Autosomal analysis of the genomes of Iron Age Britons and Anglo-Saxons

[Maciamo]

Schiffels et al. (2014) tested two Iron Age Celtic samples and four early medieval Anglo-Saxon samples, all from Hixton in Cambridgeshire, East Anglia, England. The Iron Age Britond lived approximately 2,000 years ago, while the Anglo-Saxon individuals are dated to c. 1,300 years before present.

[N.B. Fire Haired started another thread on this subject here, but I would like to post my own detailed analysis in a separate thread for the sake of clarity. In the other thread Motzart posted a comparison of the Iron Age Briton's genome using Eurogenes' K15 admixtures with other ancient European samples. Angela also posted the K12 (div3) and K12b admixtures for both the Celtic Briton and the Anglo-Saxon.]

Here is a summary of the data available, with an analysis under each category. I have coloured the admixtures using the same colours as on the maps I made, if there is a map for that admixture.

Hg	Celt 1 (ERS389798)	Celt 2 (ERS389799)	Anglo-Saxon 1 (ERS389795)	Anglo-Saxon 2 (ERS389796)	Anglo-Saxon 3 (ERS389797)
Y-DNA	R1b-L21 ?	(female)	R1b-L11	(female)	(female)
mtDNA	H1ag1	H2a2a1	K1a1b1b	H2a2b1	K1a4a1a2

Nothing very surprising here. Both Celtic and West Germanic people are predominantly R1b. Maternal haplogroups are all fairly common in north-west Europe.


Dodecad K12 admixtures

Admixture	Iron Age Celt 1	Anglo-Saxon 1	Modern Britons
West European	59.73%	41.46%	65%
East European	8.22%	24.44%	3.2%
Mediterranean	24.91%	17.31%	22.8%
West Asian	0%	1.70%	6.7%
Southwest Asian	0%	0.85%	1.2%
South Asian	0%	3.17%	0.5%
Southeast Asian	0%	0.99%	0.1%
Northeast Asian	0%	1.21%	0.1%
Northwest African	0.03%	5.55%	0.2%
East African	0%	0%	0%
Neo African	1.69%	0.63%	0.1%
Palaeo African	5.41%	2.69%	0%

The Anglo-Saxon individual is the biggest surprise. His admixtures do not match any modern population. If we look only at the the West European, East European and Mediterranean admixtures, he appears to lie somewhere in between East Germans and Poles, probably close to the modern population of Mecklenburg-Vorpommern or Brandenburg. However, he also has over 2% of East Asian admixture and virtually no West Asian or Southwest Asian admixture, which places him closer to modern Finnish people (as reported in the paper) or Northwest Russians. Another oddity is his nearly 9% of African admixture (see below), which shouldn't be there at all.

The Celtic Briton is closest to modern British people in the West European, East European and Mediterranean admixtures. The higher East European and lower West European of the ancient Briton makes them even closer to modern Belgians. This isn't unlikely considering that many Belgic tribes (Atrebates, Menapians) settled in Britain during the Iron Age.

One major difference between this ancient sample (perhaps unrepresentative) and modern Britons is the high percentage (>7%) of African admixture (and mostly Paleo African at that !) in the Iron Age individual on the one hand, and his complete lack of West Asian or Southwest Asian admixture. Modern Britons have practically the reverse proportion, with close to 8% of West/Southwest Asian and almost no African at all.

Unless the data was inverted, it is very hard to explain why Iron Age samples possessed such high African admixtures. Here are possible explanations:

A) The admixtures reported as African are in fact non-African, but do not correspond to any modern sample, or represent another regional group not included in the K12 admixtures.

B) Africans possess some admixture in common with North Europeans that came from a very ancient common source. This could be the Early Neolithic split between R1b-V88 (who migrated to Africa) and R1b-M269 (who migrated to Europe).

C) Some Paleo African could be the remainders of Palaeolithic African A1a lineages that are still found sporadically from the British Isles to Scandinavia. After all, the Mesolithic samples from Sweden, Luxembourg and Spain all had a few percent's of African admixture too.

D) The Northwest African in the Anglo-Saxon sample may actually be Germanic DNA that ended up in Northwest Africa when the Vandals settled there. Being more common in Northwest Africa today it may easily have been mislabled as Northwest African.


It is easier to explain the increase in West Asian and Southwest Asian admixtures among modern Britons. In Europe, these two admixtures are highest in Italy and Greece, and would have spread around Western Europe with the Roman colonisation. The Iron Age sample dates from 2,000 years before present, which is approximately the time the Romans first set foot in Britain - obviously not long enough for having any impact on the Celtic population.

A second source of West Asian DNA would have been the Norman conquest. France has the highest West Asian admixture in Western Europe after Italy, and the Normans, despite their partly Danish roots, had a lot of French blood. The Normans appear to have had a considerable genetic impact on the British population. That may be because of the Normans dominated the English aristocracy, and the upper classes had more surviving offspring in the long run.



Dodecad K12b admixtures

Admixture	Iron Age Celt 1	Anglo-Saxon 1	Modern Britons
North European	39.04%	49.89%	44.5%
Atlantic-Med	32.43%	30.08%	43%
Caucasus	9.12%	9.78%	1.8%
Gedrosia	5.96%	0.08%	10.4%
Southwest Asian	1.63%	0%	0.2%
South Asian	2.90%	0.07%	0.1%
Southeast Asian	0%	0%	0%
East Asian	0%	3.63%	0%
Siberian	0.02%	1.16%	0%
Northwest African	2.63%	1.31%	0%
East African	2.73%	0.79%	0%
Sub_Saharan	3.54%	3.19%	0%

The K12b shows a quite different picture, because the regional division was done very differently. Instead of splitting Europe in Northwest, Northeast and Mediterranean, the split in more on a dual North-Northeast to South-Southwest axis. Here we see that modern Britons fit nicely in between the ancient Celtic and Anglo-Saxon samples for the North European admixture. But for the rest there is hardly any match between the ancient and modern samples.

Oddly enough, the ancient and modern proportions of Caucasian and Gedrosian admixtures appear to be reversed. Both the Celtic Briton and the Anglo-Saxon individuals have about 10% of Caucasian, which is the same as modern Belgians and Germans, slightly more than modern French people (8.5%), but much more than modern English (3%), Scottish (0.5%), Irish (0.2%) and Orcadian (0%) people. The Caucasian component is similar to the West Asian in K12 in that the both peak in the Eastern Mediterranean and could have been spread by the Romans and the Alpine Celts (Hallstatt and La Tène, including Belgic tribes). That would explain why the English have some Caucasian, although less than the French, but the Scots and Irish almost don't have any. If that is the case, the Iron Age Celt would surely be of Belgic or at least Hallstatt descent.

The main difference between the Caucasian and West Asian admixtures is that the former is also very common among northern Slavs (Poles, Belarussians, Russians), while the latter is not. That explains why the Anglo-Saxon individual has 10% of Caucasian, once again an intermediary value between modern Poles and Germans.

The Gedrosian component is conversely absent from North Slavic people and is also nearly absent in the Anglo-Saxon individual. This is consistent with the K12 data showing the closest affinity to the present-day northeast German and northwest Polish populations. The Celtic Briton has 6% of Gedrosian, much less than the 10.5% found in modern English people, the 12% of the Irish and Orcadians, or the 13% of Argyll Scots. Once again, this Iron Age Briton has closer affinities with modern Belgians (6.8%) and Germans (7.3%) than with the modern Celtic populations of the British Isles.

One thing that remains to be explained is how modern Britons got more Gedrosian admixture than any other Europeans (even Southeast Europeans), and also more than these two ancient samples. I had hypothesised before that the Gedrosian admixture came with R1b people during the Bronze Age, and represents one of the original admixtures of R1b people before the crossed the Caucasus to settled as Neolithic cattle herders in the Pontic Steppe. If that is the case, then true ancient Britons (R1b-L21) should have at least as much Gedrosian admixture as modern Britons. The reason why this Iron Age guy didn't is that he was an immigrant from the continent (Belgium, southwest Germany or Switzerland).

The Southwest Asian admixture, which was completely absent in both samples in the K12, shows 1.6% in the Celt. Once again that is closer to modern Belgians (2.5%) and Germans (1.7%) than to modern Britons (0.2%). The Anglo-Saxon has 0% this time (instead of 0.85%), which is less than any modern population, but closer to the Scandinavians and Balts.

Modern Germanic people lack the South Asian admixture at the K12b, just like the Anglo-Saxon sample. However, modern British and Irish people also lack it, while the Iron Age Celt has a surprisingly high 2.9%. No modern European population has more than 1% of South Asian, and that is the Poles, Russians and Romanians. The original Bronze Age Proto-Indo-Europeans may have had higher percentages though, and this Celtic guy may be an isolated case of someone who ended up with more South Asian segments than average.

The really surprising data is the tremendous amount of African DNA in both samples and East Asian DNA in the Anglo-Saxon. This was already observed in the K12, but the regional breakdown in African admixtures is now very different, which reinforces my suspicion that something is wrong and that this may not be African DNA at all (see above).

The East Asian component in the Anglo-Saxon is clear though, and even more pronounced here (4.8% instead of 2.2%). The best explanation is that the Huns contributed a few percent's of the Anglo-Saxon gene pool, or that this individual came from a family with an especially high Hunnic ancestry. I had theorised before that the Huns were the ones who brought haplogroup Q to Central Europe and Scandinavia. It is also possible that the Anglo-Saxon got some Mongoloid admixture from the Finns or Lapps, but that seems more far-fetched because modern Finns have exclusively Siberian admixture at the K12b, while this Anglo-Saxon guy has three times more East Asian than Siberian, which is an even higher proportion of East Asian than modern Mongols (36% Siberian, 45% East Asian).



Conclusion

Of the two Iron Age samples from East England tested by Schiffels et al. (2014), the Celtic Briton individual displays the greatest affinity to modern Belgian, then German, then French people, indicating a probable Belgic/La Tène or Hallstatt origin.

The Anglo-Saxon genome fits in between the modern populations of northeast Germany and northwest Poland, which may indicate an East Germanic (Gothic ?) origin. This is further corroborated by the fact that this individual possess a substantial percentage (2-5%) of East Asian and Siberian admixture, present in a proportion that is consistent with an introgression of Hunnic (Mongol) genes.

Both individuals have unexplained level of African admixture using the Dodecad K12 and K12b (both between 5 and 9%), which are probably misclassified admixtures of unknown origin, some of which may be shared with some African populations. Note that the Eurogenes K15 admixtures show 0% for both the Northeast African and Sub-Saharan admixtures.

Finally, both samples are characterised by very low levels of West Asian and Southwest Asian admixtures compared to the modern British population. The higher modern percentage could be explained both by the Roman colonisation and the Norman conquest.


Below: Anglo-Saxon helmet from the Sutton Hoo ship-burial, England.

[Wilhelm]

I don't think so, that he is closest to Belgians. On the PCA plot of David he is closest to modern British Isles :.

https://drive.google.com/file/d/0B9o...V0OWI3N0E/view

Also using the least-squares method here is the closest populations :

v3 K=12
Anglo-Saxon : Ukranian @ 12.07
Iron Age Briton : Orkney @ 9.9

K12b
Anglo-Saxon : German @ 9.7
Iron Age Briton: German @ 11.0


Eurogenes K13
Hinxton 1 : Irish @ 4.34
Hinxton 2 : West Scottish @7.23

Eurogens K15
Hinxton 1 : Orcadian @ 6.30
Hinxton 2 : Orcadian @ 6.56

[Sile]

see link below for all the SNP's for each individual

enter google drive

http://www.y-str.org/2014/10/hinxton-dna.html

[Maciamo]

I don't think so, that he is closest to Belgians. On the PCA plot of David he is closest to modern British Isles :.

I don't like PCA plots. That's an old, outdated method of comparison that only uses two dimensions and is therefore unreliable. I prefer to compare each admixture one by one.

v3 K=12
Anglo-Saxon : Ukranian @ 12.07
Iron Age Briton : Orkney @ 9.9

K12b
Anglo-Saxon : German @ 9.7
Iron Age Briton: German @ 11.0


Eurogenes K13
Hinxton 1 : Irish @ 4.34
Hinxton 2 : West Scottish @7.23

Eurogens K15
Hinxton 1 : Orcadian @ 6.30
Hinxton 2 : Orcadian @ 6.56

A perfect illustration that the least-squares method isn't any good either. How can you get such different results across calculators, and get exactly the same closest population for two individuals as different as these Iron Age Celt and Anglo-Saxons ?

[Fire Haired14]

Maciamo, check out what Davidski at Eurogenes is doing!

http://eurogenes.blogspot.com.au/201...m-hinxton.html

Hex 1 and 2 have several percentages higher of WHG and ANE than Irish(who have the highest in NW Europe besides Norse). They have about the same EEF/WHG/ANE ratio as Belorussians, Polish, and Scandinavians. Scandinavians though have much more eastern-like ancestry. Basically Hex 1 and 2 seem to be NW European to the extreme, more than any modern ones, with very little east European and Mediterranean affinities. Hex 2's kit # at GEDmatch is F999921.

Here are her Dodecade K12b results.

Gedrosia 10.69%
Siberian 0.76%
Northwest_African 0
Southeast_Asian 0
Atlantic_Med 35.86%
North_European 50.25%
South_Asian 0.11%
East_African 0.62%
Southwest_Asian 0
East_Asian 0
Caucasus 0.99%
Sub_Saharan 0.72%

She scores about 1% higher in Gedorsia than the highest score in west Europe today. Where did you get these admixture results from? Hex 1 and 2 have more from what changed in west Europe after the Neolithic than anyone around today.

[Sile]

some people want to say that he has L151 and does not have L11, but because L151 is in the same branch as L11 , he should be noted as L11 .................I disagree with this summary...........he is R-L151 as its his only positive SNP on that branch with L11

this is what is found by others

YBrowse Build 37 at http://www.isogg.org/tree/ISOGG_YDNA_SNP_Index.html downloading all of the Complete_Y.csv files from each run of ERS389795 .

ERS389795 has C allele at position 2657176 which is a positive marker for Page65.1/SRY1532.1/SRY10831.1 in haplogroup BT and not for R1a1-Page65.2/PF6234/SRY1532.2/SRY10831.2.

ERS389795 has the following -
R1b1 M415=rs9786194 9170545 C->A
R1b1a2 CTS8728/L1063/PF6480/S13=rs9786876 18167403 C->T
R1b1a2 L265/PF6431=rs9786882 8149348 A->G
R1b1a2a1a L151/PF6542=rs2082033 16492547 C->T

[Sile]

http://www.fc.id.au/2014/10/hinxton-2-analysis.html

for ERS389796



k23b seems to be the new "better" admixture test to do

[Wilhelm]

A perfect illustration that the least-squares method isn't any good either. How can you get such different results across calculators, and get exactly the same closest population for two individuals as different as these Iron Age Celt and Anglo-Saxons ?

The least-squares method is just the best method to get the closest numerical populations. Just look at the spreadsheet, and you'll see that it makes sense, the closest population is the one that will look the most similar statistically. But you also have to take in account the distances, is not the same @11 than @3. Also, the reasons why there are such differences across different admixture runs is simply becaue of the different components in which Europe is divided (K=12 has East-Euro/West-Euro/Med, while Eurogens K13 has North Sea/Atlantic/Baltic, West-Med.). Also, the samples are different between Dodecad and Eurogenes. The K12b was much limitied in european samples, than is the Eurogenes K13/K15

[Angela]

Transferred from the other thread so everything is together:
One of many questions that remains open for me is whether this ancient Briton is representative of all the Pre-Roman Iron Age inhabitants of the British Isles.

Once again, these are the results for what is probably a "Briton" or Iron Age Celt from what is now East Anglia:
K12b
39.04% North_European

• 32.43% Atlantic_Med
• 9.12% Caucasus
• 5.96% Gedrosia
• 3.54% Sub_Saharan
• 2.90% South_Asian
• 2.73% East_African
• 2.63% Northwest_African
• 1.63% Southwest_Asian
• 0.02% Siberian
• 0.00% East_Asian
• 0.00% Southeast_Asian

Now let's look at the Cornish sample:
North European 42.8
Atlantic Med 43.8
Caucasus 2
Gedrosia 11.4

All the African and South Asian and South West Asian percentages have disappeared.

The North Euro component is only 3 points higher, which one could say might be the result of internal migration and intermarriage over the centuries.

However, how did the Cornish go from 32.43% Atlantic Med (if our ancient sample was representative of all the pre-Roman Iron Age Britons) to 43.8% Atlantic Med? Their Atlantic Med numbers are pretty "French" looking. Even the Argyl (41.2), English (41.5), and Orcadian (42.4) Atlantic Med numbers are substantially higher than those of our ancient Briton.

So, I'm still left wondering if there were populations in Britain at that time that were substantially more EEF than the sample we have, either because of differential gene flow over the centuries into the east coast that made them different, or because this sample also came from a relatively more recent migration, like perhaps the Belgae.

[Angela]

The high African scores for the Iron Age Celt could be explained away by saying that it is a low coverage genome. However, the Anglo Saxon also scores some significant African. Any one of your explanations A,B, or C could be the answer. Also, through constant recombination, minority lineages can be lost.



As to your speculations about the Romans contributing to the increase in West Asian and SWAsian in modern British people, I’m not so sure.

I did mention in the prior post that there has been a decided increase in the Atlantic Med component when comparing the ancient Briton and the modern people of Britain in the K=12b analysis. I don't know how to explain it. L21 and U-106, which account for so much of Britain's y dna, are not "Roman" or even particularly "southern". I don't know if the J2, some G2a and the U-152 and perhaps a few upstream clades would be enough to explain it. The Normans may be a better bet, or this particular Iron Age Celt just may not be representative of the entire Pre-Roman Iron Age population of the British Isles.


For the sake of completion, since that run did not include a Belgian sample, here are the results for the French:

North European 36.5
Atlantic Med 44.4
Gedrosia 7.9
Caucasus 7.9
North West African 2

The results are pretty similar to those of the Iron Age Celt, except for the fact that the French have 12 points more Atlantic Med, very close to the figure for the modern Cornish.

As to the Anglo-Saxon, I think your placement of him is spot on. However, I think the explanation need not rest with the Huns. They were just as likely, in my opinion, to just be from an area that was higher in ANE, and had all those affinities.

Edited:
I tend to agree with your opinion of PCA plots. They are only very rough tools. Of course, in combination with other tools, they can be very helpful.

As for the least squared method, it can be informative when properly done. It all depends on the quality of the underlying admixture program. In the cited examples, the results, in some instances, are pretty good, while in others, because of the poor quality of the admixture run, you can get a situation where almost identical genetic distances to the same population can be produced when in fact the two ancient samples have significant differences.

[motzart]

If you look at the k12b admixtures I did in the other thread for our other ancient samples, Loschbour/Labrana/ and Malta boy all have a trace of Sub Saharan (<0.2%). Motala and the Stuttgart(the LBK farmer) have none.

If I were to make a guess as to the origin of it in these samples I would sale a mid neolithic arrival of E1b in Europe.

[Angela]

For some comparisons on the K-12 run:
The Iron Age Celt-I removed everything under 1% and the African numbers just so we can get a clearer picture.
dv3

• 48.72% West_European
• 23.25% Mediterranean
• 9.04% East_European
• 9.00% West_Asian
• 1.31% Northwest_African

The French:

West Euro: 52.1
East Euro: 3.9
Med: 33.9
NW African .4
West Asian 7.2
It’s a decent fit except for the fact that the French have 10 points more Med.

The Germans, except for their much higher East Euro, might be even closer.

West Euro 52.7
East Euro 16.9
Med 22
West Asian 6.5
S.W Asian 1.5

So, some of those FST numbers, the ones based on good admixture calculators, may not be so bad. An ancestral population close to the Germans if you take out some East European, and close to the French if you take out rather more Med might be on target.


That still doesn’t tell us if this ancient Brit was representative of all Britons of his time.


In that regard, these are the numbers for the modern British in this run-there isn’t an English population.

West Euro: 65.4
East Euro 2.6
Med 22.8

West Asian 6.7
SW Asian 1.4

The Med is the same for the Iron Age Celt, the modern Germans and the modern English. The West Asian for the Iron Age Celt is 9%, while the combined West Asian and SW Asian for the modern Germans and the modern British is about 8%.


So, upon further reflection, I am doubtful that there was any additional influx of specifically Mediterranean or SW Asian genes into Britain after this period.


The biggest change or difference is that the Iron Age Celt, the French and the Germans on this run are all about 50% “West European”.. The ancient Anglo-Saxon was even less West European, at 41.46%. The modern English are 65.4 West European.


It seems to me that somewhere in the British Isles there was a heavily “West European” population, and that our ancient Iron Age Celt from East Anglia is not totally representative of that group.

Ed. The fact that there is an increase in the Atlanto Med component in the modern British in the K-12b run may be related. It is an increase in the "Atlanto" portion of that component, not the "Med" portion of that component.

[motzart]

Motzart says: October 11, 2014 at 4:59 am
The Eurogenes blog is saying he is R1b L11+, but you are saying that he is R1b-, are we going to get another analysis of the Y DNA sometime soon?
Reply
genetiker says:
October 11, 2014 at 5:53 am
Wesolowski also says that ERS389795 is Iron Age. Wrong.
And he says that ERS389795 is the high coverage genome. Also wrong.
Wesolowski got the idea that ERS389795 is R1b-L11 from Anthrogenica. The Y-SNP calls that they used are wrong. For example, the file they used shows R1a1-PF6234 as being positive. PF6234 is at the position 2657176, and ERS389795 had the ancestral C allele at position 2657176.
I’m not saying that I know that ERS389795 wasn’t R1b. I’m inferring that he wasn’t R1b, because he only had 0.08% of the Gedrosia component, and the Gedrosia component is strongly associated with R1b.
There’s no need to do another Y-DNA analysis. The calls are what they are.

@Sile

I'm editing my post, I cross referenced the "all Y SNPs" file against the ISOGG SNP index and also saw what you are talking about for the L 151+, I think the Anglo Saxon is definitely L151

[polako]

I don't like PCA plots. That's an old, outdated method of comparison that only uses two dimensions and is therefore unreliable. I prefer to compare each admixture one by one.

The PCA results are backed up with f3 formal stats and UNBIASED admixture results.

The Dodecad results you used for your write up are biased. They're biased because of the calculator effect.

You can read about it here. Please try and understand it, because if you do, you'll be able to analyze these sorts of samples more accurately in the future.

http://bga101.blogspot.com.au/2012/0...or-effect.html

[polako]

For some comparisons on the K-12 run:
The Iron Age Celt-I removed everything under 1% and the African numbers just so we can get a clearer picture.
dv3

• 48.72% West_European
• 23.25% Mediterranean
• 9.04% East_European
• 9.00% West_Asian
• 1.31% Northwest_African

The French:

West Euro: 52.1
East Euro: 3.9
Med: 33.9
NW African .4
West Asian 7.2
It’s a decent fit except for the fact that the French have 10 points more Med.

The Germans, except for their much higher East Euro, might be even closer.

West Euro 52.7
East Euro 16.9
Med 22
West Asian 6.5
S.W Asian 1.5

So, some of those FST numbers, the ones based on good admixture calculators, may not be so bad. An ancestral population close to the Germans if you take out some East European, and close to the French if you take out rather more Med might be on target.


That still doesn’t tell us if this ancient Brit was representative of all Britons of his time.


In that regard, these are the numbers for the modern British in this run-there isn’t an English population.

West Euro: 65.4
East Euro 2.6
Med 22.8

West Asian 6.7
SW Asian 1.4

The Med is the same for the Iron Age Celt, the modern Germans and the modern English. The West Asian for the Iron Age Celt is 9%, while the combined West Asian and SW Asian for the modern Germans and the modern British is about 8%.


So, upon further reflection, I am doubtful that there was any additional influx of specifically Mediterranean or SW Asian genes into Britain after this period.


The biggest change or difference is that the Iron Age Celt, the French and the Germans on this run are all about 50% “West European”.. The ancient Anglo-Saxon was even less West European, at 41.46%. The modern English are 65.4 West European.


It seems to me that somewhere in the British Isles there was a heavily “West European” population, and that our ancient Iron Age Celt from East Anglia is not totally representative of that group.

Ed. The fact that there is an increase in the Atlanto Med component in the modern British in the K-12b run may be related. It is an increase in the "Atlanto" portion of that component, not the "Med" portion of that component.

Here's the problem; the ancient genomes aren't Dodecad samples, so you're comparing apples to oranges.

Please try and understand this. It's really not very complicated. This is the basic rule:

Unless samples are tested under exactly the same conditions, then their results can't be directly compared.

[motzart]

Just a follow up post to my post above.

Genetiker's comment above states that the SNP's provided by Felix's analysis are incorrect because PF6234 shows up as positive.

If you actually go into the data and check the allele and cross reference that against the ISOGG Y DNA tree you will see that it says that the call is positive, but the listing for the Allele is Ancestral. I saw this with P126 as well.

For example P126 for ERS389795 has a C at its position and is listed as positive, but checking it against the ISOGG Y DNA tree shows that C is the ancestral position and G is derived.


The positive markers for R,R1,R1b, ect to R1b-L151 DO contain the correct allele in the data when they are marked as positive. (and there are many of them).

Anglo Saxon DNA = R1b1a2a1a L151

[joeyc]

Using Dodecad is useless because of the calculator effect.

On the Eurogenes they are both closest to modern Orcadians but with less West Med, East Med, West Asian and Red Sea than modern Britons.

[joeyc]

I've downloaded the excell spreadsheets with the Irisplex/Hirisplex macros. I will try to include these nacient genomes to see what it comes out.

[ElHorsto]

The biggest change or difference is that the Iron Age Celt, the French and the Germans on this run are all about 50% “West European”.. The ancient Anglo-Saxon was even less West European, at 41.46%. The modern English are 65.4 West European.


It seems to me that somewhere in the British Isles there was a heavily “West European” population, and that our ancient Iron Age Celt from East Anglia is not totally representative of that group.

If "West European" and "East European" are defined by contemporary people and assuming that both categories are essentially descendants of the same old WHG, then it is impossible that there once was a heavily ancient "West European" population. The older a sample, the more evenenly his WHG will be divided in "West European" and "East European", because WE and EE are the result of recent differentiation due to geographic separation. The iron age sample is closer to Loschbour in terms of time scale and Loschbour is about equally East and West. Therefore I think it is generally impossible to find any ancient sample with such high "West European" percentage like contemporary west Europeans.

EDIT: Maybe the same reasoning can also explain the general tendency towards exotic admixtures in ancient samples (which is incomplete differentiation)!?

[Angela]

Using Dodecad is useless because of the calculator effect.

On the Eurogenes they are both closest to modern Orcadians but with less West Med, East Med, West Asian and Red Sea than modern Britons.

Well, that's certainly a model of sophisticated analysis. Two samples, with significantly different percentages even on the seriously flawed Eurogenes runs, are both closest to modern Orcadians. Oh, and how does such a conclusion illuminate in any way the source or migration path of these two ancient groups? That is supposed to be the whole point, isn't it?

Wait a minute, what happened to the statement by the Eurogenes blogger that one of the samples had a perhaps West Germanic signature, and the other a northeastern Germanic, even Scandinavian signature? You know, that statement which is rather similar to my own conclusions? Isn't it still on that thread?

Don't bother answering. I will do better about not reading new posts without signing in, so my ignore block will be working.

[Fire Haired14]

If "West European" and "East European" are defined by contemporary people and assuming that both categories are essentially descendants of the same old WHG, then it is impossible that there once was a heavily ancient "West European" population. The older a sample, the more evenenly his WHG will be divided in "West European" and "East European", because WE and EE are the result of recent differentiation due to geographic separation. The iron age sample is closer to Loschbour in terms of time scale and Loschbour is about equally East and West. Therefore I think it is generally impossible to find any ancient sample with such high "West European" percentage like contemporary west Europeans.

EDIT: Maybe the same reasoning can also explain the general tendency towards exotic admixtures in ancient samples (which is incomplete differentiation)!?

Why there is a west-east differentiation in admixtures is up to debate. It certainly isn't as simple as geography, and recent splits. Don't take admixture results to literally. Components aren't ancient populations they're just just clusters. If I was made into a component compared to east Asian components, you would probably score 100% in it, that doesn't mean I'm and ancient population and you're 100% me. Ancient samples of good quality don't score in exotic components. Have you seen updated Eurogenes K13-15 results for Loschbour, Stuttgart, Motala-12, La Brana-1, and Otzi? They literally score 0 in just about every non-west Eurasian components. The hunter gatherers don't score 0 in Mediterranean and near eastern components. Don't take HGs scores in regional clusters too seriously either, but do take note of them.

[Angela]

If "West European" and "East European" are defined by contemporary people and assuming that both categories are essentially descendants of the same old WHG, then it is impossible that there once was a heavily ancient "West European" population. The older a sample, the more evenenly his WHG will be divided in "West European" and "East European", because WE and EE are the result of recent differentiation due to geographic separation. The iron age sample is closer to Loschbour in terms of time scale and Loschbour is about equally East and West. Therefore I think it is generally impossible to find any ancient sample with such high "West European" percentage like contemporary west Europeans.

EDIT: Maybe the same reasoning can also explain the general tendency towards exotic admixtures in ancient samples (which is incomplete differentiation)!?

I don't have the time right now to give your points the attention required, as I'm about to run out the door, so I may revisit this later. I just didn't want you to think I'm not going to respond to your post.

So, a quick note. I'm not sure that I would assume that "East" and "West European" are any more "pure" components than any of the others in admixture analyses, and that therefore they represent "pure" WHG groups.

I know you're aware of those 2012 threads on the Dienekes site which examined the components in terms of one another. I think we've even discussed them?

http://dienekes.blogspot.com/2012/08...-k12b-and.html

http://dienekes.blogspot.com/2012/09...decad-k7b.html

Southwest Asian is Caucasus
Atlantic Med is Caucasus and some North European
North Euro is Atlantic Med plus Gedrosia and a slice of North European
Atlantic Baltic on World 9 is not a "pure" WHG component either

These are geographical groupings and nothing more, not "pure" ancestral populations. Their only usefulness is in drawing comparisons between different populations and trying to infer clues from that about ancient populations.

So, in the case of my example, all that I meant was that in order to reach the "West European" levels of the modern English, given the lower levels in both the ancient samples, one hypothesis might be that there was another, perhaps geographically more western population, which, had we an ancient sample, would score higher in "West European".

[LeBrok]

I

These are geographical groupings and nothing more, not "pure" ancestral populations. Their only usefulness is in drawing comparisons between different populations and trying to infer clues from that about ancient populations.

This should be reminded to all from time to time. I always look at this as statistical averaging and classification of genome and population comparison-pooling (or was it pulling?)

So, in the case of my example, all that I meant was that in order to reach the "West European" levels of the modern English, given the lower levels in both the ancient samples, one hypothesis might be that there was another, perhaps geographically more western population, which, had we an ancient sample, would score higher in "West European".

We also have to remember that populations in the past were not nicely mixed as of today. People were not very mobile, more insulated, more traditional, more inbred.
East Anglia is special place, it is one of first places to see more newcomers from continental Europe (celtic or germanic influence in this case) than the rest of British Isle. I'm sure we are going to find more populations of EEF Neolithic descendants in rest of Britain, who gave more Mediterranean, Atlantic and less East Euro admixture to modern English demographics.

[ElHorsto]

@Angela,@FireHaired

No, that's not what I meant. To the contrary.
What I meant: There can be no ancient population which scores more "West-Euro" than modern Britons, because:
1. "East" and "West" are recent and non-"pure" categories
2. WHG is already the largest admixture in both
3. WHG (Loschbourg) scores equally in both, "East" and "West", so WHG can not be unique to either West or East, but at best must have drifted over time into "West" and "East", if at all.
4. If WHG differentiation is not responsible for most "East" and "West" differentiation, then it must be due to the minor admixtures. But they are too minor in order to increase "West-Euro" percentage up-to or above today Britain's. Hence I conclude that "East" and "West" separation is based on post-Loschbourg events. And if I recall correctly, north-western-like (f.i. "North-Sea") and north-eastern-like (f.i. Baltic, East-Euro, ...) are often very close in terms of FST distances (but I don't have them at hand to be sure right now).

So, if you Angela believe there once was a population scoring even higher in "West-Euro" than today West-Europeans, then I doubt this hypothesis. It is possible that one minor admixture is responsible for the emergence of a "West-Euro" cluster, but this admixture alone would never score higher in "West-Euro", but would be something different like Gedrosia+Atlantic_med (particular EEF+ANE-variant for instance) or something else.

[Maciamo]

http://www.fc.id.au/2014/10/hinxton-2-analysis.html

for ERS389796



k23b seems to be the new "better" admixture test to do

Really ? 1% of African Pygmy, 0.8% of Amerindian, 0.7% of Melano-Polynesian, and you think that's reliable ?