August 6, 2011

More Swedish autosomal genetics

There is a new free access paper on Swedish genetic structure:


Swedes are not too heterogeneous so this exercise is a bit of applying a microscope. Yet I'd love to see much lower resolution analysis of comparable quality on many other populations of Europe or elsewhere. 

Interestingly, the authors took some pains to, first of all, prevent that Finnish ancestry may have distorted the results because:

It may be argued that Finnish ancestry may be responsible for a large portion of the genetic stratification detectable in a Swedish sample but we chose to exclude them as the studies we combined had different inclusion criteria with respect to foreign ancestry. 
Fig S1 - shows the 'uncut' Swedish PC and how the cut was done

 The result of this surgery of sorts is this PCA:

Fig S3 - PCs by county

It is very apparent that the Norther Swedish lands, previously detected already as 'different' define PC1, while the isolated mountain county of Dalarna defines PC2. 

The impact of these two PCs in the Swedish sample is mapped in fig. 2:

Fig. 2 - color scale from dark brown to yellow

It is interesting to notice that the extremes of these two PC polarities may be caused more by their own genetic isolation than anything else, as it's apparent in fig. 5:

Fig. 5 - Homozygosity (base is Stockholm county)

Both Finnish and Norwegian samples applied to these two PCs produced negative values (supp. materials), what reinforces this idea that they are a creation of genetic isolation in Sweden and not external influences.

Further PCs (up to PC10) are mapped in fig S7 but I do not see them as very informative, excepting maybe PC6, which shows a pole near the Southern Norwegian border.

See also in this blog: Swedish autosomal genetics.

2 comments:

  1. My interpretation:

    The exclusion line should have been chosen perpendicular to the apparent Finnish principal component. But it wasn't. As such, I believe that PC1 of the "cleaned up" version still simply shows Finnish/Sami contribution. PC2 shows higher indigenous make-up (relating to parts of Norway, negative correlations notwithstanding), and the mass-center population, mostly south, simply have the highest contribution from (Northern German) agriculturalists.

    Surely there is some isolation reflected here, but in my opinion it ts rather unimportant.

    ReplyDelete
  2. Hmmmm... Northern Swedish clustered differently than Finnish since K=2 in the previous study by Salmela: they clustered with NW Europeans in K=2, Swedes in K=3 and separate from other Swedes (but also from Finns) in K=4.

    There is definitely something different about Northern Swedes, but it's not mere Finnish ancestry.

    Look also at the supplementary materials. Figs. S4 and S5 note that Finns and Swedes of Finnish ancestry display low values in the "cut" PC2, which would be, if any, the equivalent to the "uncut" PC1 (see fig. S10).

    So the Dalarna people (from which the anthropometry term "Dalic") and those from the surrounding mountain areas are also a somewhat distinct group, even if not as much marked as the Northern Swedes.

    It makes sense: the deeper you look the easier to grasp these isolate "inbred" populations as distinct. It has not been done with humans but it has been tried with cows and you can reach incredible levels of K-depth and still find more and more distinct subclusters, assuming some genetic isolation, as happens with cattle breeds.

    ReplyDelete

Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).