Eurogenes v2

[Polako]

Those are publicly available individuals from the HGDP set.

By the way, I just did the same test for the extended dataset. PCA looks alright still:



Blow-up of British swarm:



We're ready to rock and roll now. Bring on the lists of populations and/or individuals you want me to run through. It's quick and easy, believe me.

It reminds me of the ChromosomePainter PCA, so it definitely looks correct.

But the extreme IBD sharing between groups in Northeastern Europe makes it difficult to see how much IBD all the groups really share.

I think bubbles of average sharing on a map would be better. Loxias could probably make the maps, if he had the data.

[Day Tripper]

IBD for Horners, as requested by Lol_Race:

Z-scores
PCA Coordinates

PCA overview:



I had to remove TIGRAY10, as he was way off on his own for whatever reason.

[Huckleberry Finn]

Does anyone know what is the definition of the russian Vologda group, for instance in terms of geographics? Vologda, of course, but how about places such as Pskov and Archangel?

[soulblighter]

I have noticed that genetic maps in IBD6-11_SNPs.text and IBD12-22_SNPs.text are very accurate, i.e. positions of each SNP as given in column D in bp is exactly as given by
http://integrin.ucd.ie/cgi-bin/rs2cm.cgi
I have checked few thousands of SNPs

Map in IBD1-5_SNPs.text was slightly off but still accurate enough to estimate the segments length in cM.

So all the data for the script is available in the files, we do not have to query in outside genetics maps.

For those who would like to investigate their data in the mean time short guide what is to be done:

All data in extended IBD files have in column E FastIBD scores less than 1.0E-10 so this score is sufficient for IBD and should not concern us any more. The length in cM is of greatest interest. We are searching for segments longer than 1cM as only those are IBD with high probability.
How to calculate length in cM?
For example in IBD6-11_SNPs.text I select firs row:

A	B		C	D	E
NA21436	GSM536527	10	45	6,08E-11

(What NA21436 and GSM536527 are can be fund in IBD6-11_samples.text.)

10 in column C marks the beginning of the segment
45 in column D marks the end of the segmnt.
It means that we should look for beginning SNP in line 10 of IBD6-11_SNPs.text and for the end SNP in line 45 of IBD6-11_SNPs.text.

Line	Chrom	SNP		?	Distance
10	6	rs12212839	0	173345
45	6	rs2473492	0	511673

From column Distance we get beginning and end of the segment in bp.
One centimorgan corresponds to about 1 million base pairs in humans on average.
The length in cM will be (511673-173345)/1000000 =0,338328 cM.
It means the segment is too short for IBD <1 cM.

Are you sure the line values in the SNP files are not the same start values you would enter here in gedmatch?: http://gedmatch.com/GD_calc1_old2.php . For your examples, the length comes out as 1.22cM

[warwick]

What analytical tools are we using for the results text file. I've been busy with grad school and away from the boards. Thanks

Alternatively, I'm willing just to email the file to someone who's willing to do it. Pretty busy here.

[EastPole]

Are you sure the line values in the SNP files are not the same start values you would enter here in gedmatch?: http://gedmatch.com/GD_calc1_old2.php . For your examples, the length comes out as 1.22cM

I am not familiar with http://gedmatch.com/GD_calc1_old2.php.
For female it gives 0.36 cM, for male 2.08 cM, average 1.22 cM.
Isn’t it safer to assume smaller values?

Someone with more experience should verify this and tell us what to do.

[Jonny]

So what are the upcoming plans for Eurogenes?

[linkus]

Does anyone know what is the definition of the russian Vologda group, for instance in terms of geographics? Vologda, of course, but how about places such as Pskov and Archangel?

I don't think that any people from Pskov could have been included in the Vologda sample. RU40 is from Pskov and he clusters right beside Balts whereas none of the Vologdans do.

We're ready to rock and roll now. Bring on the lists of populations and/or individuals you want me to run through. It's quick and easy, believe me.

Could you please make one with:

Lithuanians
Latvian
Estonians
Chuvash
Mordvinian Erzya
Mordvinian Moksha
North Ossetians
Georgians

[Red Rover]

What analytical tools are we using for the results text file. I've been busy with grad school and away from the boards. Thanks

Alternatively, I'm willing just to email the file to someone who's willing to do it. Pretty busy here.

In Excel sorting the last column from smallest to largest.

[Loxias]

My requests to DT :
Can you email me the links from Polako.
And make Basque, France, British Isles, Romanians, Hungarians, Poles and all Caucasus?