator of 50% of the Jewish population as females were excluded. A secondary female Jewish test group would be necessary to control any variation. As there are non of the chromosomal markers used in the study in the Chromosomes of females this would be impossible but testing of other inherited female markers would have been vital in establishing whether the same ancestral bases existed or did not. For instance the entire population of the Europe can be traced back to 7 females who lived 28,000 years ago via markers in their Mitochondrial DNA, the so called 7 Daughters of Eve. See http://www.sciencedaily.com/releases/2008/07/080715204741.htm
• Thirdly, The sample group was so small that the results have to be seriously questioned as valid. Whilst it is accepted that testing 1000 individuals should yield a reliable result for the above concerns this would not be acceptable as manipulation could easily have occurred of the sample group. If a very large sample group of both male and females were taken then the chances would decrease of any manipulation to have taken place.
• Fourthly, the 'Khazar' (Khazars were a European/Asian Turkish tribe who converted to judaism in or around 980 AD and are thought to comprise 80 - 90% of the genetic stock of all people calling themselves Jews in the world today - also my own lineage) element of the Jewish population has not been examined in the study by genetic markers that would trace Khazar ancestry through a different chromosomal link. The reasoning ere is that a direct link with the Cohen, Levite and Israeli lineages but it does not mean that these populations did not then actively reproduce with Khazars thus including the Y chromosome Aaron markers in the resulting groups.
• Fifth, The sample groups are known to have the highest level of exclusive reproduction of all the Jewish grups and therefore this would naturally show the most 'pure' of genetic links back to the original Semite middle Eastern Jewish tribes and does not therefore prove that all Jewish populations carry the same link to the Aaron lineage. —Preceding unsigned comment added by Marcuslaing09 (talk • contribs) 21:33, 4 March 2009 (UTC)

Marcus, I think you're being paranoid. The best indication that Thomas and later Behar assembled a reasonably representative sample is how closely the Ashkenazi data matches the data that principally Americans have submitted to open databases.

(2) This article is specifically about Cohen lineage, not general Jewish genealogy. The result is that many Cohens appear to be particularly closely related, as might be expected from tradition.

(3) See (1).

(4) As I understand it, the haplogroup that is thought would be most strongly associated with the Khazars is Haplogroup R1a. According to the data at Y-DNA haplogroups by ethnic groups this occurs in about 13% of Ashkenazis, but only about 4% of Sephardim. Even if all of the Ashkenazi R1a was from the Khazars (which is very unlikely), this suggests that they represent much much less than 80-90% of Ashkenazi ancestry. There do seem to be quite a few R1a Levites, though.

(5) See (2). This is not an article about the genetic ancestry of "all Jewish populations". It is specifically an article about the genetic ancestry of the Cohanim.

-- Jheald (talk) 23:08, 4 March 2009 (UTC)

End of discussion with Jheald

Here is a link to Bruce Walsh's full paper, Estimating the Time to the Most Recent Common Ancestor for the Y chromosome or Mitochondrial DNA for a Pair of Individuals, Genetics 158: 897--912. As you can see, he sets out the mathematics quite generally, and it can be adapted to any number of markers.

Anatole, you may be a professional in chemical kinetics, but you are not a professional in bioinformatics, and it seems your training in statistics is shaky. The falseness of your analogy with the "last two molecules" should scream out at you. You can ignore the "last two molecules" in chemical kinetics because you are dealing with an enormous number of molecules all acting independently; besides, you are only expecting the reaction to run to equilibrium. But this system is different. The lines aren't all independent: they are strictly hierarchical. And you can't ignore the last two, because it is their coalescence we are specifically interested in, that all the lines go through. As Ken Nordtvedt pointed out, that means you cannot treat the 100 haplotypes as independent: they are correlated. You will get a fair estimate of the expectation value of the TMRCA distribution, but you will grossly underestimate its uncertainty. No matter how many haplotypes you test, at the very least you will still have an uncertainty a few times that of the uncertainty in a perfect n/n match, if you are testing n markers - because that is the minimum uncertainty in how long it will take the last two lines to merge. And if you know the last two lines have a much larger disparity than an n/n match, the uncertainty will be so much bigger.

The second point is that we are interested in the uncertainty of the distribution, not the uncertainty of its expectation value. You may have produced three (strongly correlated) estimates of the expectation value of the TMRCA, but that tells you nothing about the width of the distribution.

I don't mean to be dismissive. But in science bad claims about statistics stink. Deluding yourself into thinking you know something more accurately than you really do is a road to nowhere. Jheald (talk) 01:07, 5 March 2009 (UTC)

" We really don't know how many generations take place between those early occurrences of nodes in the tree for the sample population. There is intrinsic noise uncertainty in any ability to infer them no matter how large the sample population (which is constrained by total population). Those intervals which contribute to the total generations are shared in common by a very large fraction of the sample population haplotype pairs , so there is no way to drive down the statistical uncertainties in their contribution to total G. It requires INDEPENDENT measurements of something to invoke the 1 / square root of N diminishment of noise errors. " (Ken Nordtvedt) " Jheald

Disagree. Dear Sir or Madam, Since I do not know your name, though you know mine (Anatole Klyosov), we are not in exactly equal position. However, it does not matter. On your side I see a person, who has a rather limited knowledge in my field, that is kinetics of chemical and/or biological processes, and mutations exactly fit into this broad category. To call it “bioinformatics” is ludicrous; however, I do not care either. You can call it “astronomy”; it would not change a thing. However, it reflects your level in that field of science. You see, it is easy for me to describe your lack of knowledge, since you do not have a name. Now, I am still responding, only because I think I will publish this epistolary piece as rather educational for some audience. Now, the first question which I missed completely. The question is – "why we are doing those calculations?".

From your comment it follows that you do NOT do those kinds of calculations. However, you dare to “criticize” a person who does it by hundreds, and long ago developed an “educated intuition” what can be done and what cannot. I do those calculations for some 40 years. Having said it, I observe with bitterness how poorly people do calculations in DNA-genealogy. It is insane. And it seems that by doing this in an absolutely wrong way, their major concern is to fool themselves with “error margin calculations”. Let me explain this. In chemical kinetics, a long established field of (mutation) rates of reactions, it is very important to separate time courses of parallel reactions, or consecutive reactions, or parallel-consecutive ones. Otherwise one will obtain a mess, a bunch of (mutation) rates, some phantom parameters. That what exactly most of people do not do in mutation calculations. They grab some series of haplotypes, which represent mixtures of lineages, stemming from different “common ancestors”, and calculate something, which does not have any sense. And then they “carefully” calculate some pseudo-error margin”, being miles away with the principal result. That is what Hammer, Scorecki, Thomas et al did in their papers in Nature. In the “Origins of Lemba” paper they did not even bother to do any calculations at all, ridiculously talking on the “CMH” as “exclusively Judaic origin”. Wrong. The initial claim on the “Cohen modal haplotype” and calculations in the first Nature papers was wrong from the very beginning. It turned out that the authors took several different “CMH” lineages, recent and old, and calculated them as one pool. Wrong. The whole “CMH” premise as a “Cohen” one was wrong. This a typical Arabic lineage as well, 9,000 year-old, at least. Among the Jewish people it is some 4,000 years old, well before Cohens. Do you follow my point? DNA-genealogy is full of wrong things, wrong in principle. And a seemingly major concern is to wrap those wrong things into calculations of “error margin”! I do not deny an importance of calculations of error margins, but ONLY when the principal answer is basically correct. And even in this case one should understand what is the question addressed, and when error margins seems to be important, and when nor necessarily. I means, when sometimes a semi-qualitative result is very important. For example, there is a popular fable that current R1b (R1b1b2) folks in Europe have their common ancestor who lived some 30,000 years ago. One, on the contrary, using different populations – in the Isles, Iberia, Germany, Italy, Flemish haplotypes, etc. gets repeatedly 4200, 3600, 4400, 4500, 3800 years. Repeatedly and relatively reproducibly, using different populations. Do you understand that an error margin here would be useful, granted, but practically irrelevant to the principal question? Clearly it is not 30,000 years.

What it is important here, is to demonstrate, that one dissected those haplotypes into lineages, shown – on a haplotype tree – that it is a “homogeneous” population, having one common ancestor, that these statements are supported with quantitative criteria, that the mutation rates are calibrated, that 12-, 25-, 37, 67-haplotypes give principally the same, or similar result of calculations. That is what DNA-genealogy needs at the present time. However, “critics” do not care. They demand “error margin calculations”, despite the most ridiculous ways of principal calculations and wrong principal results. “Error margin” is what they do care about.

And you know why? Because it is easy. It is formal. Because they do not need to know more complicated things beyond their knowledge.

And, yes. One more thing. They do not want to acknowledge that those “error margin calculations” are themselves based on wrong or arbitrary conditions. There is a lot of assumptions in those calculations.

I prefer to use ACTUAL numbers, without those assumptions. I know that when haplotypes in the series are generated from ONE common ancestor for all haplotypes in the series, a two criteria should be consistent to each other: a number of base (non-mutated, presumably ancestral) haplotypes and a number of mutations. They should give the same or similar numbers of generations to a common ancestor. If both of them are the same, I got the right number. And it is ridiculous to tell me that my number has a huge “uncertainty” such as between 500 and 2500 years. People who say it do not understand principles of mutation rates and principles of working with right series of haplotypes.

And cone more thing. Let me teach you a principal thing in science. First, you line up what is known about your subject. Regarding Jewish haplotypes – write on a sheet of paper KNOWN facts and findings. Some of them might look as conflicting with each other, it is O.K. It is life.

Then, try to come up with an OPTIMISED version which would unite all those findings, or most of them, into a tentative (always tentative!) concept. You can call it a working hypothesis. Try to give an explanation why there are “facts”/findings which conflict with the others: either your concept is too narrow, or you believe that the “facts” are erroneous ones.

Having this concept, try to explain other, new findings. You might need to extend your concept to include new facts and findings.

You concept can be imperfect. In fact, it always is. However, it is your concept, your base, opened for corrections and further upgrading.

That is what I have done with the Origin with the Jews, based on their haplotypes.

If you want to change it, to modify, to upgrade, give me DATA, or a new INTERPRETATION. It is justified, I gladly accept it. It will make my concept better.

However, what are YOU trying to do? You do not give me new data. You do not give me new interpretation of know data. You just try to dismiss my concept. And not even concept! You just toothlessly trying to say that I do all wrong. What specifically?

• This is a Wikipedia article. It is not a place for presenting your new concepts. Jheald (talk) 19:55, 5 March 2009 (UTC)

Let’s take a look, specifically:

Anatole, we need to review what these papers said, and didn't say, so that people have the references as to where to find them, and know how to assess what they may find written about them elsewhere. Jheald

Fine, be my guest. Review. Make references. I have already done it, in my “Origin of the Jews”.

“As you see, there are clearly two different branches here. The “younger” have a DYS385a,b pair 13-15, the “older” 13-17. There are other rather significant differences in 37- and 67-marker formats. This pair was not determined and, hence, considered in the 2004-Behar [15]paper.” Jheald

Skorecki, and later Behar, deliberately recruited a sample of Cohens, as Cohens, by seeking out Cohens - presumably through synagogues, advertisements, etc. Jheald

Well, not exactly. Why there are so many “Israelites” on the list? If they will get only Cohens, all of them will be self-identified as Cohens. I agree, that the selecting was “doctored”, deliberately distorted, but this was not described in the paper. This is un-scientific and deceiving in that case.

• Skorecki et al delibertely recruited a Cohen sample, and a Yisrael sample to act as a control. This is abundantly clear in the paper. Jheald (talk) 19:58, 5 March 2009 (UTC)

However, if the number of “Cohens” was deliberately increased, and it was certainly increased, this does not change anything in what I have said before. We just take these haplotypes “as is”, analyze them, build a tree, dissect into lineages, and calculate timespans to common ancestors. Jheald

How is it different from what I have done? I did not consider any religious component in it. I treat the haplotypes of “J” haplogroup as regular people. And I obtained, that these 194 6-marker haplotypes split into at least five different lineages, one of them is the “CMH”. It was already clear from the tree and the haplotypes themselves, that the “CMH”, or at least a good part of them, represent a rather young lineage, 1350 years old. The 7th century. Can you tell me how can it be explained from the viewpoint of Nature papers of 1997 and 1998? Do you want to challenge this finding? Please, do. Now you see what I am talking about principles of scientific discussion? I give you fact and findings. What do you give me? Just some doubts? Based on what? Can you spell them out?

• I will defer full comment on your 1350 number until I have looked at your paper and seen exactly what data you are using. However, some initial comments.

(1) We only have high-resolution data for Ashkenazi Cohens. We don't know what high-resolution halotypes look like for the Sephardi and Mizrahi Cohens in J1 matching the CMH. Therefore, this is likely to represent a TMRCA for Ashkenazim only.

(2) The TMRCA estimates the first branch point with extant descendents. But the original founder could have lived rather earlier than that.

(3) The TMRCA only reflects the common ancestor of people in the sample. This is not an unbiased estimate of the TMRCA of the whole Cohen population, because less populous descendent groups may survive from earlier branchings may survive that weren't picked up in the sample.

(4) Even if less populous groups were picked up at low frequency, they would be masked by your statistical method, which doesn't treat non-homogeneous groups well and is dominated by the TMRCA of the most doominant subset.

(5) Finally, you have no serious estimate of the large uncertainties, particularly the heavy tailed distribution towards earlier times.

For all these reasons I regard your 1350 number as preliminary and not conclusive.

I will have more comments once I have seen your data and had a chance to analyse it. Jheald (talk) 21:42, 5 March 2009 (UTC)

It is possible that the selection wasn't completely reliable, but it is clear that it must have been pretty reliable just from how distinct the Cohen population was from the Yisrael in the studies. Jheald

It is a confusion. It does not follow from the paper at all. It is HOW it was presented to the general public.

What the paper said? First, there is a high proportion of the “CMH” in Jewish population and among presumed “Cohen” first of all. However, the “CMH” is not just 14-16-23-10-11-12. It is all mutations around. Mutations can happen to anyone at any generation. What, they ceased to be Cohens? Second, it essentially shown but not said, that those “CMH” among the Jews is a very recent lineage. It does not fit Aaron times at all. Otherwise it would not be so many identical, non-mutated haplotypes. There are 91 base, “CMH” haplotypes among those 194. Even if ALL of those 194 haplotypes are slightly mutated CMH, even then a common ancestor of those CMH would have lived some 2,000 years ago, beginning of the AD. Still too far from Aaron. It turned out, the common ancestor of a majority of them lived 1350 years ago. The 7th century. I know only one truly prominent person, who lived in the 7th century, an who had a very influential and powerful offspring, his sons who were dukes and knights of France, Spain, and some other countries. It was Bustenai haDavid, the head of the Diaspora, direct descendant of King David. Hence, my hypothesis. I know story of his life, and that is why I came up with this hypothesis. I will stand for it unless someone will justify that it was impossible. But why impossible? He had sons and their descendants. Well, enough about it. I know what I am talking about.

I challenge anyone to disprove it. Remember the principles of a scientific discussion?

And yes, not all CMH people are Cohens. And not all religious Cohanim are CMH (nor all of them from the same family). Jheald

That´s what I have said. Again, what is new that your are talking about? Where is substance?

As to your paper, I haven't seen your data so I can't assess your conclusions. I am surprised if you have found such a substantial cluster of Cohens other than those that match CMH-12 haplotype, when there is so little trace of any such in the databases above; but why not publish your data on a website, and then we can all look more closely at it, and see whether you are right. Jheald

The paper is available. (lulu.com/content/2677603)

Sorry, it costs as much as $5.00, and I do not have rights to send it around myself. The edition belongs to the Academy, as well as all proceedings from it. I do not sell it, and I did not earn a penny from it. However, we are walking in circles. I provided a haplotype tree of 37-marker “CMH” haplotypes, which obviously splits into two principal branches. The same happens with 37-, 25- and 12-marker haplotype trees. You know what it is of interest. I give you data (!), the picture (!), and you respond “I am surprised if you have found…” Hello!! Can’t you see?

“The existing article itself discusses how having the CMH for a typical Jew only raises their probability of Cohen ancestry to a little over 20%, i.e. still leaving an almost 80% chance that they are not a Cohen. And the article makes clear that the CMH is widespread elsewhere in peoples where there is no suggestion of Cohen ancestry." The calculation that gives the 20% probability is clearly set out in the article. How do you get your 11%? Jheald

Because we calculated thing differently. The amount of Cohen according to the Behar’s paper supposed to be 4% among the Jews. The article itself gave 36%, 9 times higher. That is, actual amount of Cohens is 9 times smaller, hence, 11%.

• Your calculation makes no sense. What is this 36%? And why should 0.04 * (1 / 0.36) be a sensible indicator of anything at all? Jheald (talk) 20:09, 5 March 2009 (UTC)

However, what is a point to discuss, 20% or 11%, particularly when we refer them to different things? Clearly that the amount of “Cohens” in the paper grossly exaggerated. Clearly that the “CMH” is not the exclusive Cohen feature, and vise versa, clearly that the CMH is presented among Arabs and others. So what are we talking about?

It is, as you have noted above, necessary to distinguish Cohanim from others who also happen to have a 6-marker haplotype close to J1. The claim is that a very large number of Cohanim, close to FTDNA's CMH-12 are closely related and share a comparatively recent common ancestor. The FTDNA CMH-12 cluster is the one that is most apparent in the public databases. If there is another cluster nearby, specifically of Cohens, please present the data for it. However, even if such a cluster does exist, there would still be the possibility that both ancestors were original priests; or alternatively that this was an early branching that just happened to experience a run of early mutations. But the bottom line is that for your claim to be verifiable, you data needs to be verifiable. Jheadl

Again, we are moving in circles. Give me your clear version. I prefer to discuss data and concepts, not fuzzy words. I refer you again to my description of a scientific discussion above.

The claim is that a very large number of Cohanim, close to FTDNA's CMH-12 are closely related and share a comparatively recent common ancestor. Jheald

I do not know who made this claim. I do not know who has haplotypes of “a very large number of Cohanim”. How many? Where did you get them? What was a principal criteria? The last name?? A self-identification? The “CMH” haplotype? And you are making comments on an “error margin”? I have data on 6 (six) J1-Cohanim, and 12 (twelve) J2-Cohanim. Most of them have variable but close to each other (in their haplogroups) haplotypes. Still, I can make some preliminary conclusions. If you want to challenge them – go ahead, you are welcome. But, again, please, do not give me just fuzzy words. I hope, by now you understand that we stay at different levels. I have data, though sometimes limited, you do not have them, or do not show. Can you see the difference?

The FTDNA CMH-12 cluster is the one that is most apparent in the public databases. If there is another cluster nearby, specifically of Cohens, please present the data for it. Jheald

There are two. I told you about them, I list them, I gave you a haplotype tree. And here we go again. Are you really serious? Can’t you see the data? Can you give me yours?

However, even if such a cluster does exist, there would still be the possibility that both ancestors were original priests; Jheald

Sure it is possible. Everything is possible. However, it is funny that you desperately cling onto the original and wrong premise that the “CMH” is the priest signature. Arabs with the “CMH” are also originated from a Jewish priest? Including those some 9,000 years ago? By the way, my data shows that the both lines (you call them clusters) originated from their common ancestor 4000±200 years ago. Again, ±200 – because different calculations with different series come to the range of 3800-4200 years back. This common ancestor is called Abraham in the Bible. I do not know what was his actual name, but he clearly existed as a person which split the Jewish and the Arabic lines. It is not a religious, but a scientific statement.

or alternatively that this was an early branching that just happened to experience a run of early mutations. But the bottom line is that for your claim to be verifiable, you data needs to be verifiable. Jheald

Please come back and re-read my principles of scientific discussion.

Here is a link to Bruce Walsh's full paper, Estimating the Time to the Most Recent Common Ancestor for the Y chromosome or Mitochondrial DNA for a Pair of Individuals, Genetics 158: 897--912. As you can see, he sets out the mathematics quite generally, and it can be adapted to any number of markers. Jheald

I have already explained. Feel free to calculated a probability that 12-, 25-, 37- and 67-markers haplotypes will give just about the same figure, as well as a number of base haplotypes in the 12-marker format. Then we will talk. I would appreciate it. Jheald

Let me give you an example. I have published, in the pending paper http://precedings.nature.com/documents/2733/version/1/html data on a common ancestor of 15 (fifteen) Basque 25-marker haplotypes (R1b1b2). It gave me 3600 years to a common ancestor. This figure was calculated by five different approaches using the ASD in three variants (using deduced base haplotype, using formally calculated base haplotype, using the permutation method with no base haplotype), also by “linear” (not square) mutations, and estimated logarithmic method. All of them gave 3600 years plus-minus very little. 15 haplotypes, mind you. Then I received the Adams paper on Iberian haplotypes, which contained 750 19-marker haplotypes. I calculated a timespan to a common ancestor using all 750 haplotypes, and obtain 3625 years to a common ancestor. Well, maybe now you understand why I do not care about the Adams paper which considers PAIR of individuals? Did you read the paper yourself? Did you read the TITLE of the paper which said “pair of individuals”? Do you understand a difference between a pair and 750 haplotypes? Or a pair and even 15 haplotypes on that matter? Do you know that the whole paper is largely considers just 1 (one) mutation in a pair of haplotypes, which gives the highest uncertainty indeed? Did you read that when 20 markers and higher are considered, “is sufficient to obtain reasonably precise estimates for two individuals separated by 200 or less generations”? Have you noticed that even for two (!) individuals with 20 markers (40 markers together) when half of them are the same, their common ancestor lived 192±59 generations ago, while with 50 markers (100 markers total) and 30 markers are identical he lived 134±30 generations ago? And for 200 markers (in two haplotypes together) with 60 of them identical a common ancestor lived 158±31 generations ago? 158gen ±20% error margin. In my case 15 25-marker haplotypes provide 375 markers, and 750 19-marker haplotypes provide 14,250 markers. And you cite me the Wash paper for two individuals? You, dismissers, pick the worst case scenario from his paper and cite it around insisting that it invalidates all calculations in DNA-genealogy. Shame.

• No. Fifteen 25-marker haplotypes only provide 25-marker precision about the early stages of the tree; which is why the uncertainties reflect the uncertainties in TMRCAs for 25-marker close matches. Seven hundred and fifty 19-marker haplotypes still only provide 19-marker precision about the early stages of their tree.

See my comments immediately below. Jheald (talk) 21:52, 5 March 2009 (UTC)

Anatole, you may be a professional in chemical kinetics, but you are not a professional in bioinformatics Jheald

I have already explained. Read above.

And you can't ignore the last two (individuals), because it is their coalescence we are specifically interested in, that all the lines go through. Jheald

Wrong. We do not make those calculations “to the last two”. When I consider that from 100 haplotypes 35 are still intact (base), I do not care about “the last two”. You use the wrong concept.

In order to see that the swimming pool is half-empty (or half-full), you do not have to wait until it is empty, and they talk about waiting until it is completely dry. That is your distorted logic. What we do consider is for how many generations the amount of mutations would drop, say, two times. If the system has only one common ancestor, this is already enough to get an answer.

As Ken Nordtvedt pointed out, that means you cannot treat the 100 haplotypes as independent: they are correlated. You will get a fair estimate of the expectation value of the TMRCA distribution, but you will grossly underestimate its uncertainty. Jheald

So what? If you have a chemical compound dissolved in water, all molecules of this compound are not only “correlated”, but identical. Nevertheless, we consider their individual “mutations” (aka breaking of their bonds) as independent events. The same with haplotypes. Each mutation is an independent event, there is no “correlation” between them. As a result, we have a “mutation accumulation” curve, and can treat it as a sequence of independent events. That is what mutation rates are about.

• No. It depends whether the mutations (or non-mutations) are on different lines. If we're talking about segments of the ancestry which are distinct for the two haplotypes, then the mutations (or non-mutations) on them are independent. But for segments of their ancestry which are shared by two haplotypes, so are the mutations (or non-mutations).

This is where things are different from chemical kinetics, where you are interested in the average behaviour of many many molecules. With the haplotypes there is only one tree, and you in effect have only one independent data point for what's going on at the top of it.

That is why, as well as the uncertainty in your estimation of the expectation value of the TMRCA (which is what you are calculating), you also have to consider the uncertainty because of what Ken calls the intrinsic amount of noise at the top of the tree -- something you don't have to deal with in chemical kinetics calculations. Jheald (talk) 21:15, 5 March 2009 (UTC)

You folks are confusing yourself. Starting with “bioinformatics”.

About “uncertainty” – read above. There is some, of course, but it should be attributed not there where you think, but at wrong calculations of haplotype series.

No matter how many haplotypes you test, at the very least you will still have an uncertainty a few times that of the uncertainty in a perfect n/n match, if you are testing n markers - because that is the minimum uncertainty in how long it will take the last two lines to merge. And if you know the last two lines have a much larger disparity than an n/n match, the uncertainty will be so much bigger. Jheald

Read above.

… we are interested in the uncertainty of the distribution, not the uncertainty of its expectation value. You may have produced three (strongly correlated) estimates of the expectation value of the TMRCA, but that tells you nothing about the width of the distribution. Jheald

Read above. Apparently, we talk pass each other. As I told you, repeatedly, when I calculate, say, R1a1 haplotypes all over Europe, I get figures between 4200 and 4800 years from a common ancestor, - from Greece and Italy to the Isles, Scandinavia, Poland, Russia. Mind you, I consider different populations, not “strongly correlated estimates”, as you imply. I am talking about the width of the distribution. I perfectly understand that those numbers might be the same, within the same error margin, that is some 4500±300 years. However, it is still some ±7% error margin. A big deal? Also, I never say that 4500 is “younger” than 4800 years in that regard. I treat them as approximately equal. So, what is the problem? Bit clearly, they are not 15,000 years, as a popular fable about a “Ukrainian R1a1” says. THAT is the problem. Not an error margin, which is much less certain than the principal number itself.

"I don't mean to be dismissive." Jheald

- Maybe you don’t. But you are. This is VERY unproductive. Deluding yourself into thinking you know something more accurately than you really do is a road to nowhere.

You are confusing me with somebody else. At this level of our knowledge we discover new things. And I know more than many people in the area (and less than some others in their area of expertise) not because I pretend that I have some more accurate data then others. This is a level of a technician. I know more with principal things. You apparently do not understand. O.K. When I took the recent data by Adams et all about Iberia, and build haplotype trees of all haplogroups in Iberia, even the patterns themselves tell a remarkable story. Without any calculations. Calculations is a bonus. You can tell right away from those haplotype trees which populations are VERY young, just a few hundreds years ago, and which are ancient, some 12,000 years ago. You cannot possibly mix them. They are obvious without any calculations. And when I have done calculations, and, surely enough, I obtained some 200-300 year old common ancestors for some populations, and 11,500 years for others, it all fit with the patterns on the trees.

And someone like you come up and said – no, no, no, it is all very uncertain, it is a road to nowhere, it is all about distribution, - I do not know, what to do, to laugh or just to turn around and walk away. However, unfortunately, too many people repeat like parrots the same thing, not being able to do that kind of research.

This is pity. --Scienceprogenetic (talk) 17:31, 5 March 2009 (UTC)

Who are the editors?

The brief answer is, "everyone". Therefore edit summaries like "Done. Now it is up to Wikipedia`s editors to approve the reliable sources. (Not Jheald)" are incorrect. There is no board of editors, no official peer review or approvals process, and no one is really in charge (though about a thousand users, "admins", are given enhanced privileges that let them enforce Wikipedia's policies - including blocking or banning users). Wikipedia is a volunteer project with over 6 million contributors, some of whom are just 10-year old kids, some are retired bus drivers, soldiers, housewives, scientists, and so on. Few will have any expertise in genetics, but all (except blocked/banned users) will be able to edit the article text as they see fit. To specifically try to exclude another user from editing this article is not acceptable. Jheald is certainly much more experienced in this subject matter than (at a guess) most other editors, and you should welcome his contribution to the article and the ongoing debate on this talk page. Astronaut (talk) 03:28, 7 March 2009 (UTC)

Astronaut is very correct, but I am sure no one really mean to exclude Jheald out of the context, but inviting others editors to read the dicussion and to help make the correct decisions, to improve the quality of the article. As a matter of fact, Jheald is very welcome to participate in every aspect of the main article. I made few contacts with him, on his talk page, and Jheald did treat me very well, answering my questions and pointing out his comments. I gave him my word that we will work together always, including the other sections as weel. Jheald can count with me anytime, and I believe that with this "team", the article will be much more improved and gain quality every month. --Chris Cohen (talk) 13:42, 7 March 2009 (UTC)

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5