CMH Question: I was but am I still?

[HapJ1]

I'm J1 and 6 marker Cohen Modal Haplotype but not 12 marker Cohen Modal Haplotype. Can I still describe myself as CMH, even if FTDNA has now delineated 12 markers? Is the 12 marker CMH widely accepted? If not, why not?

Were others "disqualified" as CMH by the expansion to 12 markers? Am I correct to assume that all bona fide Kohanim match the 12 markers?

[Rick]

I'm J1 and 6 marker Cohen Modal Haplotype but not 12 marker Cohen Modal Haplotype. Can I still describe myself as CMH, even if FTDNA has now delineated 12 markers? Is the 12 marker CMH widely accepted? If not, why not?

Were others "disqualified" as CMH by the expansion to 12 markers? Am I correct to assume that all bona fide Kohanim match the 12 markers?

Over 3000 years I suppose we might expect a mutation or two from a 12-marker founding haplotype, so I'm certain the answer to your last question is "no". But, the big question about CMH, as I recall from previous discussions on the topic, is whether it is J1 or J2. With the short 6-marker haplotype, there were matches from both J1 and J2 individuals. Since these clades separated (supposedly) thousands of years before Aaron, only one of them should contain the "authentic" CMH. Has this question been resolved yet?

[vineviz]

Were others "disqualified" as CMH by the expansion to 12 markers?

Yes, though I don't think "disqualified" is the best term to be using.

Am I correct to assume that all bona fide Kohanim match the 12 markers?

No, it would be incorrect to assume that.

The CMH (or any modal, for that matter) is merely a form of statistical average for a population. In this case, the population is Kohanim.

Some Kohanim will not match the modal, but you can infer that the further away from the modal you are the less likely you are to be Kohanim.

Here's an analogy. The average height of professional basketball players is 6'7". If some one is 6'5" are they automatically disqualified from being in the NBA? No.

The Kohanim question is a little different, of course, because you either are or are not already Kohanim. You are merely trying to use your haplotype to figure out which is more likely to be true.

[josh w.]

Alan, there was extensive discussion of this issue on Roots Web. Most seemed to agree with FTDNA that the CMH was J1. However, it was also noted that FTDNA has never presented the empirical basis for its 12 marker set in any public forum. A related issue is that no one has any data on the dna of ancient Cohens. On another thead I also noted that any empirically based rule for inclusion or exclusion will have some false positives and false negatives (Since FTDNA has not published their data the percentage of false conclusions is anyones guess). The most serious error would be a false negative, i.e. informing someone that they are not from the Cohen line when they really are. Perhaps in consideration of these issues, FTDNA has a broad definition of the CMH allowing up to three mismatches.

[HapJ1]

CMH and other haplotypes

I appreciate all your replies.

Another question: I see that the Atlantic Modal Haplotype is 6 markers. The Dinaric and Nordic 5 markers. Why does it "require" 12 markers to ascertain who is and who is not CMH?

[vineviz]

Another question: I see that the Atlantic Modal Haplotype is 6 markers. The Dinaric and Nordic 5 markers. Why does it "require" 12 markers to ascertain who is and who is not CMH?

A modal haplotype is simpy an artifice. If you can define the population (e.g. these people are all Kohanim) then the modal for the population is only limited by the number of markers for which the population has been tested.

You could make a modal with 3, 9, 11, or 63 markers. It really makes no difference.

The choice to show a 12 marker modal only means that FTDNA feels they have enough data to do so. Nothing more.

[josh w.]

But based on the data for the 6 marker CMH, my guess is that less than 50% of Cohens will meet a 12 marker test (depending on the number of possible mismatches). That's why published data is so important.

[vineviz]

But based on the data for the 6 marker CMH, my guess is that less than 50% of Cohens will meet a 12 marker test (depending on the number of possible mismatches). That's why published data is so important.

Published data is important.

However, don't focus too much on the whole "perfect" match thing. An 11/12 match is more-or-less equally informative with a 6/6 match.

[josh w.]

I have no problem with extending the definition of the haplotype from 6 to 12 markers, provided the data support the extension. The problem is the one raised by HapJ1, namely that people who are truely Cohens will be informed that they are not on the Cohen genetic line, what I have described as "false negatives". Expanding the number of markers is similar to raising the cutoff point or threshold for a diagnostic test, e.g. the PSA test for prostate cancer. Such a strategy is designed to reduce the number of "false positives", in this case informing people that they are on the Cohen line when in fact they are not. However such a strategy usually increases the number of false negatives, the more serious of the two errors in this situation. Using the REO standard of a minimum of a 10/12 match, I wouldn't be surprised if less than one third of Cohens are informed that they have the CMH, i.e. two thirds of Cohens will not be designated as CMH. I could be wrong but only published data will show that. I haven't gone into the assumption that present day Cohens have the same haplotype as ancient Cohens, an assumption that was inconsistent with the data in the case of Ashkenazi Levites.

[HapJ1]

This just in from Dr. Skorecki

Dr. Skorecki informs me that he is working on a research project which he hopes will determine if the CMH is to be found within J1 or J2 and which determines "higher levels of resolution."

Stay tuned.

[vineviz]

The problem is the one raised by HapJ1, namely that people who are truely Cohens will be informed that they are not on the Cohen genetic line, what I have described as "false negatives". Expanding the number of markers is similar to raising the cutoff point or threshold for a diagnostic test, e.g. the PSA test for prostate cancer. Such a strategy is designed to reduce the number of "false positives", in this case informing people that they are on the Cohen line when in fact they are not.

I think it is important that people not misunderstand what a modal represents.

Being an exact match to the Cohen Modal Haplotype (CMH) is in no way to be considered a "test" for being in the Cohen line. Having the CMH does NOT tell you that you are Cohen, nor is it designed to.

Very few actual Cohen will match the CMH exactly, regardless of how many markers are used.

At six markers, I'd guess that less than a third of people who are Cohen descendants will exactly match the CMH. And given haplotype convergence, probably only 50% of those who match the CMH exactly at six markers are actually Cohen.

The CMH is an interesting statistical constructs, and some really smart people are working on better ways to deal with some of these issues, but nothing with those error rates should be considered a "test" at all.

I not defending the 12 marker modal, only trying to point out that anyone who judges themselves to be Cohen or not-Cohen based merely on a modal of any length is more likely to be wrong than right.

[HapJ1]

If I may interject...

If I may interject, I am not a Cohen nor am I assuming that having the original six markers delineated as CMH made me a Cohen. I am not a Jew and there is no tradition in my family going back 250 years that we are Jews or kohanim.

Almut Nebel uses the six marker CMH as the core haplotype of J and builds his network of haplotypes around it. I wonder if he can still do that with FTDNA's 12 markers.

[J Man]

If I may interject, I am not a Cohen nor am I assuming that having the original six markers delineated as CMH made me a Cohen. I am not a Jew and there is no tradition in my family going back 250 years that we are Jews or kohanim.

Almut Nebel uses the six marker CMH as the core haplotype of J and builds his network of haplotypes around it. I wonder if he can still do that with FTDNA's 12 markers.

I am one marker off of the original 6 marker CMH but I am way off of the 12 marker CMH. I am a J2 and my family also has no history of being Jewish.

A lot of things may change now with the 12 marker CMH.



Y-DNA: J2a*

[josh w.]

I have no problem with developing a modal haplotype, but why call it the Cohen Modal haplotype if it is not designed to see if one falls on the Cohen genetic line. I understand that FTDNA went to 12 markers in order to reduce the number of J2s who would qualify as CMHs. This is what I meant by reducing the number of false positives. The issue is that by resolving one problem they probably exacerbated another problem namely that of false negatives. That is, getting higher resolution in one direction may well produce lower resolution in the other direction. I agree that very few will base their religious identity on genetic data, although this kind of information might be helpful for those who suspect that their family was once Jewish or who suspect Cohen ancestry. Nevertheless it might be disconcerting for a family with a Cohen genealogical history to be informed that they are no longer CMHs in the shift to a 12 marker definition.

(I hope my numerical example was clear. By going from 6 to 12 makers the number of Cohen families with CMHs will be reduced. This is what happens when one raises thresholds. Current data suggests that around 50% of Cohens share the 6 marker CMH. Generally when there is an increase in the number of markers in a haplotype the number of matches will be reduced. This is what most people discover when they upgrade from 12 to 37 markers. This is how I came up with the guess of 33% Cohens having the CMH on a 12 marker test).

[vineviz]

I have no problem with developing a modal haplotype, but why call it the Cohen Modal haplotype if it is not designed to see if one falls on the Cohen genetic line.

A modal is simply a way to describe one characteristic of a population, in the same way that a series of numbers can be described by the mean and variance of that series.

Since the modal is a characteristic of the population, not any individual in that population, it is not designed to "see if one falls on the Cohen genetic line". That's just not what a modal does.