East Finns, best fits

I have two East Finnish sample groups, one (EFinn) consisting of my project members from Eastern Finland, another (FIN-East) derived from the first one using samples from the 1000 genomes project.  The 1000 genomes gives gives better result quality due to sample amount and quality.  My 1000 genomes samples are grouped into 3 groups:  Southwest Finns resembling my project members from Finland Proper,  West Finns resembling Tavastians and East Finns resembling people from the Savo district.  Now I have a new group resembling medieval Estonians under abbreviation Estonian_MiddleAge_N1c1.  N1c1 comes from the allocation of researchers.  There is no reason to assume that medieval Estonians could have been genetically extremely diverse though.  The new Finnish group is named as  Estonian_med_Finns.

 

Some explanations.  Z-score values over 3 or below -3 are considered to be meaningful.  It doesn't mean that all other results are not "true", only that they are not mathematically reliable.  The reliability depends on SNP amount among other things ( ! mark). Good  Z-score doesn't tell about a superiority of the result itself, neither the margin of error, column 5. 

F3 uses only two sources and admixed results can be achieved by several combinations.  Real admixture could be achieved by some combination of these sources giving admixed results, or not, or even by a combination of fitting and non-fitting results.  Using qpAdm I could have found these combination, but qpAdm is a challenging tools to use with already admixed modern populations.