The Cimbri in Northern Italy

The Cimbri were a Germanic and/or Celtic tribe that attacked the Roman Empire, but were defeated in 101 BC at the Battle of Vercellae. Some accounts suggest that the Cimbri, including women and children, all perished. Others suggest that some of the Cimbri fled to the surrounding country and that their descendants survive in Italy today.

A forthcoming paper by Anders Borglum, et al. (2006), No signature of Y chromosomal resemblance between possible descendants of the Cimbri in Denmark and Northern Italy concludes that the Cimbri in Italy today are genetically distinct from Cimbri living in Himmerland or Denmark.

From the abstract:

Two European populations are believed to be related to the ancient Germanic tribe Cimbri: one living in Northern Italy, the other living in Jutland, Denmark. The people called Cimbri are documented in the ancient Roman historical record. Arriving from the far north their movements can be tracked from successive battles with the Romans. The Cimbri finally entered Italy from the northeast and were defeated at Vercellae (present day Vercelli) in 101 BC by Gaius Marius and his professional legions. Classical sources from the first centuries AD relate the homeland of the Cimbri to the coasts around the Elb estuary (northern Germany) or specifically towards the north (Himmerland in northern Jutland). In the alpine parts of Veneto, northeast of the historical battlefield, local traditions dating back to late medieval time, identify a local population as Cimbri living in Terra dei Cimbri. They are considered the descendents of the Germanic combatants that fled the battlefield at Vercelli. As the defeated Cimbri that possibly fled to the mountains of Northern Italy most likely would have been male (warriors), the present study investigated the possible Y chromosomal diversity of the two present populations using microsatellite markers and single nucleotide polymorphisms. While Cimbri from Himmerland resembled their geographical neighbors from Denmark for the Y-chromosome markers, Cimbri from Italy were significantly differentiated both from Cimbri from Himmerland and from Danes. Therefore, we were not able to show any biological relationship for uniparentally transmitted markers.

The authors looked at both SNP and STR data, and found no relationship between the populations. While it is not possible to dismiss the chance that some limited genetic ties exist, no large scale y-chromosome similarities were found.

Haplogroup R2

At the moment, one of the least common y-chromosome haplogroups in the Italy DNA Project is haplogroup R2.

This haplogroup has been in the spotlight a little bit recently, due to a paper by Sengupta et al. (2006) entitled Polarity and Temporality of High Resolution Y-chromosome Distributions in India Identify Both Indigenous and Exogenous Expansions and Reveal Minor Genetic Influence of Central Asian Pastoralists. Sengupta proposes an Indian origin for haplogroup R2 (along with F* and L1). R2 is found with high frequency throughout South Asia and with diminishing frequency to the west.

Haplogroup R2 is exceedingly rare in Europe, with one notable exception: the Sinte or Sinti. The Sinti are a historically nomadic people, often considered to be a subgroup of the Roma. It is estimated that there are 120,00 to 150,2000 Roma, Sinti, and Caminanti people in Italy. The Sinti populations tend to be genetic isolates, and R2 frequencies as high as 53% have been observed among them.

Recently, an outbreak of measles was found to be clustered among Roma and Sinti Italian populations.

R2 is also found among non-Roma Indian populations, of course, as well as Central Asians (Pakistanis, Kurds, Caucasians, etc) so there are several possible paths that R2 could have taken on its way to Italy.

Our project also has two members in haplogroup L, which is also found in India but is almost never observed among the Sinti or the Roma.

Haplogroup J2a1 Origins in Greece?

I have been slow to link to Dienekes Anthropology Blog, largely because some of religious and racial views that occassionally crop up there. Nonetheless, Dienekes does a great job of keeping up with the science of genetic and anthropological research in the Mediterranean and elsewhere and I read his blog regularly.

A post entitled Did haplogroup J2a1 originate in Greece?, from December 2005, is a great example of the kind of blogging he does and is relevant to many of the participants in our project. An extract:

However, it could be argued that mobility within the Byzantine and Ottoman Empires may have introduced J2a1 from Anatolia to the Balkans. However, this does not explain the high diversity of J2a1 in Romania and Italy which were little if at all affected by Anatolian populations.Moreover the idea that J2a1 originated in Greece also explains the coastal distribution of J2 in the Mediterranean, observed by Di Giacomo et al.. It is well-known that Greek colonization was especially maritime.

Read the whole thing here.

mtDNA 15497 G/A Polymorphism

A recent paper from Liquori et al., The mtDNA 15497 G/A polymorphism in cytochrome b in severe obese subjects from Southern Italy, looked at

whether the 15497 G/A mtDNA polymorphism (G251S) in the cytochrome b subunit of respiratory complex III, which has been associated with obesity-related variables and lipid metabolism in a Japanese population, is associated with severe obesity also in adult Caucasians from southern Italy.

15497 is a SNP that would be revealed in a full mtDNA sequence, not the more common HVR1 or HVR2 tests done for genealogical purposes. The G>A mutation had been associated with severe obesity, but his study found very weak support for a relationship.

The Liquori study concluded:

The mtDNA 15497 G/A polymorphism in cytochrome b was present in 0.6% obese subjects, two females whose lipid parameters and BMI were similar to those of the overall group. Therefore, this mutation may appear to contribute in rare instances to severe obesity but does not explain the majority of cases in our population. A more extensive genetic haplogroup characterization is required to identify associations to obesity in Caucasians.

What do we know about R1b1b?

One of the interesting things about a project like the Italy DNA Project is that we occasionally turn up some interesting data. Because of Italy's geographic and historical position, we occasionally find haplotypes and/or haplogroups that are not common in Europe as a whole. Many of these likely result from gene flow from Eastern Europe, the Middle East and Central Asia in both historical and prehistoric times. The existence of two R1b1b, one from Lazio and one from Sicilia, in our project is an example of this phenomenon.

R1b1b is a subclade of haplogroup R1b1, the most common haplogroup in Europe. Haplogroup R1b1 is defined by SNPs M343 and P25 (although P25 is not particularly reliable, as it is prone to back mutation), and contains four subclades:

• R1b1a, defined by the SNP M18
• R1b1b, defined by the SNP M73
• R1b1c, defined by the SNP M269
• R1b1d, defined by the SNP M335

The vast majority of R1b males in Europe are in subclade R1b1c (M269+), but the Italy DNA Project has two members (one confirmed and one pending confirmation) in subclade R1b1b (M73+). R1b1b is very rare in Europe (in fact, as far as I know these two Italians are the only M73+ samples discovered in Europe to date), but much more common in Central Asia.

M73 refers to the deletion of 2 basepairs (GT) at nucleotide position 260 on Y-chromosome locus G65537. The SNP was apparently discovered by P.J. Oefner and the first examples were reported in a paper by Underhill et al. (2000). Subsequent examples have been reported in papers by Cinnioglu et al. (2004) and Sengupta et al. (2006).

The Underhill paper reported six samples from Central Asia/Siberia. The Cinnioglu paper reported four samples from Turkey. The Sengupta paper reported ten samples from China, one from Japan, and eight samples from Pakistan. Combined with the two Italian samples from our project, we have identified a grand total of 31.

The six Underhill samples were never haplotyped (not publicly, anyway), but the Cinnioglu and Sengupta samples have all been haplotyped at 10 loci (DYS19, DYS388, DYS390, DYS391, DYS392, DYS393, DYS389I, DYS389II, DYS439, and DYSA7.2).

However, I believe (based on personal correspondence with a reliable source) that at least two of the samples identified as M73+ in the Sengupta paper are not actually R1b1b.

Additionally two of the M73+ in the Cinnioglu paper (haplotypes 442 and 443) have interesting values for DYS390. These two samples have DYS390=19, which is far below the reported value for this marker in any other published R1b samples. DYS390 is a multi-segment locus, and in rare cases the normally invariant flanking segments (DYS390.1, DYS390.2, and DYS390.3) mutate or suffer a deletion (partial or comple). The phenomenon was described by Peter Forster in an excellent 1998 paper, Phylogenetic Resolution of Complex Mutational Features at Y-STR DYS390 in Aboriginal Australians and Papuans.

I have excluded the Japanese sample from my analysis, but am including the remainder at the present time. Haplotype data for the 24 samples in this study are available here.

One of the two Italian samples has been haplotyped at 25 loci and the other haplotyped at 37 loci, but testing is still in progress. Neither has been tested for DYSA7.2 (a.k.a. DYS461) yet. When both samples are done, each will be haplotyped at 38 loci including all ten loci for which the Cinnioglu and Sengupta samples were tested.

Based on the results we have accumulated (i.e. nine common markers for 21 M73+ samples), it looks like the age of the R1b1b subclade is approximately 10,00 to 12,500 years. I made that estimate based on the pairwise genetic distance of the samples and an effective mutation rate (0.00069) from Zhivotovsky (2oo4), but such an estimate should be considered with caution at this stage of the investigation.

Given the distribution of samples, it seems probable that R1b1b arose in Central Asia (possibly near Kazakhstan, Kyrgyzstan, or Uzbekistan which were along the Silk Road connecting China, Turkey, and Italy) and then spread east and west from there.

Based on the limited haplotype data we currently have, there appear to be two distinct branches within R1b1b. The Italian, Anatolian, and Pakistani samples cluster into one branch, while the Chinese samples cluster into another. You can view a preliminary phylogenetic tree showing these clusters. I excluded the Japanese sample from the Sengupta paper, mentioned earlier, but included all the Chinese samples since it is not yet clear which of them is suspect. The reader should be aware that because our sample size is so small, one or two mistaken inclusions/exclusions could have a dramatic impact on the shape of the phylogenetic tree.

At this point, there is no clear defining modal for the group. The modal values for the few samples we have collected are similar to the Atlantic Modal Haplotype (AMH), with one weak exception: DYS393. The R1b1b samples tend to have slightly higher values (e.g. 14) at DYS393 than the AMH variety of R1b1c which has a modal of 13. But this marker is highly variable in both R1b1b and R1b1c, so DYS393 is not predictive for R1b1b.

Our hope is that more complete haplotyping of the existing samples combined with more systematic sampling of Central Asian populations (of the kind being done by the National Geographic Genographic Project) will allow us to refine our understanding of haplogroup R1b1b.

A Little Housekeeping

Over the weekend, I refined the subgrouping on the Y-DNA results page.

I added a grouping for R1b1c and split the grouping for J2 into J2a and J2b. This makes the size of each grouping much more manageable.

In a few cases I was able to rely on SNP testing to provide the new groupings, but in some cases I had to rely on my own judgement. Some results could not be reliably subgrouped, so there are still some cases of J2 and R1b1. If you have some reason to suspect that I have grouped your result erroneously, please email me.

Early Ethnic Italy

Last night I was Googling around and I found a really cool map of early Italy on the National Geographic website. If you click the link at the bottom of the page, you can download a higher resolution version as a PDF.

Oh, it is not too late to order Genographic Project public participation kits as holiday gifts for your friends and family.

150th Member

Over the weekend, we reached 150 participants!

We currently have 106 Y-DNA participants and 56 mtDNA participants.

Population Structure in the Mediterranean

A paper by Capelli et al. entitled "Population Structure in the Mediterranean Basin: A Y Chromosome Perspective" was published in March by the Annals of Human Genetics.

The authors identified four population clusters in the Mediterranean: North Africa, Arab, Central-East, and West. The Italian and Sicilian samples in the study fall into the Central-East Mediterranean cluster.

The conclusions are a little arcane, but the study found "limited genetic contribution of North African population" to the Italian and Sicilian samples, and also found that the Iberian (Spanish, French, etc) samples were distinct from those in the Central-East cluster. From page 217:

Of interest is the genetic separation that West Mediterranean samples from Iberia display vs. Central and East Mediterranean samples, as shown in this study by Y chromosome SNPs and STRs analyses and by autosomal data (Rosenberg et al. 2002). Investigation of the mtDNA distribution of genetic variation instead seems to support a more homogeneous situation for European Mediterranean populations. Higher female than male gene flow across populations and/or difference in population sizes of breeding individuals between the two genders have been suggested as possible explanations for this observation (Seielstad et al. 1998; Dupanloup et al. 2003).

The authors also concluded that a combination of Neolithic migration, Phoenician and Greek expansion, and Arab conquest played an important role in the genetic structure of the Central and Eastern Mediterranean basin.

Of course we know that genetic influences from more northern populations (Celts, Lombards, etc) are detectable as well, but examining that was beyond the scope of this particular study.

What's With The Ads?

We are experimenting with the addition of advertisements on the blog. You'll notice them on the right side of the page, at the bottom of the sidebar.

This space was empty before, so I thought I'd try to get some use out of it. I intend to limit the ads to relevant (I hope) products at Amazon.com and text ads from Google. So far, I've not been happy with the Google ads so there are none now but don't be surprised if they show up at some point.

If the ads generate any revenue, that money will flow to the Italy DNA Project's General Fund. The General Fund, in turn, will be used (and can ONLY be used) to subsidize tests and upgrades for project members at Family Tree DNA. I expect the Fund will be used to help folks who have genealogical hurdles that DNA could help them clear but who lack the means to pay for the proper tests. Also, the Fund will be used to pay for tests that are of limited genealogical benefit to the participant but of broad interest to the larger genetic community.