Updated Europe analysis

New updated Europe regional analysis. I have removed the Orcadians as they give odd clusterings in the PCA because of relatedness, but one British group who have in earlier analysis clustered with them work fine as stand-ins. In the regional analysis more substructure is revealed. Caucasus not included.

As we can see for the Fennoscandians much is as before, but the addition of more continental Europeans have revealed more structure. We see that the 2 new Germans have clustered the dispora individuals Ux who have at least partly German and English ancestry into one group close to one group of the British, also containing a few French and one British.

The 1 new Austrian appears to cluster closest with the Romanians. We see 3 of the 4 new Polish cluster with the Ukrainians and Belorussian cluster while the last cluster with the Lithuanians. Most of the new Russians appears to cluster together with the 2 Estonians already in the project. However note the Estonians appears to pull away from the Russian cluster on the main PCA D1-D2 plot.

CC Europe Aggregated

CC Euro Populations

CC Euro Raw

(PCA plots in order D1-Dx+)

Updated World Analysis

This time I reduced the samplesize for Basque, Sardinians, Italians, Bulargians and Spanish to one each because of capacity problems as I would like to limited the analysis process to two threads to reduce the work with adding and formatting input filer, however they didnt cluster to their own as I expected like with the more distant populations. At the same time I increased the number of individuals to geographic closer population to make sure sample size is large enough to give good phasing.

The new thing this time is that we have have new samples from Germany, Austria, Poland and Russia. In the world analysis as said before its more difficult to differentiate similar European populations but also small sample sizes play a role her as Spanish, Iberian, Italians and French Basque didnt seperate into seperate clusters as in the previous analysis. This analysis will be presented in the next post.

Note that Romanian7 and Romanian3 appears to be Rom people as they cluster partly with different tribes in Pakistan.

CC World Aggregated

 CC World Raw

CL World Aggregated

CL World Raw

Updated local Fennoscandian analysis

New individuals added. We see much the same structure as seen before. However we see two Finns who share recent Saami ancesty groups together. We can as before differentiate between different influences between the different groups. The main sources is from Scandinavians, Finns and what appear to be two different kind of Saamis, North-Saami group and a South-Saami group. The South-Saami seems to diivide Swedes with ancestry from Dalarna and northwards out of the main Scandinavian cluster in different grades. Also Swedes with partly Finnish ancestry pull towards Finns. Finns again have different grades of sharing with Scandinavians.

ChunkCounts Aggregated Populations

 ChunkCounts Aggregated PCA 

 ChunkCounts Aggregated Individuals 

 ChunkCounts Raw Individuals  

ChunkCounts PCA Individuls

 ChunkLenght Aggregated Populations 

  ChunkLenght Aggregated PCA  

  ChunkLenght Aggregated Individuals 

Oceania influence in East-Asians

This is offtopic regarding the project goals but its so interesting I have to make a post about it. It appeared already in my first analysis including East-Asians, Siberians and Native Americans that these had increased sharing with my Sub-Sahara African individuals than many Europeans and therefore shared branch. I suspected then it may have something to do with Oceania like Papuans and Melanasians. In the most recent Chromopainter/Finestructure run it appears to confirm this.

Here we see first of all that the closest relatives to the cluster of Pygmy, San and a Bantu from Kenya outside Africa are a Papuan and a Melanasian. Then secondly to a Cambodian and a Panyian. This influence is also to a weaker extent seen in a clustered groups from Pakistan through i India to Malaysia.

However we can also see that among that the East-Asians in particular the populations from or close to China all have the highest level of sharing with the Africans, then to a weaker extent Siberian groups and at weakest with the Native Americans. This pattern is also seen vs the Papuan and Melanasian but to much stronger extent.

This pattern probably show that this Oceanic must be the influence of ancient migration as its even shows among all the Native Americans and is also shared between San, Pygmys and Papuan and Melanasian.

Updated Chromopainter/Finestructure World analysis

Same settings as before but added more populations (or basically almost the rest of populations I have in my reference database) from Middle-East, Caucasus and all the way to Pacific/Oceania. To properly view the images open them and copy them to a image program like MS Paint or something similar or just download them directly from the source webpage where you will get them in full size.

All members should have received the image through email, this time I have put them into Excel windows that make it easier to view individual result. I have only added 1 individual from each of the more distant populations, but it seems even with only 1 individual the analysis software manage to cluster and tree them quite well. Notice also that in worldview some individuals and populations that is close to each other will merge into the most similar cluster.

So what new information can be seen by adding more populations?

We now got a range of main European clusters seen in ChunkLenght (most of these can be further divided):

1. Common Scandinavian-Orcadian-PartBritish cluster
2. Common Lithuanian-Belorus-Ukrainian-Mordovian-Russian cluster
3. Common Saami-Finnish cluster
4. Common Sardininan-PartIberian-Basque cluster
5. Common Hungarian-Romanian-French-PartBritish-PartIberian-Spanish-Italian cluster
6. Common Caucasus-Middle-East cluster

From the heatmaps its possible to infer the following:

1. Scandinavians appears to have more similarity vs Middle East and Caucasus cluster than Saamis and Finns. Saamis and Finns share more with the Chuvash.

2. Saamis appears to show considerable sharing vs a group of northern Siberians however this appears to be mostly attributable to only 1 Saami individual. The single Aleut in the panel is partly mixed with Europeans and therefore shows more genetic sharing vs Europeans.

Whole analysis heatmaps:

ChunkCounts Aggregated World 



ChunkCounts Aggregated Raw World

ChunkLenght Aggregated World



Looking again at Fennoscandia

This is a further local study of the Fennoscandian participants using the previous Chromopainter analysis data.

Overview

As we can see here the Finns and Saamis clearly deviate from each other in this "local" analysis. This is something we also have seen in earlier MDS analysis. In earlier analysis using wider population panels Saamis and Finns have shown a tendency to be close to each other as regional differences is emphasised more in such analysis. We also see the same in this analysis as the Saamis and Finns share the largest branch with each other opposite to Scandinavians. What appears to be East-Finns (Finn1) and North-Saamis appear to be most distant to Scandinavians.

South-Saamis and Scandinavians

The difference between this analysis and previous similar MDS analysis is that the software to larger degree is able to differentiate the different groups. The Saamis appears much the same as in previous analysis but what is interesting here is what appear to be South-Saami influence on Scandinavians especially Swedes and to some parts Finns too "from" SWE7 in particular.

SWE7 appears to have a special position. This indvidual appears to show strongest affiliation with the subgroup of Scandinavians in particular Swedes that appears to show South-Saami admixture and to much lesser degree to "normal" Scandinavians. The intermediate position of SWE7 between North-Saami and Scandianvians also seem to suggest SWE7 is some kind of North-Saami and Scandinavian admixture but this is not supported when comparing to NO6 and NO7 who appear to be straightforward Scandinavian-North-Saami admixted individuals. Actually SWE7 appears to be much less affiliated with the North-Saamis than NO6 and NO7. So SWE7's pull away from Scandinavians cant therefre be explained directly as direct North-Saami like admixture.

Software not able with current resolution to detect obvious Saami admixture in some Finns

However the software have not been able to detect obvious Saami influence on FI26 and FI5 among Finns that is very observable on the PCA plot as a strong pull toward the North-Saami individuals also seen in the CC raw heatmap. It may be due to the fact that Finns and Saamis are closer to each other than Scandinavians.

Finns

We also see that Finns splinter into 3 Finn only groups and 2 Finns joined into a mixed group. Finn1 group appears to be East-Finns, Finn2 appears to be West-Finns and Finn3 group appears to also be a West-Finn group. The third mixed group containing two Finns appears to be a group that appear to share branch with the Saamis and the South-Saamis influenced individuals in particular.

Heatmaps and PCA plots

ChunkCount Aggregated

ChunkCount Populations Aggregated

ChunkCount Raw Aggregated

ChunkCount PCA Populations Aggregated

ChunkCount PCA Populations Aggregated

(To view images, click and copy image to MS Paint or similar program.)

Extending reference populations - looking beyond Europe

I am now in a first stage to extend the reference populations to the Chromopainter analysis. In this first run I have increased the number of individuals from different populations, but because of constraints in how many individuals my computer system can handle I have only included 1 or 2 reference individual from the more distant populations from Asia, America and Africa. However as one can see from the heatmap even with only 1-2 individual these distant populations do cluster nicely and as expected. I have also in this run modified the color scales to concentrate on the lowest scales to get a clearer picture of the dissimilarities among the populations especially for the more closely related populations.

Whats do the addition of more distant reference populations give?

In earlier analysis using ADMIXTURE we have seen that Saamis appear to have the highest level Siberian like admixture. In the ChunkLenght analysis we appear to see the strongest sharing for all the Saamis with the Nganassans (Nenets). Note that this affiliation also appears among East-Finns, Vologda-Russians, Moldavians and Chuvash. A faint influence also appears among Scandinavians.

There also appears to be sharing with a group of Kets and Selkup but these individuals have also in earlier ADMIXTURE analysis shown to be about 50% Siberian clustering with over 20% European ancestry and considerable Central-Asian ancestry so thats why we also see increased affiliation with populations that do not affiliate with the Nganassans individual. This increased sharing is also reflected among Finns, Vologda Russians, Mordovians and Chuvash.

Among the Saamis SA2 appears to have an additional affiliation that the other Saamis do not have. Its the only Saami individual that show increased affiliation with a cluster containing Daur, Xibo and Mongolian. The Chuvash and a group of Adigey is the only other groups that shows a similar profile. SA1 and SA2 also appears to have an increased affiliation with the Koryak and Yukagir than the remaining Saamis.

We also see that Finns and especially East-Finns appears in general to have a increased affiliation to Siberian populations but less than both the Saamis and Vologda Russians and Moldovians.

Also notice that the Saamis appears to have increased assymetric relationship between receiptant and donor to the Siberian populations compared to the Chuvash appears to be much more symmetrical. This indicate that the Siberian like ancestry among the Chuvash must be more recent or similar to these Siberians than the Saamis relationship to the same populations. Also the higher color intensity suggest closer relationship to the Siberians for the Chuvash.

Heatmaps:

ChunkLenght Aggregated

.

ChunkLenght Raw

 

ChunkCounts Aggregated

 

ChunkCount Raw

 

Average SegmentSize Aggregated (ChunkLength divided by ChunkCounts)

Note here that the Africans share very small segmentsize with each other compared to any other group, This is because of higher genetic diversity. Segmentsize include both identical and related or mutated segments.