Agree or disagree?

[J. Hugh Sullivan]

DNA testing has changed the parameters of genealogy. The time when I
could assume I was Irish because the first Sullivan, O'Suilebhain
b.862, was Irish no longer exists. In fact my Y-DNA test indicates
that my ancestors were Anglo-Saxon from northwest Europe who wandered
into England after the last Ice Age. My test also indicates that my
line mutated about 3000 AD.

Let me illustrate the potential error in accepting surname as proof of
ethnicity or genealogy. Suppose my earliest ancestor was a Sullivan
and…
a. That Sullivan had a baseborn son by a woman named Wyatt and…
b. The son of that Wyatt had a baseborn son by a woman named
Hornblower and…
c. The son of that Hornblower had a baseborn son by Frau von
Braun and…
d. The son of Frau von Braun had a baseborn son by Senorita
Castro.
…and each time the youngster was given the surname of his mother as
was the norm.

Then Senor Fidel would have the same YDNA as me. So would Herr Werner,
Captain Horatio and Sir Wyatt of Earp. Now suppose the first man in
line was a Wyatt or a Hornblower or a von Braun or a Castro?

I am a Sullivan only if my earliest proven ancestor was a Sullivan.
Without DNA testing we become family historians. If we are lucky we
are also genealogists.

Hugh (surname unknown)

[Charlie Hoffpauir]

Sort of agree....

Let me explain... I can trace my direct paternal line back only six
generations, that's my 3rd great-grandfather, Thomas Hoffpauir
(actually he spelled it Hoffpauer) In attempting to link farther back,
two cousins and I have done the Y-DNA tests (and we all match), so we
think we "know" Thomas' Y-DNA. However to say that once we identify
the ethnicity of Thomas, that "sets" our ethnicity is a bit of a
stretch, since the amount that Thomas contributes to my personal
genitic makeup is only 1/64. (Actually, it's somewhat more than 1/64
for me because Thomas appears 3 times in that 6 generation pedigree.)
But the point is that Y-DNA only represents a small portion of our
genetic makeup.

[J. Hugh Sullivan]

If I understand DNA tests, your 1/64 is based on autosomal tests which
represent all of your ancestors including female spouses. Y-DNA
represents the DNA you inherited from your father, he inherited from
his father - the Y-DNA in your case follows the surname from Thomas.

Is that not correct?

Hugh

[Charlie Hoffpauir]

On Sun, 13 Sep 2015 18:05:17 GMT, Ea...@bellsouth.net (J. Hugh

My 1/64 is based only on the fact that I have 64 3rd
great-grandparents, so my genes from any one is only 1/64 of my total
genetic makeup. The Y-DNA test I did enable me to link directly only
to my direct line male ancestor, or only one of those 64 ancestors.

[Charlie Hoffpauir]

On Sun, 13 Sep 2015 13:18:27 -0500, Charlie Hoffpauir

Or, to put it another way, although I am Y-DNA identical to Thomas,
genetically, I'm as related to any (all ) of the other 63 3rd great
grandparents as I am to Thomas.

[J. Hugh Sullivan]

On Sun, 13 Sep 2015 13:21:31 -0500, Charlie Hoffpauir

That gets into autosomal. As I understand you would not only be
related to the 63 others, you would be related to everyone they were
related to. Your mother was related to 64 and your father was probably
related to 64 others. They passed a recombination of those 22 genes
plus the x and y. That's 128 just in 2 preceding generations.

I'm not trying to explain anything to anyone . I'm just trying to make
sure I understand DNA testing and what it means.

Hugh

[Charlie Hoffpauir]

On Sun, 13 Sep 2015 20:09:27 GMT, Ea...@bellsouth.net (J. Hugh

Maybe one of us isn't understanding this. My parents are already in
that ancestry, as are their parents, etc. So yes, I'm related to
everyone they are related to, but calculating my relations by adding
their 6 generation calc to me is the same as just calculating my
relationship for 7 generations (I get 128 either way).

The only point I was trying to make is that when someone talks about
their Y-DNA tests indicating a certain area, or tribe, or some other
trait or whatever, that if it's 6 generations back (for example) their
are probably 63 different locations, tribes or whatever that the
someone is ALSO a genetic part of. The point I guess is that while
Y-DNA testing is useful, it only tells a very small part of one's
genetic background.

[J. Hugh Sullivan]

It appears that there are two conditions - genealogy and culture.
Genealogy is the history of the family and DNA can assist with that.

Since my DNA doesn't match any Sullivan but one known cousin either a)
enough Sullivans have not tested or b) I am not a Sullivan. But I am a
Sullivan from 1789 - possibly from 1690.

My DNA tells me that my paternal ancestors were from northwestern
Europe but not Scandinavia. Generally speaking those people migrated
to England and then to America. It tells me where I should be looking.
It tells me that my maternal ancestor was probably Irish (Sullivan)
but my paternal ancestor was not (R1a1a1a). Without the DNA test I
would still be researching Ireland.

Doesn't it seem that DNA testing is a tool just like census and court
records. While the written record tells me who my gg grandfather was
the DNA test tells me that 5000 years ago my ancestor was probably an
Anglo Saxon.

How much does possibly being from 63 other cultures matter? I am a
Southern WASP. Do I need to know anything else about my culture?

Am I the only one who feels a lack of knowledge in this complex area?

Hugh

[Charlie Hoffpauir]

On Mon, 14 Sep 2015 12:34:03 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote:
<snip>

>
>How much does possibly being from 63 other cultures matter? I am a
>Southern WASP. Do I need to know anything else about my culture?

I guess it's just natural to be concerned with the one ancestor that
has given one his surname. In my case, since that's the one that I've
hit a brick wall with, I tend to put more emphasis on the other 63.
accordingly, I'm interested in Foreman, Aucoin, Comeaux,Stutes,
Perilliard, Benoit, Boudrot, Savoie, Faulk, Hargrave, Perry,
Harrington, Hachley, Derouen, Prejean, Barre, Latiolais, Fontenot,
Thibodeaux, Brown, LaGrange, Pirella and Burnett.

On a related issue, as a user of Rootsmagic since it was first issued,
I have been requesting that they include a calculation of coefficient
of relationship (degrees of kinship)in the program, without any
success so far. It would seem to be an easy calculation to implement,
since one can do the calculation by "hand" by doing laborous simple
calculations. As a user of Legacy, do you know if it will make such a
calculation? The degree of kinship is useful for us southerners where
we have many cousin marriages in our ancestry.

[Richard Smith]

On 13/09/15 19:18, Charlie Hoffpauir wrote:

> My 1/64 is based only on the fact that I have 64 3rd
> great-grandparents,

Really? I only have 32.

Richard

[J. Hugh Sullivan]

On Mon, 14 Sep 2015 17:49:17 -0500, Charlie Hoffpauir
<inv...@invalid.com> wrote:

>On Mon, 14 Sep 2015 12:34:03 GMT, Ea...@bellsouth.net (J. Hugh
>Sullivan) wrote:
><snip>
>>
>>How much does possibly being from 63 other cultures matter? I am a
>>Southern WASP. Do I need to know anything else about my culture?
>
>I guess it's just natural to be concerned with the one ancestor that
>has given one his surname.

That might have been my point! :)

>In my case, since that's the one that I've
>hit a brick wall with, I tend to put more emphasis on the other 63.

Mine is pretty much limited to the 16 great greats. Beoyd that it not
much more than how many pages is the printout.

>On a related issue, as a user of Rootsmagic since it was first issued,
>I have been requesting that they include a calculation of coefficient
>of relationship (degrees of kinship)in the program, without any
>success so far.

I used Family Origins and I found Bruce very receptive to the 2 or 3
suggestions I made. But that was before he bacame famous. I would
still use RM AND Legacy if RM didn't make a change in name source when
using GEDCOMS back and forth. A name source in Legacy comes back as
"unassigned" from RM and name doesn't have a source. It prints
correctly. Sometimes I get nit-picky.

>It would seem to be an easy calculation to implement,
>since one can do the calculation by "hand" by doing laborous simple
>calculations. As a user of Legacy, do you know if it will make such a
>calculation? The degree of kinship is useful for us southerners where
>we have many cousin marriages in our ancestry.

It doesn't get any more Southern than me - Cheryl may be close.
Greenville MS is the only place I was ever born! And a little
University in Tuscaloosa knew me in the 40s.

Legacy does relationship calculations but I'm not sure what you mean
by degrees. If cousins married wouldn't their children still be
cousins to the children of the parents siblings?

Hugh

[Charlie Hoffpauir]

On Tue, 15 Sep 2015 14:51:50 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote:

<snip>

>Legacy does relationship calculations but I'm not sure what you mean
>by degrees. If cousins married wouldn't their children still be
>cousins to the children of the parents siblings?
>
>Hugh

There are a number of web sites that describe it, some highly
technical. A good one is at
http://www.genetic-genealogy.co.uk/Toc115570135.html
It's a bit different from the standard kinship that RM calculates (1st
cousin, 2 times removed, etc) in that it asigns a numerical value to
the relationship(s). For example, brothers would normally have a
coefficient of relationship of .5. However, it happens that their
parents were first cousins, their coefficient becomes 0.5625 (this
example is one that the cited web site uses). So, to someone like me
who has lots of cousin marriages in my tree, such information would at
least be interesting. Basically the calculation is:

R(XY) = S (1/2)n
(where n is supposed to be a superscript, ie 1/2 to the nth power)

Where R(XY) is the coefficient of relationship between the two
relatives X and Y and n is the number of connecting links or paths
separating them. S (meaning 'the sum of') refers to the fact that if
there is more than one connecting path, the paths are computed
separately and their coefficients are then added together. The type of
relationship, i.e. whether direct or collateral, affects the
procedure.

[Charlie Hoffpauir]

Right, I should have said 4th great- grandparents. I get mixed up
when looking at my ancestor Thomas Hoffpauir because he is in both the
4th g-grandfather spot and the 3rd g-grandfather spot.

[J. Hugh Sullivan]

I don't think my cousins actually married ...

Look for a program titled "Out Behind the Barn". ;)

Hugh

[Richard Smith]

On 14/09/15 13:34, J. Hugh Sullivan wrote:

> My DNA tells me that my paternal ancestors were from northwestern
> Europe but not Scandinavia.

On what basis do you say not Scandinavia? I think you've said
previously that your haplogroup is R1a and has the L664 mutation. From
what I've read, although L664 is most commonly found in Germany, the Low
Countries and the British Isles, Germanic peoples with the L664 mutation
also went to Scandinavia in the 2nd millennium BCE. Is there another
marker that distinguishes the Scandinavian bearers of L664 from the
Germanic ones?

Richard

[Charlie Hoffpauir]

On Tue, 15 Sep 2015 18:47:27 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote:

<snip>
>

>I don't think my cousins actually married ...
>
>Look for a program titled "Out Behind the Barn". ;)
>
>Hugh

LOL. If we're really talking genetics, it doesn't matter if the
married or not.

[J. Hugh Sullivan]

The following e-mail caused me to look at the possibility... He and
Dr. Klyosov appear to be experts on the "Tenths".

Martin Voorwinden, Co-administrator R1a1 Project (subgroup Tenths) by
e-mail… Because your haplotype shows DYS388=10 this means you are part
of the subgroup 2. (North-Western European Branch). This subgroup is
also called the Tenths because of the unique 10 for DYS388=10 which is
normally 12. Nearly all members of this subgroup (97%) have their
origin in the countries around the North Sea (British Isles,
Norway/Sweden, Denmark, NW-Germany, Netherlands). This subgroup is
further identified by the SNP L664 and belongs to one of the oldest
subgroups of R1a1 in Europe. On the British Isles you find about 72%
of the Tenths (0.4% of the population), Scandinavia 11%, NW-Germany
11%, and Netherlands 3%. The Tenths (DYS 388=10) can be subdivided in
several subclusters. You belong most probably to the largest
subcluster 2.D with typical DYS464=12-14-14-17 and most probably you
will also have the very rare DYS492=14 (normal value is 12). This sub
cluster covers about 50% of the Tenths and the far majority we find on
the British Isles. In our subclass R1a1-L664 we can distinguish 4 main
subgroups 2.A, 2.B, 2.C and 2.D. The largest one is 2.D (about 50%).
This subgroup 2.D has a unique mutation on DYS492 with the value 14
(nearly all other R1a1 have the value 12). Many members of subgroup
2.D also have a typical DYS464=12-14-14-17 (this combination is absent
in other R1a1-L664 members). Therefor I am quite sure you belong to
subgroup 2.D, but to be absolutely sure you must test 67 markers You
will most probably belong to 2.D2 and not 2.D1 (because your DYS447=25
and not 24).

I did test 67 markers and he is correct.

What I found on Internet...
R1a men consist genetically and geographically of two main STR types.
One can be called GENERAL and the other SCANDINAVIAN. The General type
is original; the Scandinavian type was developed later by a number of
STR mutations. The below chart compares the most essential two types
of markers. (the chart won't copy so I post the numbers)

DYS19
DYS391
YCAIIb
DYS442
DYS594
DYS565

The Scandinavian type numbers are 15,11,21,12,11,12
My numbers were 16,10,23,13,10,13

As you can see I don't match any of the numbers. From that I draw the
conclusion that I can't be Scandinavian.

I had rather be corrected if I am wrong than continue to think I know
something.

Hugh

[Denis Beauregard]

On Tue, 15 Sep 2015 21:10:26 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote in soc.genealogy.computing:

>What I found on Internet...
>R1a men consist genetically and geographically of two main STR types.
>One can be called GENERAL and the other SCANDINAVIAN. The General type
>is original; the Scandinavian type was developed later by a number of
>STR mutations. The below chart compares the most essential two types
>of markers. (the chart won't copy so I post the numbers)
>
>DYS19
>DYS391
>YCAIIb
>DYS442
>DYS594
>DYS565
>
>The Scandinavian type numbers are 15,11,21,12,11,12
>My numbers were 16,10,23,13,10,13
>
>As you can see I don't match any of the numbers. From that I draw the
>conclusion that I can't be Scandinavian.
>
>I had rather be corrected if I am wrong than continue to think I know
>something.

Some people would consider STR as statistics to help to chose SNP
tests. In other words, STR would suggest for example you are R1b
but more testings with SNPs can say you are not.

SNP are the mutations themselves.

As a comparison, you can say DNA testings are to identify the
language of a book. SNP would be looking for specific words
while STR would be counting the numbers of A, B, C, etc.

Have you made a test with SNPs like Geno2 ?


Denis

--
Denis Beauregard - généalogiste émérite (FQSG)
Les Français d'Amérique du Nord - www.francogene.com/genealogie--quebec/
French in North America before 1722 - www.francogene.com/quebec--genealogy/
Sur cédérom à 1785 - On CD-ROM to 1785

[J. Hugh Sullivan]

On Tue, 15 Sep 2015 17:52:36 -0400, Denis Beauregard
<denis.b-at-f...@fr.invalid> wrote:

>Some people would consider STR as statistics to help to chose SNP
>tests. In other words, STR would suggest for example you are R1b
>but more testings with SNPs can say you are not.
>
>SNP are the mutations themselves.
>
>As a comparison, you can say DNA testings are to identify the
>language of a book. SNP would be looking for specific words
>while STR would be counting the numbers of A, B, C, etc.
>
>Have you made a test with SNPs like Geno2 ?

HELP! MAYDAY! ...---...

I tested 37 then 67 markers. I am also L-664+. Isn't that an SNP?
Doesn't it say I can't possibly be R1b? If not what do I test to know
for certain?

I have heard of Geno2 and since it is National Geographic I trust it.
But that's all I know.

Given a choice I would rather be R1b. I did a lot more study on the
Irish. And it is possible in theory because the English and Irish
moved between the islands a lot.

Merci, Denis,

Hugh

[J. Hugh Sullivan]

On Tue, 15 Sep 2015 17:52:36 -0400, Denis Beauregard
<denis.b-at-f...@fr.invalid> wrote:

Additional predictions include:
If DYS426 is 12 and DYS392 is 11, one is probably a member of
haplogroup R1a1.
If DYS426 is 12 and DYS392 is not 11, one is probably a member of
haplogroup R1b.

After reading your post I looked at Wikipedia and found the above. I
am 12 and 11. How can I be anything but R1a1? I do see "probably" but
that;s my problem. Everyone hedges about DNA. I am a math major and I
like precise and exact.

If a squirrel runs between two holes in a tree and reduces his time by
1/5 second each lap, he CAN stick his head out of both holes at the
same time. But, if you call a dog's tail a leg, he still only has 4
legs.

Hugh

[Denis Beauregard]

On Wed, 16 Sep 2015 01:57:44 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote in soc.genealogy.computing:

>On Tue, 15 Sep 2015 17:52:36 -0400, Denis Beauregard
><denis.b-at-f...@fr.invalid> wrote:
>
>>Some people would consider STR as statistics to help to chose SNP
>>tests. In other words, STR would suggest for example you are R1b
>>but more testings with SNPs can say you are not.
>>
>>SNP are the mutations themselves.
>>
>>As a comparison, you can say DNA testings are to identify the
>>language of a book. SNP would be looking for specific words
>>while STR would be counting the numbers of A, B, C, etc.
>>
>>Have you made a test with SNPs like Geno2 ?
>
>HELP! MAYDAY! ...---...
>
>I tested 37 then 67 markers. I am also L-664+. Isn't that an SNP?
>Doesn't it say I can't possibly be R1b? If not what do I test to know
>for certain?

67 markers = YSTR tested
L-664 = SNP presumed from YSTR, and more exactly, R-L664
you can test directly this SNP (1 SNP is less expensive but you have
to know which one).

A SNP is not exclusive. It may occur in 2 different lines.

Please check this page :

http://www.isogg.org/tree/ISOGG_HapgrpR.html

You will see the hierarchy of the Haplogroup R. This line :

R1a1a1a CTS7083/L664/S298

So, R1a1a1a is the nodal notation. It represents the series
of mutations, i.e. R1a1 > R1a1a > R1a1a1 > R1a1a1a

CTS7083/L664/S298 are 3 equivalent mutations (SNP). I think they
are different codes for the same SNP. In comparison,
L146/M420/PF6229, L62/M513/PF6200, L63/M511/PF6203, L145/M449/PF6175
this would be 4 SNPs, each having 3 different names.

See also

http://www.isogg.org/tree/ISOGG_YDNA_SNP_Index.html

You will see the SNPs and their alternate names.


To be R1a1a1a, you should have those mutations and those of the
preceding groups (or subclads).

In other words, you would be part of :

R M207/Page37/UTY2, P224/PF6050, P227, P229/PF6019, P232, P280,
P285, PF6014/S9 M734/PF6057/S4/YSC0000201
• R1 M173/P241/Page29, M306/PF6147/S1, P225, P231, P233/PF6142,
P234, P236, P242/PF6113, P245/PF6117, P286/PF6136, P294/PF6112
• • R1a L146/M420/PF6229, L62/M513/PF6200, L63/M511/PF6203,
L145/M449/PF6175
• • • R1a1 M459/PF6235, L120/M516/PF6236, L122/M448/PF6237,
Page65.2/PF6234/SRY1532.2/SRY10831.2
• • • • R1a1a M512/PF6239, L168, L449/PF6223, M17, M198/PF6238,
M514/PF6240, M515
• • • • • R1a1a1 M417, Page7
• • • • • • R1a1a1a CTS7083/L664/S298

and have all those SNPs.

>I have heard of Geno2 and since it is National Geographic I trust it.
>But that's all I know.

Geno2 has about 150,000 SNP tested but in the 23 pairs and mt.
They try to have geographical/ethnical SNPs, those that are more
stable (compared to medical SNPs for example, that may be
associated to some illness).

The results from Geno2 is a sample of SNPs.

mt - the SNPs that are not standard and could be used to identify
the mt haplogroup (but set is less complete than only mt
at FTDNA for example)
Y - some SNPs, enough to identify the haplogroup with more accuracy
than the STR estimate, but may have false positive(s)

>Given a choice I would rather be R1b. I did a lot more study on the
>Irish. And it is possible in theory because the English and Irish
>moved between the islands a lot.
>
>Merci, Denis,

Bienvenu !

[Richard Smith]

On 15/09/15 22:10, J. Hugh Sullivan wrote:

> The following e-mail caused me to look at the possibility... He and
> Dr. Klyosov appear to be experts on the "Tenths".
>
> Martin Voorwinden, Co-administrator R1a1 Project (subgroup Tenths) by

> e-mail… [...]

That's all consistent with my understanding of the R1a haplogroup --
though that's not surprising as a lot of my knowledge has come from
reading Anatole Klyosov's papers on the subject.

However none of Martin Voorwinden's email (as quoted) implies a
non-Scandinavian origin. I think within the context of the R1a
haplogroup, the L664 SNP and the DYS388=10 STR are the same group of
individuals, and they say it's found in the British Isles, Scandinavia,
North West Germany and the Netherlands. Although Voorwinden discusses
subgroups of L664, he makes no comment on their geographic
distributions, and I've not seen any discussion on that elsewhere.

> What I found on Internet...
> R1a men consist genetically and geographically of two main STR types.
> One can be called GENERAL and the other SCANDINAVIAN. The General type
> is original; the Scandinavian type was developed later by a number of
> STR mutations. The below chart compares the most essential two types
> of markers. (the chart won't copy so I post the numbers)

This appears to be a comment by Kalevi Wiik on page 58 of the following:

http://www.jfpalmen.nl/files/Wiik_Haplogroup_R1a.pdf

Wiik is saying that R1a as a whole can be divided using STRs into two
genetic group which he calls "general" and "Scandinavian". Despite the
names, both groups are found in Scandinavia, as the maps on pp 60 & 62
show. (Note that the two figures in Norway correspond to genetic rather
than geographic divisions of the population.) What the maps show
somewhat, and the tables on pp 59 & 61 clarify, is that there is strong
correlation between Wiik's "Scandinavian" STR group, and the "Old
Scandinavian" branch of R1a, abbreviated Nor(OSc) by Wiik.

"Old Scandinavian" is the name often given to the Z284-Z287 branch of
R1a, and is far removed from your L664 branch, which is sometimes called
the "North Western" branch. The fact that you are an almost textbook
non-match for Wiik's "Scandinavian" group is little more than a
statement that you don't belong to the "Old Scandinavian" (Z284-Z287)
branch. That's as expected, as L664 is part of Wiik's vastly bigger
"general" group. (His division of R1a into two such unevenly-sized
groups makes sense in context of the paper which is the Finnish R1a
population, where Z287 is common.)

To summarise, the STR results tell you nothing you didn't already know.
Voorwinden's statement that "the British Isles [contains] about 72% of
the Tenths [R1a-L664], Scandinavia 11%, NW-Germany 11%, and Netherlands
3%" is still the best handle you have on its geographical distribution.
Within the British Isles, they're pretty well distributed between
England and Ireland.

From the information you've given, there's no reason to suppose your
recent paternal ancestry is English rather than Irish, or further back
that it is Anglo-Saxon rather than Viking. Perhaps you can make
progress by finding out more about the four subgroups of the Tenths that
Voorwinden refers to; maybe they will give more of a clue.

Richard

[J. Hugh Sullivan]

On Wed, 16 Sep 2015 10:50:32 -0400, Denis Beauregard

It will take me several 40 hour workweeks to digest all that but I
really appreciate you trying to make me less ignorant..

I will now add beaucoup to my previous merci. But mon cher is out of
the question!

Hugh

[Richard Smith]

On 16/09/15 03:59, J. Hugh Sullivan wrote:
> On Tue, 15 Sep 2015 17:52:36 -0400, Denis Beauregard
> <denis.b-at-f...@fr.invalid> wrote:
>
> Additional predictions include:
> If DYS426 is 12 and DYS392 is 11, one is probably a member of
> haplogroup R1a1.
> If DYS426 is 12 and DYS392 is not 11, one is probably a member of
> haplogroup R1b.

Y-STRs are only indicators of the haplogroup. The difference between
DYS392=11 and DYS392=12 is a change in the length of a short tandem
repeat (STR). Such mutations are rather more common than the single
nucleotide polymorphisms (SNPs) that define the haplogroups. A
particular SNP such as L664 may have occurred more than once in human
history, but a STR like DYS392=11 has almost certainly happened many times.

DYS392=11 is indicative of R1a1, but can occur in other haplogroups too.
Y-STR tests therefore rely on statistical analysis of several STR
markers. They know the correlation between STR markers and haplogroups,
and which markers are independent of each other. This is used to
calculate the probability that the individual belongs to each
haplogroup. With enough data, the calculation usually produces a very
high (even near-certain) correlation with one particular haplogroup,
even if one apparently indicate another haplogroup.

Testing a few dozen STR markers will identify the haplogroup with
reasonable specificity; to get a comparable degree of specificity from
SNP tests, you'd need to test hundreds or thousands of SNPs. The former
is currently cheaper which is why Y-STR testing is more popular.

Richard

[J. Hugh Sullivan]

The hurrider I go the confusider I get. :)

From a psychological standpoint that's good news. Sullivan is an Irish
name. I am a Sullivan, I want to be Irish - nothing against the
stalwart men of John Bulldom. Can I be Irish and not frequent taverns
very often?

Tne Celtics settled Ireland and they apparently descend from King
Milesius who died in Espana. The northwest Europeans did not settle
Ireland although everybody was everywhere.

I think my question now would be... Almost all Sullivans are R1b and
those who can trace their ancestors back to Ireland - in theory to
O'Suilebhain b. 862 AD. How did I become R1a?

Bottom line - maybe it doesn't matter. I have had an Irish surname for
5 provable generations so I had an Irish ancestor somewhere. Whether
my ancetors were Saxons or Neanderthals 5000+ years ago doesn't
matter. All I should be looking for in DNA is an MRCA who has the
genealogy that enables me to go back one or more generations before
1789.

I'm Irish. If I can't be proven wrong am I right?

Hugh

[Denis Beauregard]

On Wed, 16 Sep 2015 19:25:30 +0100, Richard Smith
<ric...@ex-parrot.com> wrote in soc.genealogy.computing:

>Testing a few dozen STR markers will identify the haplogroup with
>reasonable specificity; to get a comparable degree of specificity from
>SNP tests, you'd need to test hundreds or thousands of SNPs. The former
>is currently cheaper which is why Y-STR testing is more popular.

Not necessary cheaper.

Smith is a common name but mine is not. About 98% of Beauregard
living in Quebec or descendants thereof have one ancestor and his
signature is known. So, when the ancestor is known with a signature,
you can merely check one SNP to confirm he is the ancestor.

With 2 or 3 ancestors all with a signature, then you can also
test one SNP for each if the line is not known, or the one
with a documented trail.

So, when there is a good database of tested descendants, then the
SNP test can be cheaper. But before that, a few people were already
tested so they paid to clean the path to the right SNP. And some
SNP can appear twice or more, so testing 2 or 3 SNP, when they
are already detected, is better than one, but still cheaper than
a 37 or 67 markers YSTR test.

[J. Hugh Sullivan]

On Wed, 16 Sep 2015 18:47:55 GMT, Ea...@bellsouth.net (J. Hugh
Sullivan) wrote:

Looking at the below URL I think I have answered my own question.

http://www.khazaria.com/genetics/irish.html

>The "Celtic" Irish people of the emerald isle of Ireland are closely related to the Scottish people of nearby Scotland, and Irish and the partly Frisian-Anglo-Saxon English people from England are also significantly related. This shows the limitations of assuming we know everything about somebody's ancestry merely based on what language their ethnic group traditionally spoke (in this case, Irish Gaelic versus English). Also, some Irish people moved to Iceland and are thus partly related to modern Icelanders.

>R1b, which originated in western Europe, is the most common Y-DNA haplogroup among Irish men, at a frequency of about 81.5%. I1 is the second most common with 6%, followed by I2b at 5%, R1a at 2.5%, and E1b1b at 2%. G2a is found in only about 1%. Also rare are I2a (1%) and J2 (1%).

So, 2.5% of Irish men are R1a.

If I am to presume, the better presumption is that I am a
comparatively rare Irish Sullivan. If no one can determine who my ggg
grandfather was, knowing that I descend from a culture that mutated
5,000 years ago won't help.

What will help is a match that has genealogy that will enable me to
determine our MRCA and reconstruct who my ggg grand was.

IOW I need to redirect my focus.

Hugh

[Richard Smith]

On 16/09/15 19:47, J. Hugh Sullivan wrote:

> Tne Celtics settled Ireland and they apparently descend from King
> Milesius who died in Espana.

Milesius himself is almost certainly a later invention, and I think the
idea of significant Iberian settlement in Ireland is now largely
discredited. About 10 years ago, Brian Sykes gave the theory a boost in
credibility using Y-DNA results, and from memory this is described in
/The Seven Daughters of Eve/. However, I believe more recent and
detailed analysis has not borne it out.

> The northwest Europeans did not settle Ireland although everybody was everywhere.

The Vikings raided Ireland and founded a colony in Dublin. They may
well have brought the L664 line with them.

It's generally believed the Celts were responsible for the Hallstatt
culture in the 6th century BC in what is now Bavaria and Austria, and
expanded from there. The migration from central Europe to Ireland would
have taken generations, and perhaps on the route a Celtic woman had a
child with a man of north-west European origin?

> I think my question now would be... Almost all Sullivans are R1b and
> those who can trace their ancestors back to Ireland - in theory to
> O'Suilebhain b. 862 AD. How did I become R1a?

Most likely you are not a male-line descendant of the 9th century
Súilleabháin. That could be through illegitimacy, adoption, a
non-paternity event, or someone assuming the Sullivan surname for other
reasons. And it could have happened at any time between 862 and 1789.
Perhaps it happened relatively recently, while your family was in
America. But it could equally have happened many centuries earlier in
Ireland. I think you've said previously that you've found a few R1a
(O')Sullivans, even though they're not particularly close matches to
you. Were any R1a-L664? If so, maybe the dodgy link was centuries ago,
back in mediæval Ireland, and you all descend from it.

I think you've said you've traced Sullivans in Virginia back to the
1690s, even though you've not proven your descent from them. I don't
know all that much about American immigration at that time. How common
was Irish immigration to America that early? I'm wondering whether it's
more likely that those Sullivans came to America directly from Ireland,
or whether they spent a few intervening generations in Britain. It's
seems pretty uncommon in Britain at that time -- there are no
(O')Sullivans in any surviving Hampshire baptism or marriage register
until 1740, and just one earlier burial (a visiting soldier).

Richard

[Ian Goddard]

On 16/09/15 21:54, Richard Smith wrote:
> The Vikings raided Ireland and founded a colony in Dublin. They may
> well have brought the L664 line with them.

And probably not just Dublin. Strangford & Carlingford and Viking names.

--
Hotmail is my spam bin. Real address is ianng
at austonley org uk

[J. Hugh Sullivan]

On Wed, 16 Sep 2015 21:54:10 +0100, Richard Smith
<ric...@ex-parrot.com> wrote:

>The Vikings raided Ireland and founded a colony in Dublin. They may
>well have brought the L664 line with them.

From what I have read Vikings raided but didn't establish settlements.
No matter because going back that far doesn't help me like I thought
it might.

>It's generally believed the Celts were responsible for the Hallstatt
>culture in the 6th century BC in what is now Bavaria and Austria, and
>expanded from there. The migration from central Europe to Ireland would
>have taken generations, and perhaps on the route a Celtic woman had a
>child with a man of north-west European origin?

That's what I thought all along. But I didn't think the man could be
R1a and Irish.

>> I think my question now would be... Almost all Sullivans are R1b and
>> those who can trace their ancestors back to Ireland - in theory to
>> O'Suilebhain b. 862 AD. How did I become R1a?
>
>Most likely you are not a male-line descendant of the 9th century
>Súilleabháin.

I agree and even if I was I would need to identify about 33 more
generations to link - I can't find 1 more.

>That could be through illegitimacy, adoption, a
>non-paternity event, or someone assuming the Sullivan surname for other
>reasons. And it could have happened at any time between 862 and 1789.
>Perhaps it happened relatively recently, while your family was in
>America. But it could equally have happened many centuries earlier in
>Ireland. I think you've said previously that you've found a few R1a
>(O')Sullivans, even though they're not particularly close matches to
>you. Were any R1a-L664? If so, maybe the dodgy link was centuries ago,
>back in mediæval Ireland, and you all descend from it.

I now see that. I thought everything was cut and dried if I could just
find it. It may be science but it's not an exact science. Now they
tell me!

>I think you've said you've traced Sullivans in Virginia back to the
>1690s, even though you've not proven your descent from them. I don't
>know all that much about American immigration at that time. How common
>was Irish immigration to America that early?

John O'Sullivan came to VA in 1655 but he is R1b. Others who might be
my ancestors came closer to 1700. I find three immigrant Sullivans
that no one links to. I think I link to one and I think I know about
the others. I think I have done the genealogy of every Sullivan in the
1790 NC census - plus Wyatts and Tomlinsons. Most have not been done
by others that I can find.

I think I am convinced that many early ancestries are too convoluted
and controversial to give much attention.

I plan to concentrate on 3 areas for my website and will revise it.
I will should what I know on my proven ancestor. I will show who I
believe are the 5 generations before him including the immigrant.

Because I am a Sullivan I will presume I am Irish since there seems to
be no way to prove I am not - or am. But that has little to do with
genealogy. Genealogy is waiting for another match to learn whether I
can determine an MRCA and link from there.

I think I need to find another hobby because this one is done. If
there is a needle in a haystack I can find it. I'm not interested in
learning whether one is there or not.

Thank you very much for the help.

Hugh

[Ian Goddard]

On 14/09/15 23:49, Charlie Hoffpauir wrote:
> On a related issue, as a user of Rootsmagic since it was first issued,
> I have been requesting that they include a calculation of coefficient
> of relationship (degrees of kinship)in the program, without any
> success so far. It would seem to be an easy calculation to implement,
> since one can do the calculation by "hand" by doing laborous simple
> calculations. As a user of Legacy, do you know if it will make such a
> calculation? The degree of kinship is useful for us southerners where
> we have many cousin marriages in our ancestry.

I've thought about this from a different direction: estimating extent of
pedigree collapse.

My first approach was this: one would expect, for example, 8 ggparents.
Call these roles. If there's a cousin marriage such that one pair of
ggparents appears in different lines there will be 6 individuals filling
the 8 roles so there are 2 less individuals than expected. 2/8 = 25%
collapse. In another situation one might have 1 ggparent marrying twice
with a line to a child of each of these marriages. In that case there
is only one individual less than the number of roles so the collapes is
1/8 = 13.5%

However, like Charlie, I have the situation of the same pair appearing
in different generations. The next version was to say that with 3
generations of parents one would have 2 (parents) + 4 (grandparents) + 8
(ggparents) = 14 roles so that the cousin marriage above is missing 2
out of 14 giving 2/14 = ~14.3% and in the remarriage example it's 1/14 =
~7.1%.

This enables us to deal with the situation where a couple appear as
ggparents in one line and gggparents in another. If we extend the count
to the gggparents there are now 2+4+8+16=30 roles. Assuming it's just
one couple who are duplicated there are 28 individuals filling these
roles and the collapse is 2/30 = ~6.7%.

The general formula, therefore, is to count r, the number of roles for
which an ancestor has been identified and i, the number of individuals
identified as filling those roles and the pedigree collapse is given by
(r -i)/r.

[Ian Goddard]

On 15/09/15 18:12, Charlie Hoffpauir wrote:
> There are a number of web sites that describe it, some highly
> technical. A good one is at
> http://www.genetic-genealogy.co.uk/Toc115570135.html
> It's a bit different from the standard kinship that RM calculates (1st
> cousin, 2 times removed, etc) in that it asigns a numerical value to
> the relationship(s). For example, brothers would normally have a
> coefficient of relationship of .5. However, it happens that their
> parents were first cousins, their coefficient becomes 0.5625 (this
> example is one that the cited web site uses). So, to someone like me
> who has lots of cousin marriages in my tree, such information would at
> least be interesting. Basically the calculation is:
>
> R(XY) = S (1/2)n
> (where n is supposed to be a superscript, ie 1/2 to the nth power)
>
> Where R(XY) is the coefficient of relationship between the two
> relatives X and Y and n is the number of connecting links or paths
> separating them. S (meaning 'the sum of') refers to the fact that if
> there is more than one connecting path, the paths are computed
> separately and their coefficients are then added together. The type of
> relationship, i.e. whether direct or collateral, affects the
> procedure.
>

How does this handle the situation where an ancestor has married twice
and there is a descent from each marriage? This is clearly a lesser
degree of kinship than both lines going back to a single marriage.

[J. Hugh Sullivan]

On Wed, 16 Sep 2015 10:50:32 -0400, Denis Beauregard
<denis.b-at-f...@fr.invalid> wrote:

I think I have been getting the "hind part before".

The Bible starts "In the beginning..."; genealogy starts at the end.
Genealogy doesn't get me from there to here, it gets me from here to
there. The repopulation of the British Isles started when the ice
started melting; my genealogy starts with me. Constantly doing
descendant printouts vice ancestral further clouds the issue.

My surname is Irish; my earliest proven ancestor has the same Irish
surname. So, since 1816 (when he first appeared, not when he was born)
I am Irish. The Irish aren't really Irish and the British aren't
really British. Go back far enough and they were both something else -
Celtic, Basque, Saxon, Norman, Viking...Scythian? with a sprinkle of
Neanderthal.

My focus should not be whether my ancient ancestors were members of a
specific tribe of settlers - and it has been. I was trapped by my
Y-DNA test because I misread the meaning and tried to start at the
beginning. What I need to look for is a matching test that will help
me get 1 step further back.

It just took me longer to catch on than most people.

Hugh

[Richard Smith]

On 17/09/15 13:56, Ian Goddard wrote:

> How does this handle the situation where an ancestor has married twice
> and there is a descent from each marriage? This is clearly a lesser
> degree of kinship than both lines going back to a single marriage.

That's what the sum is for. Let's take a simple example of two full
siblings. Their coefficient of relationship is 0.5 = (1/2)^2 + (1/2)^2.
One (1/2)^2 term comes because they share a mother, the other because
they share a father. The power of two comes because there are two steps
in the relationship: from one brother, up one step to the parent, and
then down one step to the other brother.

Had they been half-siblings, they would only have shared a single parent
-- either the father or the mother -- and the coefficient would only
have been 0.25 = (1/2)^2.

Richard

[Charlie Hoffpauir]

That's an interesting approach. I'll have to look into that in more
detail. In my case, the actual situation is a bit more complicated
than I've stated (not wanting ot overcomplicate the calculation). The
actual case is that I have multiplse common ancestors, the most common
being Hoffpauir and Foreman. Since this occurs mainly on my father's
side, I'll describe that. A Hoffpauir ancestor appears 3 times in the
generation that is my father'gg-grandparents, and a Foreman appears as
well 3 times in that generation. But some of these appearances are
children of others who also appear. So the goal at the start was to
find out whether I'm "more" Hoffpauir or Foreman.

So I generated a pedigree chart, and arbitrarily went 7 generations
beyond my father, assigning everyone in that generation a "1".The next
generation (next descendant) then gets 1/2 of that first generation's
genes, so I assigned them 1/2, and so on until I reached my father,
who receives 1/128 of his genes from each person in that starting
generation. Now some people appear multiple times, and so would
receive multiple assigned portions, depending on "when" they
appeared.... so I took the highest fraction that they were assigned,
and applied that to their other appearances in the pedigree.* Then I
calculated each line of descent, from the earliest Hoffpauir, and from
the earliest Foreman, and summed the fractions that each path
produced, and calculated the fraction that represents the portion of
my father's inheritance of each line.

The results are: he is .0078125 of any one ancestor in that starting
generation, .0195312 of Foreman, and .03125 of Hoffpauir. Of course,
that makes me 1/2 of each of those numbers.

So, the calculation wasn't very difficult, just simple division and
addition, something that a computer can do easily.... so why can't
some genealogy program include something like this?

*My logic here is that if they receive a higher portion of the genes
then even if they appear later in the pedigree, they still pass on
that correspondingly higher portion to their descendant.

[Richard Smith]

On 17/09/15 13:53, Ian Goddard wrote:

> I've thought about this from a different direction: estimating
> extent of pedigree collapse.

It's not quite the same thing, but a standard measure of pedigree
collapse is the coefficient of inbreeding.

http://www.genetic-genealogy.co.uk/Toc115570144.html

Genetically speaking, it can be interpreted as the probability of two
alleles being identical through descent, which (loosely) is the property
that gives rise to the harmful effects of inbreeding. The coefficient
is a number between 0 (entirely uninbred) and 1 (impossibly inbred), and
is often expressed as a percentage. I don't recall having seen a value
above 40%, even in the most inbred dynasties such as the Habsburgs in
the 17th century, the Ptolemies in the 1st century BC, or the Eighteenth
Dynasty in the 14th century BC.

Where it differs from your idea of pedigree collapse (as I understand
your idea), is that inbreeding is solely about how related a persons
parents are. If an individual's parents are each individually very
inbred, but are not related to each other, the child is not considered
to be inbred at all. In an extreme example, one could imagine a man
with only four great grandparents, because his paternal grandparents
were siblings, as were his maternal grandparents. This has a huge
amount of pedigree collapse -- 50% in that generation -- but no
inbreeding, because his mother and father are not related to each other.

> My first approach was this: one would expect, for example, 8
> ggparents. Call these roles. If there's a cousin marriage such that
> one pair of ggparents appears in different lines there will be 6
> individuals filling the 8 roles so there are 2 less individuals than
> expected. 2/8 = 25% collapse.

A nice property of this definition is that it is not a function of
which generation of ancestors you look at: two out of eight, four out of
sixteen, and eight out of thirty-two all work out at 25%.

For comparison, the coefficient of inbreeding in this case is 6.25%.

> In another situation one might have 1 ggparent marrying twice with a
> line to a child of each of these marriages. In that case there is
> only one individual less than the number of roles so the collapes is
> 1/8 = 13.5%
>
> However, like Charlie, I have the situation of the same pair
> appearing in different generations. The next version was to say
> that with 3 generations of parents one would have 2 (parents) + 4
> (grandparents) + 8 (ggparents) = 14 roles so that the cousin
> marriage above is missing 2 out of 14 giving 2/14 = ~14.3% and in
> the remarriage example it's 1/14 = ~7.1%.

An unfortunate property of this definition is that the pedigree collapse
is now a function of the number of generations considered, even when the
earlier generation contain no additional intermarriages. When
considering three generations you have 2/14 = 14%, but if you consider
another generation, you have 6/30 = 20%. As the number of generations
considers increases, the value converges on 25%.

> This enables us to deal with the situation where a couple appear as
> ggparents in one line and gggparents in another. If we extend the
> count to the gggparents there are now 2+4+8+16=30 roles. Assuming
> it's just one couple who are duplicated there are 28 individuals
> filling these roles and the collapse is 2/30 = ~6.7%.
>
> The general formula, therefore, is to count r, the number of roles
> for which an ancestor has been identified and i, the number of
> individuals identified as filling those roles and the pedigree
> collapse is given by (r -i)/r.

An alternative, and one that I've used in other contexts, might be to
consider a graphical representation of an ancestor table as a unit
square. Divide the square in quarters. The two left quarters represent
the father and mother, while the two right quarters will themselves be
quarter. This process continues with smaller and smaller squares for
increasingly distant ancestors. Parents occupy 1/4 = 25% of the square,
grandparents 1/16 = 6.25%, great grandparents 1/64 = 1.56%, etc.

If the base person's parents are cousins, the two great-grandparents are
shared. Imagine blocking out one of every duplicated ancestor on the
unit square. This blocks out two great-grandparents (1/64 each), and
all their ancestors (another 2/64). In all, 6.25% of the diagram is
blocked out, which is the same as the coefficient of inbreeding.

If the duplicates are on different generations, which should you block
out? For a simple (if extreme) example, consider an individual whose
parents were uncle and niece. This means two grandparents are also
great-grandparents. Do you block out the grandparents (25%) or the
great-grandparents (6.25%)? Intuitively, the right numeric answer seems
to be the geometric mean (12.5%), which again is the coefficient of
inbreeding.

It's worth noting that this method doesn't always give the coefficient
of inbreeding. Imagine a person whose paternal grandparents were
siblings. That results in two duplicate great-grandparents, or 6.25%,
but the coefficient of inbreeding is 0 because the individual's parents
are unrelated.

Richard

[Tony Proctor]

"J. Hugh Sullivan" <Ea...@bellsouth.net> wrote in message
news:55f59c80...@news.eternal-september.org...
> DNA testing has changed the parameters of genealogy. The time when I
> could assume I was Irish because the first Sullivan, O'Suilebhain
> b.862, was Irish no longer exists. In fact my Y-DNA test indicates
> that my ancestors were Anglo-Saxon from northwest Europe who wandered
> into England after the last Ice Age. My test also indicates that my
> line mutated about 3000 AD.
>
> Let me illustrate the potential error in accepting surname as proof of
> ethnicity or genealogy. Suppose my earliest ancestor was a Sullivan
> and.
> a. That Sullivan had a baseborn son by a woman named Wyatt and.
> b. The son of that Wyatt had a baseborn son by a woman named
> Hornblower and.
> c. The son of that Hornblower had a baseborn son by Frau von
> Braun and.
> d. The son of Frau von Braun had a baseborn son by Senorita
> Castro.
> .and each time the youngster was given the surname of his mother as
> was the norm.
>
> Then Senor Fidel would have the same YDNA as me. So would Herr Werner,
> Captain Horatio and Sir Wyatt of Earp. Now suppose the first man in
> line was a Wyatt or a Hornblower or a von Braun or a Castro?
>
> I am a Sullivan only if my earliest proven ancestor was a Sullivan.
> Without DNA testing we become family historians. If we are lucky we
> are also genealogists.
>
> Hugh (surname unknown)


Re: "My test also indicates that my line mutated about 3000 AD"? Wow! Those
DNA tests have become prescient. Cool! :-)

Tony Proctor

[J. Hugh Sullivan]

What took you so long?

Hugh

[Tony Proctor]

"J. Hugh Sullivan" <Ea...@bellsouth.net> wrote in message

news:5659bd2c....@news.eternal-september.org...

I was waiting for that mutation to occur, and had some spare time on my
hands to reply Hugh! :-)

Tony

[J. Hugh Sullivan]

On Sun, 29 Nov 2015 15:10:42 -0000, "Tony Proctor"

I would have corrected but I lost my newsgroup eraser.

Hugh