Genetic Distance Tables for Europeans

[Neon Knight]

I found this:



Are there any more?

[Grace O'Malley]

This one is not exactly a genetic distance table but still interesting.

[Neon Knight]

Do you know the source for the top one? And what does the bottom one demonstrate?

[Neon Knight]

I did some proximity lists from the table I posted:

British
Belgian 0.0005
French / German 0.0006
Swedish 0.0013
Norwegian / Czech 0.0014
Hungarian 0.0016
Slovak 0.0017
Polish 0.0017
Spanish 0.0024
Romanian 0.0028
Russian 0.0032

Norwegians
Swedish 0.0010
British 0.0014
German 0.0016
Belgian 0.0019
French / Czech 0.0024
Slovak / Hungarian 0.0026
Polish 0.0034
Russian 0.0036
Romanian 0.044
Spanish 0.0047

Swedes
Norwegian 0.0010
German 0.0011
British 0.0013
Czech 0.0016
Belgian 0.0018
Slovak 0.0019
Hungarian 0.0020
Polish / French 0.0023
Russian 0.0025
Spanish 0.0047

Germans
Czech 0.0003
Hungarian 0.0004
Slovak / Belgian 0.0005
British 0.0006
French 0.0008
Swedish 0.0011
Polish 0.0012
Norwegian / Russian/ Romanian 0.0016
Spanish 0.0025

French
Belgian 0.0002
British 0.0006
Spanish / German 0.0008
Hungarian 0.0015
Czech 0.0016
Slovak / Romanian 0.0017
Swedish 0.0023
Norwegian 0.0024
Polish 0.0032
Russian 0.0037


Is it correct to say that genetic distance (fst) gives the most accurate picture of genetic relationships?

[PunhetaDeBacalhau]

This one is not exactly a genetic distance table but still interesting.

GEDMatch's calculators put Portugal slightly more close genetically to Northwestern Europeans than to the Balkans, and that table puts us closer to the Balkans than to Northwestern Europeans. One of them has to be wrong... Which is it?

[Petalpusher]

Is it correct to say that genetic distance (fst) gives the most accurate picture of genetic relationships?

Fst are not exactly the same than distances as we imagine them in Oracle for example. Fst are more a value of shared genetic or differentiation which is different than an absolute distance. For example if i'm half Austrian half Polish i ll have a high proportion of shared genetic with both population but not with the population i end up the most similar to in distance, let's say randomly Czech. Fst describes a relationship that can be recent or very old, distances more a present similarity, in the end it it's still kind of the same result but both cannot be compared. Fst would vary more at individual level depending on regions than distances in a population. Another typical example would be Ashkenazi and S.Italians, different fst but very similar distances.

[Neon Knight]

Fst are not exactly the same than distances as we imagine them in Oracle for example. Fst are more a value of shared genetic or differentiation which is different than an absolute distance. For example if i'm half Austrian half Polish i ll have a high proportion of shared genetic with both population but not with the population i end up the most similar to in distance, let's say randomly Czech. Fst describes a relationship that can be recent or very old, distances more a present similarity, in the end it it's still kind of the same result but both cannot be compared. Fst would vary more at individual level depending on regions than distances in a population. Another typical example would be Ashkenazi and S.Italians, different fst but very similar distances.

I think I see what you mean. The Swede is red, the German is purple and the Frenchman is blue. But the woman who is half Swedish & half French is not German purple but is a jumble of red and blue like a chessboard pattern; she averages out as a German on a genetic map but does not fit into the shoes of a German - rather, one foot fits a Swedish shoe and her other foot fits a French shoe (to simplify). Right?

I had been led to believe that Fst was genetic distance (GD). So Fst, the Oracle score and GD are three different ways of measuring genetic similarity? Which measure is the most meaningful for showing us the genetic relationships between modern European populations?

[Harkonnen]

Orcadians and Sardinians descent from small founding populations, added with subsequent isolation, drift adds to their fst distances. Orcadians, a subset of British genepool can not be more distant from the Irish than what Swedes are. You'd probably see this with Dstat. Likewise it is very unlikely that Greeks are more closely related to Swedes than they are to Sardinians. Also Sardinians get very similar fst scores with the Irish to the Druze/Bedouin posse. Again genes gone bad effect. Ladies like bad boys. That's why.

[Petalpusher]

I think I see what you mean. The Swede is red, the German is purple and the Frenchman is blue. But the woman who is half Swedish & half French is not German purple but is a jumble of red and blue like a chessboard pattern; she averages out as a German on a genetic map but does not fit into the shoes of a German - rather, one foot fits a Swedish shoe and her other foot fits a French shoe (to simplify). Right?

That's a good analogy i think (except she would average as English!) Another example, Welsh are certainly closer to a France avg in distances than the rest of UK even though they might have the least shared dna from there. Or to really push the analogy, a fellow countryman would be closer to you in distance than one of your cousin who could be half German, but you would still have more shared dna with your cousin than your neighbour.

I had been led to believe that Fst was genetic distance (GD). So Fst, the Oracle score and GD are three different ways of measuring genetic similarity? Which measure is the most meaningful for showing us the genetic relationships between modern European populations?

It's what i would rather call distance of relatedness, the whole postulat of fst is that migration counteracts drift which also create problems for certain isolated populations. We can see it as a distance for convenience or we do actually translate it in another form of distance through admix percentages, then in actual Oracles distances. Each component has a specific fst value to each of the population and other components, a calculator is simply assessing based on your genome how much % you share with this component and all its fst.

There is just many ways to end up exactly like X country without having any ancestry from this country even though you could have exactly the same proportions of genes than them, therefore still could have been made of the same thing one way or another. The last stage of Europe's creation is in practical terms a mix of late neo cultures with some Yamnaya-like ones, through Bell Beaker and the likes, specially in the west. I can imagine some extreme cases of south Americans, where their fst and regular distances would differ greatly but we can probably say the more "ethnically packed" your components are, the least distances and fst will diverge.

I see IBD like fst (since it is the consequence of IBD), as giving you a more intricate view, on how much was shared with a country in more or less recent times but it would be more meaningful for that matter to look at it by regions, or even at an individual level, which is usually the case in IBD. In a way this is also what triggers clines in Oracle. Do i get more often Italian+German or Spanish+English doesn't have much to do with my actual position (it could be exactly the same) it has to do with fst, because that individual is rather sharing more with one or the other pair. Personally i know i trigger always the same kind of axis (cause eastern France), im sure it's also the case for most people.

[Grace O'Malley]

GEDMatch's calculators put Portugal slightly more close genetically to Northwestern Europeans than to the Balkans, and that table puts us closer to the Balkans than to Northwestern Europeans. One of them has to be wrong... Which is it?

The first one is from Cavalli-Sforza so not likely to be incorrect. Gedmatch has many different calculator programs and while informative and fun is not like a genetic study or comparable to one. It is a good guide to what populations your genome is closest to but is run through a calculator program.

The second one is from this study on 5 European populations.

https://www.karger.com/Article/Pdf/313854

Here is a write up by Razib Khan on this study.

From that study this is interesting. This confirms previous findings that the largest component of variation in Europe is north-south (at least evaluating to the west of a particular geographical cutoff), with a secondary east-west dimension.

http://blogs.discovermagazine.com/gn.../#.V-YZP_B97IU