Totalt antall sidevisninger

onsdag 14. august 2013

Gok4 and modern European variation

The Gok4 individual is as the earlier analysed Ste7 individual in the previous posts belonging to the Funnelbeaker culture finding in Gökhem, Sweden and the first farmers in Fennoscandia. Gok4 belonged to mtDNA haplogroup H. However the autosomal DNA coverage is for this ancient individual far better than for Ste7. I managed this time to match 294k SNPs vs the 1000 genome autosomal panel that I used to impute the missing markers for a standard selection of project participants and reference populations using phasing and imputation software Beagle. These was further LD pruned down to 76k in Plink to get a set of unlinked SNPs. The data was then run in Chromopainter-Finestructure pipeline unlinked mode using a world panel and then the European panel was extracted from the outpul files giving around 63k remainng SNPs for analysis.

As the results belows shows Gok4 appears to resemble quite much the earlier Ste7 individual. The strongest affiliation is with the Sardinians, then to Italians, Iberians, Albanians, Croat and Spanish. We also see an increased affiliation with populations in Central and Northern Europe like British, Scandinavians and Austrians. The most distant populations appears to be Saamis, Mordovians and Finns.

This appear to suggest some continuity between the first farmers of Fennoscandia and the modern Scandinavians. This was not seen in the earlier Ste7 analysis but may be due to the low number of markers (6k SNPs) giving high margin of error and possible partly different ancestry in Ste7 (possible Saami like or La Braña like ancestry). This analysis however has 63k SNP's used in the analysis comparable to the earlier La Braña of Northern Spain using 68k.

CC Euro Unlinked 63k

CC Euro Unlinked 63k detailed

CC Euro PCA Unlinked 63k (haploid individuals)

EDIT 23/9-13

2 kommentarer:

  1. Could you explain the A/B thing, please? For example, why are FI19_A and FI19_B so far apart?

  2. The data was made haploid to be able to be properly compared to Gok4 who is also "haploid" so _A and _B represent each side of your autosomal genome. I see now I should have merged these haploid PCA plot into a diploid plot taking away the confusion.