My Blog List

Monday, May 16, 2011

The hierarchy of clusters

Hierarchical layouts of clusters


  1. Isn't it very strange that the Finns are set so far apart, even more than Chuvash and not even clustering with them? I noticed the same issue in the previous post, the PCA one.

    In other research I have seen the Finnish samples do not behave so extremely. I wonder if your sample is from some rare highly inbred group. Often highly isolated populations behave that way: Hadza in Africa, Khalash and Mlabri in Asia, etc. But most Finns are not so isolated.

  2. Yes, this is indeed very strange situation, because my Finnish sample is from the same source, which others used in their research. Even more strange that Finns in my sampled, being detected as complete outliers in PLINK analysis, have very low level of homozygosity. And this observation completely contradicts to the high distribution of ROH (runs of homozy-gosity), revealed by the analysis of contiguous homozygosity (runs of homozygous loci). Since the distribution of ROH records information about this history of intermarriage or isolation and consanguinity among ancestors, we should expect high levels of homozygosity in Finns. This is very strange.
    On other hand, for PCA/MCD calculations we are using a pruned dataset of genotypes(removing low quality SNPs and SNPS in linkage disequilibrium) where all SNPs in approximate linkage equilibrium with each other. This approach helps to infer the stratification of populations. The (relative) isolution of Finns in comparasion to other European populations is best illustrated by the IBD segment sharing. In one of more previous posting, i presented IBD segment sharing in a form of Igraph (where each vertice represents a particular IBD shared segments). The IBD segment sharing in most pouplations from my dataset form very complex reticulate networks, which consist of the IBD sharing links, while the visualization of IBD segment sharing in Finns produces a "quasi-genealogical" structure.For example, there are two "IBD communities" in Finnish "population sample" in our project (black circles on the graph). These two enigmatic "communities" appear to be "things in themselves" being focused on two sharing "midpoints" in Finnish sample (V151;V156), while V151 and V156 appear to be only Finns linked by means of IBD segment sharing to North_Russians, and, to lesser extent, Balto-Slavic cluster.