Haplogroup R1b (Haplotype 35)
R1b is by far the most common haplogroup in Europe, and population geneticists are
striving to find ways to split it into subclades. Scientists have thus far been able to subdivide
the major portion of R1b (known formerly as R1b3, but perhaps now best known as R1b1c)
into two main groups, based on an RFLP test of the 49a,fTaq/I locus. The modal
haplotypes of these groups are Haplotype 15 - or the Atlantic Modal Haplotype, which
predominates in Western Europe - and Haplotype 35, which is found throughout Europe,
but occurs most often in Southeastern Europe and Western Asia.
Unfortunately, the RFLP test cited above is not commercially available. DNA genealogists
who seek to distinguish Haplotype 15 from Haplotype 35 must, for the time being, rely on
other differences between the two haplotypes. One major difference is that the most common
values for the DYS393 marker are 13 for Haplotype 15, and 12 for Haplotype 35. Otherwise,
the basic six markers for their respective modal haplotypes are identical, as shown below.
(DYS # 19/388/390/391/392/393)
14-12-24-11-13-13 (ht15 or AMH)
14-12-24-11-13-12 (Haplotype 35)
This variation in the values of the DYS393 marker may not seem like much, but it
could be quite defnitive. Studies conducted on the frequency of mutation events among
Y chromosome DNA markers indicate that DYS393 has one of the lowest mutation rates of the
12 main markers used to identify a haplotype. The Kayser study of 2000 demonstrated a
mutation rate of zero for DYS392 and DYS393, while the Heyer study of 1997 recorded a
mutation rate of zero for DYS390 and DYS393. Only DYS393 had a mutation rate of zero
in both studies. That does not mean that DYS393 never mutates, but it does suggest that
it mutates very rarely.
We may conclude that any given R1b haplotype from Western Europe with a DYS393
value of 12 could have acquired it through a rare one step mutation, but also that many -
if not most - of these haplotypes exhibited a DYS393 value of 12 for many generations,
and ultimately originated in Eastern Europe or Western Asia.
Research conducted by the administrators of the Border Reivers DNA Project
has identified numerous haplotypes in persons of British descent that show Haplotype 35
markers. Moreover, most of these haplotypes appear to originate from areas of Britain
near the Antonine Wall, Hadrian's Wall and other places of Roman fortification or settlement.
These areas include Galloway, Dumfries, Ayrshire and The Borders in Scotland, and
Cumbria, Yorkshire, Lancashire, Shropshire and Staffordshire in England.
Many of the Roman troops stationed in these areas came from Southeastern Europe
or Western Asia. They included Sarmatians, Dacians, Goths, Syrians, Mesopotamians,
Thracians and Anatolians. The Capelli study has shown that these areas also exhibit higher
E3b is found most commonly in North Africa, Iberia, the Mediterranean and the Near East,
and J2 occurs most frequently in the Near East, the Mediterranean and Western Asia.
The fact that all three groups - E3b, J2 and Haplotype 35 - have a similar origin in territories
of the Roman Empire, and occur at comparable frequencies in parts of Britain with a known
history of Roman settlement, suggests that they arrived in Britain through the same means.
Haplotype 35 Distribution In Eurasia
The chart below shows the locales in the Y-chromosome Haplotype Database where the
modal ht35 values (DYS # 19/392/393/389i = 14-13-12-13) occur. These locales are sorted in
order of match frequency. Among the top twenty, five fall in Greece or Macedonia, five fall in
Turkey and an additional five fall in other areas of Southwest Asia, particularly the Caucasus.
At least two more - Romania and Italy - fall in Southeast Europe or the Mediterranean. The
last three are scattered across Western Europe.
Of the next twenty, two thirds fall in Southeast Europe, the Mediterranean or Southwest Asia.
This spread is entirely consistent with the known occurrence of ht35.
Haplotype 35 Distribution Relative To Haplotype 15
Haplotypes that resemble R1b ht15 and ht35 occur all over Eurasia, but in differing proportions.
The chart below compares the frequency in YHRD of a basic ht35 haplotype
(DYS # 19/392/393/389i = 14-13-12-13) with a basic ht15 haplotype (DYS # 19/392/393/389i = 14-13-13-13).
The chart displays the match frequencies for the ht35 haplotype and the ht15 haplotype, and the
ratio of the former to the latter. As you may note, the highest ratios occur in Southwestern Asia
and Eastern Europe, within the vicinity of the Black Sea, the Caucasus and the Pontic Steppe.
Haplotype 35 Distribution In The British Isles
We have examined the Capelli data set for any haplotypes classed as R1b that resembled variants
of Haplotype 35. Very simply put, these would include any haplotype with a DYS393 value of 12.
There are several caveats we must observe in regard to our analysis. These include:
1) The high percentage of DYS393=12 R1b in a given area may simply reflect
the high proportion of R1b in general.
2) The areas where R1b is plentiful and indigenous will also be places where
R1b haplotypes exhibit the largest number of variations - including
haplotypes that resemble Haplotype 35.
3) DYS393=12 haplotypes in Britain are no doubt at least partially attributable
to mutations, and many of those will be more alike in "descent" to indigenous
British ht15, even though they may resemble Anatolian ht35 in "state".
4) One way to identify which parts of the British Isles may exhibit DYS393=12
haplotypes that are legitimately of Eastern European or Western Asiatic descent
is to calculate the ratios of ht35 to ht15. The areas that show ratios significantly
higher than the norm will be those where DYS393=12 R1b haplotypes of
continental origin could have supplemented those that arose from mutations.
The chart below calculates these ratios for the Capelli data set, and ranks
them in descending order. From these results, we have made the
1) Among those areas with the top ten ratios, two fall in areas of Central
England with a well-documented history of Roman settlement (Uttoxeter
and Southwell), and two fall in the Anglo-Scottish Border region (Penrith
and among the Border Reiver sample itself). It is well worth noting that
both Uttoxeter and Southwell also exhibit higher than average percentages
of E3b and J/J2, which may be associated with Roman settlement.
2) Pitlochry, in Perthshire, shows both a high ratio of ht35 to ht15, and
the highest proportion of J/J2 in the Capelli study. The center of
Roman settlement in Scotland (to the extent that there was one) was
3) Three areas in the top ten - Dublin, Argyll and Pembrokeshire
in Wales - exhibit high overall levels of R1b. Since these areas
have a strong Celtic history, and virtually no Roman settlement
(except for Wales), the high levels of apparent ht35 may simply be
due to mutations. However, Dublin and Argyll do have a history of
Norse colonization, and Haverfordwest had a large Flemish colony
from the 12th century onward. Since both Western Norway and
parts of the southern Netherlands exhibit fairly high levels of
ht35-like haplotypes in YHRD, this may explain the high ratios
here. (We duly note, however, that the Capelli Norway sample
shows no ht35 at all.)
4) The elevated ratio of ht35 in the Orkneys, and Cumbria to some degree,
may also have a Norse origin.
We tentatively conclude that ht35-like haplotypes in the British Isles
may have at least three sources:
1) Mutations occurring in an R1b population that is plentiful,
indigenous and highly various.
2) Roman settlement, especially in areas where higher than
average levels of E3b and J/J2 also occur.
3) Norse or Norman origin - possibly. Despite its absence from
the Capelli Norway sample, ht35-like haplotypes do occur
frequently enough in the YHRD "Western Norway" sample, and
among haplotypes of Scandinavian or Baltic origin in Sorensen.
|Geographical Locale||% ht15||% ht35||ht35 / ht15|
|Border Reiver Families||64.8||2.7||.042|
|Midhurst (West Sussex)||72.75||1.25||.017|
|Isle of Man||71||.000|
Haplotype 35 Correlated With Haplogroups J/J2
This link at Dienekes' Anthropology Blog describes how the frequency of R1b has been positively correlated with the
frequency of J2 in Southeastern Europe. According to Dienekes himself, the R1b in question is almost entirely of the
ht35 variety. This suggests that both ht35 and J2 came to Europe from Asia Minor and diffused through the Balkans and
the coastal areas of the Eastern Mediterranean to other parts of Europe.
Below are some charts that show the overall distribution of R1b in Italy (first panel), the distribution of DYS393=12
R1b in Italy (second panel), and the distribution of one sub-clade of J2 in Italy (third panel). Note how both DYS393=12
R1b and J2a tend to cluster in Southern Italy, the vicinity most influenced by the Eastern Mediterranean, while R1b in
general is concentrated in the North, closer to the Celtic" regions of Western Europe.
If ht35 and J2 are associated, it stands to reason that those samples with the highest correlation of J2 and R1b
DYS393=12 will include the highest proportion of R1b DYS393=12 haplotypes that are true ht35, as opposed to ht15
haplotypes that may simply have mutated on the DYS393 marker.
Proceeding on this basis, we have created the chart below. This compares the percentage of R1b DYS393=12
haplotypes to the percentage of haplotypes belonging to haplogroups J and J2 - which are all related, but of which
the vast majority are J2 in the Border Reiver database. We calculated the percent difference between the proportion
of suspected ht35 and the proportion of haplogroups J and J2. Then we sorted each locale in ascending order by that
difference, so that those locales whose proportions were most similar appear at the top.
(Please note that those locales which had suspected ht35, but no J2, or J2 but no ht35, were marked N/A and were
grouped at the bottom of the chart. Interestingly, many locales that had relatively high proportions of suspected ht35,
but no J2, included heavily Gaelic parts of the British Isles that saw no Roman settlers or troops, such as Ireland, and
Western and Northern Scotland. If ht35 and J2 are indeed fellow travelers, one might conclude that the R1b DYS393=12
haplotypes in these locales are most likely the result of random mutations from native Western European ht15.)
Of the top ten samples, three were from the Borders area - Penrith, Morpeth and the Reivers, three were from Southern
England - Cornwall, Norfolk and Chippenham, and three were from the central midlands or the Welsh Marches - Llangefni,
Uttoxeter and Southwell. All of these locales saw substantial Roman settlement or troop deployment.
Pitlochry, in Perthshire, has high proportions of both ht35 and J2. This is remarkable, and at first seems
anomalous, since Pitlochry is in the Central Highlands. However, when one remembers that the Antonine Wall stretched
across Scotland just south of Pitlochry, and that the city of Perth itself was originally a Roman town, these proportions -
although still surprising - do make a little more sense.
|% R1b DYS393=13 (i.e., ht15)||% R1b DYS393=12 (i.e., ht35)||% J/J2||% Diff. J/J2 from ht35|
|Border Reiver Families||64.8||2.7||4.2||55.6|
|Midhurst (West Sussex)||72.75||1.25||5||300.0|
|Isle of Man||71||N/A|
Haplotype 35 As A Marker Of "Traditional" Celtic Origins
Many DYS393=12 haplotypes show their highest match frequencies in places like
Turkey, Poland, Romania, Ukraine and the Caucasus. Sometimes the affinity with
such regions is quite compelling.
However, even if we accept an origin in, or a greater closeness to, samples in Eastern
Europe or Western Asia, that does not necessarily dispel the assumption that these R1b
haplotypes are fundamentally "Celtic". Indeed, ht35 haplotypes may ultimately be more
"Celtic" than the indigenous ht15 R1b haplotypes of Britain.
The strong showing of DYS393=12 R1b haplotypes in the Black Sea region may
reflect merely the eastern wanderings of the Celts, who become known as Galatians
when they settled in Turkey.
The easterly prevalence of DYS393=12 R1b haplotypes may even reflect an affinity
with other groups even further afield. According to "The Mummies of Urichi" by
Elizabeth Wayland Barber, the Tocharian mummies discovered in the Tarim Basin
of Western China not only resembled the tall, fair-haired Celts of Gaul. They also
showed striking similarity in material culture, especially in their fondness for plaid
woolen fabric, a universal Celtic trait. Barber also contends that these people may
have been Tocharians, and that the Tocharians spoke an Indo-European tongue
with a close if archaic similarity to the Celtic languages.
Even beyond the possible history of the Celts in Asia, there is a long-standing tradition
among many of British Celtic descent that the Celts originated in Scythia, and made their
way westward across Europe to Britain. This thesis was famously propounded in the
Declaration of Arbroath, and persists to this day.
The current thinking is that the British Celts were, by and large, not Celts at all - but merely
aborigines who acquired the Celtic language through cultural diffusion. Nevertheless, it is not
unreasonable to assume that the Celtic language was brought to Britain by someone. Maybe
those vectors of linguistic change also brought R1b haplotypes from the east.
The Problem Of "Structure" With Haplotype 35
One problem inherent in distinguishing between Haplotype 15 and Haplotype 35 is their extreme similarity.
Like the vast proportion of R1b, they both belong to that large group known alternately as R1b3 or R1b1c.
Their modal haplotypes, at least on a 12 marker level, are identical - save for marker DYS393.
Ken Nordtvedt, after research in the Sorenson database, claimed not to have found any "structure" among
DYS393=12 R1b haplotypes that would otherwise distinguish them from DYS393=13 R1b haplotypes.
Intrigued by Ken's results, we on the Border Reivers team attempted some analysis on our own, using the
data in Ysearch. We focused on DYS437, for which we compared the frequency of the values 14 and 15
for pairs of R1b haplotypes that otherwise differed only on marker DYS393.
Our results (derived from Ysearch on 11/4/2005) appear below:
The frequency of DYS437=14 is substantially higher among R1b haplotypes with DYS393 values of 12 than
among R1b haplotypes with DYS393 values of 13 - but only when the DYS390 marker equals the modal value of 24.
The frequency is 3 to 4 times higher in this subset for all the common values of DYS391.
However, the frequency of DYS437=14 is 50 percent higher among R1b haplotypes with DYS393 values of 13
when the DYS390 marker is 23, and there are zero qualifying R1b haplotypes with DYS393 values of 12 when
the DYS390 marker is 25.
The frequency of DYS437=14 among DYS393=13 R1b haplotypes actually remains fairly constant irrespective
of DYS390 marker values. That is not the case with its frequency among DYS393=12 R1b haplotypes. This may
be a function of the small number of such haplotypes that have DYS390 values of 23 or 25, but it is suggestive.
Our rationale for attempting to associate DYS437=14 with DYS393=12 R1b is that DYS437=14 is modal among
R1a haplotypes, and as such may reflect the ancestral value of this marker before the R1 haplogroup diverged into
R1a and R1b. There has been discussion to the effect that DYS393=12 R1b (or ht35) may be the earlier form of
R1b, and hence may be closer in some ways to the ancestral haplogroup.
Among R1b, a DYS390 value of 25 seems more closely associated with Irish and Iberian populations than with others,
almost as though it represents a "hyper" form of Western R1b. If DYS390=25 R1b was more likely than other varieties to
be Western in origin, then it would be reasonable to assume that a higher proportion of the DYS393=12 haplotypes in this
group would be the result of local mutations rather than legitimate Eastern R1b. Consequently, there might be less
correlation of DYS393=12 with DYS437=14, if the two together are truly associated with ht35.
The distribution of DYS437=14 is closest between DYS393=12 and DYS393=13 R1b when the DYS390 marker
equals 23. This makes sense to us. A number of researchers have found DYS390 values of 23 to be more common
along the North Sea coast and in the German-speaking areas of Central Europe. This would place DYS390=23 R1b
closer to the geographical homeland of Haplotype 35 than the Atlantic Modal Haplotype, suggesting that it might be a
slightly older variant of R1b. As an older variant, it might then be more likely to exhibit the older DYS437 value of 14
regardless of the value of the DYS393 marker.
Considering that there is a definite bias in the value of DYS437 between DYS393=12 and DYS393=13 R1b haplotypes
when the DYS390=24, this variant of DYS393=12 might be more likely to be true Haplotype 35 rather than the result
of local mutations.
This is all highly speculative, of course, and we are only presenting it here because - 1) DYS437 is indeed a slowly
mutating marker, and 2) it is thus far the only evidence of "structure" so far that we have been able to find. Please
take what we say here with a grain of salt, as DYS393=12 R1b is a pet concern of the Border Reiver Project
administrators and we are doing whatever we can to make sense of it. DNA genealogists are like the characters in
The Big Chill, who couldn't last a day without making a rationalization - except for us the quota is more like ten.
Ken Nordtvedt's Research On Eastern R1b
Prominent DNA genealogist Ken Nordtvedt has performed some analysis on "Eastern R1b" haplotypes.
He has meticulously collected STR values from DYS393=12 R1b individuals in the Ysearch and SMGF
(Sorenson Molecular Genealogy Foundation) databases who claim patrilineal ancestry in Eastern Europe or
the Eastern Mediterranean. These, if anyone, are likely to have authentic ht35 haplotypes.
Dr. Nordtvedt identified STR values he believes are typical of this variation, and these are shown in the table below.
(* As calibrated by DNA Heritage)
We don't have values for DYS441, DYS446 or DYS461 for most of our participants. However, we do have values
for DYS393, DYS458 and DYS464,a,b,c,d. Of the 34 DYS393=12 R1b for which we have at least 25 markers,
17 - or 50 percent - have DYS458 values of 16 or lower. Of these, 8 - or about a quarter of the total - also have
the tell-tale DYS464a value of 14. These include members of the Bell, Maxwell, Turner and Kerr clans.
My collaborator David Strong also has both characteristics, even though he himself is not listed in this database.
The bottom line is that approximately a quarter of the DYS393=12 R1b "Border Reivers" for whom we have
extended haplotypes also exhibit at least two additional criteria for an authentic ht35.
Sean Silver's Research On Jewish R1b
Another DNA genealogist, Sean Silver, has started a project to study Jewish R1b haplotypes, many of which are
found among American Jews of Eastern European descent. Mr. Silver has speculated that some of these R1b
haplotypes may be of Iberian origin, entering the Sephardic Jewish population during their sojourn in Spain and
Portugal and later entering the Askenazi population after the expulsion of the Sephardim from Iberia.
This is definitely plausible. Mr. Silver has identified a very distinct R1b haplotype with a DYS393 value of 12
that may have originated in either Iberia or the Middle East. The most common variation of this haplotype
is shown below.
The most common DYS458 value that accompanies this haplotype is 17, but other variations that exhibit DYS389i,ii
values of 12,27 or 13,28 rather than 13,29 have DYS458 values of 15 or 16. These may be the original values of the
ancestral haplotype and conform with the Ken Nordtvedt's modal DYS458 value for ht35.
The most common DYS464a,b,c,d values for this haplotype and its variations range are 15,15,16,16 - 15,16,16,16
- 15,15,16,17. These are a far cry from Ken Nordtvedt's modal DYS464a,b,c,d values of 14,15,16,18, and may
comprise an example of an ht35 haplotype that doesn't follow Dr. Nordtvedt's formula.
But is this haplotype really ht35? We don't know. Its ancestral place of origin among Sean Silver's
participants is almost invariably Eastern Europe - Poland, Russia, the Ukraine and Lithuania.
Could it actually be a mutated ht15 haplotype of Iberian origin? Possibly. The geographical match pattern for the
most common variant of this haplotype is shown below. (Please note that only Old World matches are included.)
|Krusevo, Macedonia [Aromun]||4.65|
|Nizhnii Novgorod, Russia||1.89|
|Adana, Southern Turkey [Eti]||.98|
|Santiago de Compostela, Spain||.97|