DNA-R1B1C7-L Archives

Archiver > DNA-R1B1C7 > 2011-08 > 1314493359


From: "Robert Reid" <>
Subject: Re: [R-M222] DNA-R1B1C7 Digest, Vol 5, Issue 284
Date: Sat, 27 Aug 2011 21:02:42 -0400
References: <mailman.21.1314428405.28642.dna-r1b1c7@rootsweb.com>
In-Reply-To: <mailman.21.1314428405.28642.dna-r1b1c7@rootsweb.com>


Clan MacTavish Seannachie websie:

The Modern MacTavishes claim descent from the line of Kings of the Kingdom
of IrGuill and Ros Guill, via descent from Conn of the Hundred Battles,
being a tribe of Picts (Cruithne) mentioned in the second century by Ptolemy
in his map materials. They were the Windukatii (Ouenniknioi) Picts, who
later fell under the dynastic Ui Niells of the North.

-----Original Message-----
From:
[mailto:] On Behalf Of

Sent: Saturday, August 27, 2011 3:00 AM
To:
Subject: DNA-R1B1C7 Digest, Vol 5, Issue 284



Today's Topics:

1. Re: The origin of R-M222 and the peopling of Ireland - ohboy
here we go again! (Sandy Paterson)
2. Re: The origin of R-M222 and the peopling of Ireland - ohboy
here we go again! (Bill Howard)


----------------------------------------------------------------------

Message: 1
Date: Fri, 26 Aug 2011 08:39:51 +0100
From: "Sandy Paterson" <>
Subject: Re: [R-M222] The origin of R-M222 and the peopling of Ireland
- ohboy here we go again!
To: <>
Message-ID: <000301cc63c3$572225d0$05667170$@com>
Content-Type: text/plain;charset="us-ascii"

Right, it's coming back to me.

Sometime during the past 6 months there was a shortish discussion in
between Dave Johnston and Ken Nordtvedt.

DJ wrote that he could show mathematically that the self-variance method of
estimating group TMRCA in fact produced an answer that was an unbiased
estimate NOT of the group TMRCA, but of the mean of the pair-wise TMRCA's
within the group. Since that is necessarily lower than the group TMRCA, it
means that the self-variance method it is biased low.

He went on to reason that in order to get an unbiased estimate of group
TMRCA, variance should be calculated with reference to the putative marker
value of the eponymous group "founder" rather than the mean marker value of
the group.

I don't wish to incorrectly attribute views to DJ or KN, but I got the
impression that KN agreed with DJ, but added the caveat that the marker
value of the founder is unknown. Yes, we can calculate modals, but although
we can regard the modals as the 'best estimates' of the founder markers, we
don't actually know them. That was my interpretation of the discussion.

Unfortunately, my computer crashed soon after that, so I'd be hard-pressed
to track down the exchange.

As I recall, DJ suggested that simply estimating the pairwise TMRCA of the
two seemingly most distantly related haplotypes within a group (read Conroy
and Doherty in M222), would probably be insightful, but the answer would be
biased high.

I think the discussion between DJ and KN was probably the most useful
exchange I've ever read on the subject on TMRCA estimation, but it seems to
have died off without much follow-up. Perhaps they continued the discussion
off-list.

So.

I think the above supports your contention that we should be concentrating
on the extremes rather than the group in its entirety.

Sandy











-----Original Message-----
From:
[mailto:] On Behalf Of Paul Conroy
Sent: 25 August 2011 19:33
To:
Subject: Re: [R-M222] The origin of R-M222 and the peopling of Ireland - oh
boy here we go again!

Sandy,

I just visually scanned the chart I linked to.

What I'd like to see is analysis done of a 67-Marker, M222 SNP Tested
sample, that excludes the well known Ui Neill families - that way we could
see what relationships can be determined from the outliers, which might
provide more clues?!

On Thu, Aug 25, 2011 at 1:57 PM, Sandy Paterson <
> wrote:

> Hi Paul
>
> I'm interested to know why you include Gillespie in the group of outliers.
> My interest in them is that I share DYS481,487 = 26,14 with them, and
> also that Gillespie and Ewing share DYS442 = 11.
>
> I agree with your general approach though - a better understanding of
> the outliers is likely to provide a better understanding of M222.
>
> Sandy
>
>
> -----Original Message-----
> From:
> [mailto:] On Behalf Of Paul Conroy
> Sent: 25 August 2011 17:11
> To: ;
> Cc:
> Subject: Re: [R-M222] The origin of R-M222 and the peopling of Ireland
> -
oh
> boy here we go again!
>
> Susan,
>
> Yes, it has been discussed many times.
>
> One of the topics most near and dear to many on this forum is whether
> M222 is Irish or Scottish, and especially whether particular families
> are of Irish or Scottish descent. The truth is that Western Scotland
> and Northern Ireland were almost the same country throughout history,
> separated by a water super-highway, that far from inhibiting travel,
> aided it. Almost every time a Scottish family can be shown to have
> Irish roots, those Irish roots may have been Scottish still earlier,
> and vice-versa - so to me it's pointless to try and distinguish, as
> they were one breeding population for millennia.
>
> Also, as I don't seem to be related closely to Northern Ireland or
> Western Scotland - though I do seem to possibly have share some
> Off-Modals to the Nith Valley Cluster - what I'm more interested in is
> the deep structure of the M222 group. I think I was the first person
> about 3 or 4 years ago, to point out that the Ui Neill families were
> extraordinary fecund, especially as their descendants were Medieval
> rulers in Ireland and Scotland, and
very
> early colonists in the US, they have created a very lopsided
> distribution of
> M222+. But to determine the true source of the entire group, we need
> M222+to be
> able to account for the placing of all the outliers, such as:
> 1. Conroy (myself)
> 2. Galyean
> 3. Gillespie
> 4. McCord
> 5. Cruden
>
> Check out this chart:
> http://dna.cfsna.net/R-M222.jpg
>
> And all the unrepresented East Ireland, West Ireland, South Ireland,
> North East Scotland, Central England, South England, Northern France,
> South West France, North West France (Brittany), Belgium, Holland,
> Germany, Norway, Denmark, Sweden - which to date are barely included in
this project.
>
> As I fairly closely some Daltons and some Stewards - both Norman
> French families - it's my opinion that M222 came in more than one wave
> from the Continent to the Isles, and probably has a source somewhere
> between
coastal
> between Brittany and Holland. Coming over first to South England and
> South East Ireland and possibly West Ireland as the Belgae tribes, and
> to South Ireland as the Erainn tribes, then later again as Norman
> French, to Scotland, England and Ireland.
>
> Cheers,
> Paul
>
> On Thu, Aug 25, 2011 at 11:34 AM, Susan Hedeen <
> > wrote:
>
> > The debates we've been reviewing regarding M222 have been around for
> > awhile. See this discussion. I'm sure you have probably seen it
> > before, but almost dejavu (spelling an example of phonetics)
> >
> > http://www.worldfamilies.net/forum/index.php?topic=8846.10;wap2
> > R1b1c7 Research and Links:
> >
> > http://clanmaclochlainn.com/R1b1c7/
> > -------------------------------
> > To unsubscribe from the list, please send an email to
> > with the word 'unsubscribe' without
> > the quotes in the subject and the body of the message
> >
> R1b1c7 Research and Links:
>
> http://clanmaclochlainn.com/R1b1c7/
> -------------------------------
> To unsubscribe from the list, please send an email to
> with the word 'unsubscribe' without
> the quotes in the subject and the body of the message
>
> R1b1c7 Research and Links:
>
> http://clanmaclochlainn.com/R1b1c7/
> -------------------------------
> To unsubscribe from the list, please send an email to
> with the word 'unsubscribe' without
> the quotes in the subject and the body of the message
>
R1b1c7 Research and Links:

http://clanmaclochlainn.com/R1b1c7/
-------------------------------
To unsubscribe from the list, please send an email to
with the word 'unsubscribe' without the
quotes in the subject and the body of the message



------------------------------

Message: 2
Date: Fri, 26 Aug 2011 09:00:24 -0400
From: Bill Howard <>
Subject: Re: [R-M222] The origin of R-M222 and the peopling of Ireland
- ohboy here we go again!
To:
Message-ID: <>
Content-Type: text/plain; charset=windows-1252

Just a short note about the determination of the time of origin of groups of
haplotypes, be they a SNP group or a surname group.

?The RCC correlation technique now has a time scale determined from
pedigrees.
?t appears to be linear over tens of thousands of years.
This means that the mutation rate has not changed
significantly over that time period.
?Haplotype analysis of only one set of markers is meaningless; it
must be compared with others.
?By the nature of this comparison, we can only determine the TMRCA of
the oldest pair.
But we want the time of origin of the progenitor, not of a
pair of people.
?If you form a phylogenetic tree, it contains ALL the testees that
are in the group you select.
?You can plot the run of descendant lines as a function of time. It
will be an exponential plot due to the growing number in the
population whose markers have separately mutated.
?A plot of the Log of the number (Log N) against the date will be
very nearly a straight line.
?You merely have to extrapolate that straight line to the point where
Log N=0. That's when the progenitor lived.

The main point here is that the straight line results by using every testee
on the plot, not just the ones who are more distant, AND this leads to only
a very short extrapolation on the plot, so it can be trusted. The R^2 value
of the plot is of the order of 0.98, yielding a very nice relation that you
can work with and trust.

The downside? Easy -- you need a very large number of testees in the group
to assure that you are not just determining the TMRCA of the group you
chose. How do you know when you have included enough testees? Well, as the
number of testees you analyze gets larger and larger, the inclusion of more
results tends to push the TMRCA further back in time, even with the
extrapolation. If you run the tree process on larger and larger numbers,
you will see the effect. But if you have two large groups, one of which
contains, say 340 testees and the other contains 680 testees, and IF the
derived date of the former is LONGER AGO than the one derived from the
latter, then you know you have had enough in the sample because the "law of
diminishing returns" has set in.

That happened when John McLaughlin and derived Figure 3 in our paper on
M222. We mentioned that effect there. It can be found at:
http://mysite.verizon.net/weh8/sitebuildercontent/sitebuilderfiles/M222Paper
.pdf
and it has been submitted to the journal Familia, connected to the
Ulster Historical Foundation where the readership in Ulster and Scotland who
carry M222 tends to peak up!

I truly think that this is now the best way to determine the TMRCA of groups
of haplotypes. The more the group shares the identified characteristic of a
group, the more meaningful the determination will be. You should not apply
the process to a random group of testees because the result will not be
meaningful.
-----------
By the way, prior to publishing my Paper 1 introducing the RCC approach,
Whit Athey and I had considerable discussions about how to determine the
TMRCA of a surname cluster whose membership was known to be incomplete. This
is EXACTLY the case here. In my Paper 1, it was determined that to determine
the time of origin, you had to find the TMRCA of KNOWN members of the
cluster by a factor of about 52.7/43.3 = 1.22 in order to estimate the TMRCA
of a cluster if all the members were present. When there are more testees
in the sample, that factor will be lower, as we see in the case of the M222
extrapolation (below).My Paper 1 is at:
http://mysite.verizon.net/weh8/sitebuildercontent/sitebuilderfiles/Howard1.p
df

Now, go to Figure 3 of the M222 paper. There you will see the last measured
point was at a date of about 1300 BC. The same sort of extrapolation applied
here would take you to the TMRCA of the M222 SNP, about 1680 years ago (SD~
300 years). If I had done this process on a number of surname clusters using
a phylogenetic tree approach, it would have given about the same result but
would have been more trustworthy. But I had not known of the tree approach
at the time I wrote that paper introducing the RCC correlation technique.
This is just a more clever way to determine the SNP => using the number of
descendant lines on the phylogenetic tree. Every testee contributes to the
resulting straight line on the tree!
----------
- Bye from Bill Howard


On Aug 26, 2011, at 3:39 AM, Sandy Paterson wrote:

> Right, it's coming back to me.
>
> Sometime during the past 6 months there was a shortish discussion in
> between Dave Johnston and Ken Nordtvedt.
>
> DJ wrote that he could show mathematically that the self-variance
> method of estimating group TMRCA in fact produced an answer that was
> an unbiased estimate NOT of the group TMRCA, but of the mean of the
> pair-wise TMRCA's within the group. Since that is necessarily lower
> than the group TMRCA, it means that the self-variance method it is biased
low.
>
> He went on to reason that in order to get an unbiased estimate of
> group TMRCA, variance should be calculated with reference to the
> putative marker value of the eponymous group "founder" rather than the
> mean marker value of the group.
>
> I don't wish to incorrectly attribute views to DJ or KN, but I got the
> impression that KN agreed with DJ, but added the caveat that the
> marker value of the founder is unknown. Yes, we can calculate modals,
> but although we can regard the modals as the 'best estimates' of the
> founder markers, we don't actually know them. That was my interpretation
of the discussion.
>
> Unfortunately, my computer crashed soon after that, so I'd be
> hard-pressed to track down the exchange.
>
> As I recall, DJ suggested that simply estimating the pairwise TMRCA of
> the two seemingly most distantly related haplotypes within a group
> (read Conroy and Doherty in M222), would probably be insightful, but
> the answer would be biased high.
>
> I think the discussion between DJ and KN was probably the most useful
> exchange I've ever read on the subject on TMRCA estimation, but it
> seems to have died off without much follow-up. Perhaps they continued
> the discussion off-list.
>
> So.
>
> I think the above supports your contention that we should be
> concentrating on the extremes rather than the group in its entirety.
>
> Sandy
>
>
>
>
>
>
>
>
>
>
>
> -----Original Message-----
> From:
> [mailto:] On Behalf Of Paul Conroy
> Sent: 25 August 2011 19:33
> To:
> Subject: Re: [R-M222] The origin of R-M222 and the peopling of Ireland
> - oh boy here we go again!
>
> Sandy,
>
> I just visually scanned the chart I linked to.
>
> What I'd like to see is analysis done of a 67-Marker, M222 SNP Tested
> sample, that excludes the well known Ui Neill families - that way we
> could see what relationships can be determined from the outliers,
> which might provide more clues?!
>
> On Thu, Aug 25, 2011 at 1:57 PM, Sandy Paterson <
> > wrote:
>
>> Hi Paul
>>
>> I'm interested to know why you include Gillespie in the group of
outliers.
>> My interest in them is that I share DYS481,487 = 26,14 with them, and
>> also that Gillespie and Ewing share DYS442 = 11.
>>
>> I agree with your general approach though - a better understanding of
>> the outliers is likely to provide a better understanding of M222.
>>
>> Sandy
>>
>>
>> -----Original Message-----
>> From:
>> [mailto:] On Behalf Of Paul Conroy
>> Sent: 25 August 2011 17:11
>> To: ;
>> Cc:
>> Subject: Re: [R-M222] The origin of R-M222 and the peopling of
>> Ireland -
> oh
>> boy here we go again!
>>
>> Susan,
>>
>> Yes, it has been discussed many times.
>>
>> One of the topics most near and dear to many on this forum is whether
>> M222 is Irish or Scottish, and especially whether particular families
>> are of Irish or Scottish descent. The truth is that Western Scotland
>> and Northern Ireland were almost the same country throughout history,
>> separated by a water super-highway, that far from inhibiting travel,
>> aided it. Almost every time a Scottish family can be shown to have
>> Irish roots, those Irish roots may have been Scottish still earlier,
>> and vice-versa - so to me it's pointless to try and distinguish, as
>> they were one breeding population for millennia.
>>
>> Also, as I don't seem to be related closely to Northern Ireland or
>> Western Scotland - though I do seem to possibly have share some
>> Off-Modals to the Nith Valley Cluster - what I'm more interested in
>> is the deep structure of the M222 group. I think I was the first
>> person about 3 or 4 years ago, to point out that the Ui Neill
>> families were extraordinary fecund, especially as their descendants
>> were Medieval rulers in Ireland and Scotland, and
> very
>> early colonists in the US, they have created a very lopsided
>> distribution of
>> M222+. But to determine the true source of the entire group, we need
>> M222+to be
>> able to account for the placing of all the outliers, such as:
>> 1. Conroy (myself)
>> 2. Galyean
>> 3. Gillespie
>> 4. McCord
>> 5. Cruden
>>
>> Check out this chart:
>> http://dna.cfsna.net/R-M222.jpg
>>
>> And all the unrepresented East Ireland, West Ireland, South Ireland,
>> North East Scotland, Central England, South England, Northern France,
>> South West France, North West France (Brittany), Belgium, Holland,
>> Germany, Norway, Denmark, Sweden - which to date are barely included in
this project.
>>
>> As I fairly closely some Daltons and some Stewards - both Norman
>> French families - it's my opinion that M222 came in more than one
>> wave from the Continent to the Isles, and probably has a source
>> somewhere between
> coastal
>> between Brittany and Holland. Coming over first to South England and
>> South East Ireland and possibly West Ireland as the Belgae tribes,
>> and to South Ireland as the Erainn tribes, then later again as Norman
>> French, to Scotland, England and Ireland.
>>
>> Cheers,
>> Paul
>>
>> On Thu, Aug 25, 2011 at 11:34 AM, Susan Hedeen <
>> > wrote:
>>
>>> The debates we've been reviewing regarding M222 have been around for
>>> awhile. See this discussion. I'm sure you have probably seen it
>>> before, but almost dejavu (spelling an example of phonetics)
>>>
>>> http://www.worldfamilies.net/forum/index.php?topic=8846.10;wap2
>>> R1b1c7 Research and Links:
>>>
>>> http://clanmaclochlainn.com/R1b1c7/
>>> -------------------------------
>>> To unsubscribe from the list, please send an email to
>>> with the word 'unsubscribe' without
>>> the quotes in the subject and the body of the message
>>>
>> R1b1c7 Research and Links:
>>
>> http://clanmaclochlainn.com/R1b1c7/
>> -------------------------------
>> To unsubscribe from the list, please send an email to
>> with the word 'unsubscribe' without
>> the quotes in the subject and the body of the message
>>
>> R1b1c7 Research and Links:
>>
>> http://clanmaclochlainn.com/R1b1c7/
>> -------------------------------
>> To unsubscribe from the list, please send an email to
>> with the word 'unsubscribe' without
>> the quotes in the subject and the body of the message
>>
> R1b1c7 Research and Links:
>
> http://clanmaclochlainn.com/R1b1c7/
> -------------------------------
> To unsubscribe from the list, please send an email to
> with the word 'unsubscribe' without
> the quotes in the subject and the body of the message
>
> R1b1c7 Research and Links:
>
> http://clanmaclochlainn.com/R1b1c7/
> -------------------------------
> To unsubscribe from the list, please send an email to
> with the word 'unsubscribe' without
> the quotes in the subject and the body of the message




------------------------------



End of DNA-R1B1C7 Digest, Vol 5, Issue 284
******************************************


This thread: