Getting The Ld Data For The 1000 Genomes

Question

[Michael]

Hi folks,

In another question, the OP mentioned that he was able to find some preliminary LD data for the 1000 Genomes project. I was unable to find the file that was mentioned there, but I got the impression, that LD data for the project already exists. Where is it, can you help me?

[Ryan D]

The 1000G files from the website are best approached with something that keeps the phasing information in-tact--one of the benefits of 1000G data. For that, haploxt is recommended: http://genome.sph.umich.edu/wiki/Haploxt It can be used on the files downloaded from here.

If you want to do some PLINK calculation, I can try to help by offering a script for extraction of such data from the phased file. Working on that at the moment.

[zx8754]

It has been almost 3 years since this question was asked, did you manage to find LD data?

[Michael]

I calculated it myself from the phase 1 vcf, using this: A: 1000 genomes LD calculation

Answer 1

[Jorge Amigo]

As far as I know, there is no such information currently available to the public. at least on the official ftp site (you can always check the whole up-to-date site tree here).

Ryan D commented this on the "1000 genomes LD calculation" question mentioned above:

Some data exists on this LD. I ended up pulling data from files with code like this: zcat ~/1000Genomes/2010-06/CEU/LD/xt/chr19.xt.gz | grep -w rs11671664 | awk '$4 > 0.5' rs11670375 rs11671664 0.8961 0.5673 A,G chr19:50848886 rs11671664 0.8990 0.7340 C,G rs11083777 rs11671664 0.8990 0.7340 G,G chr19:50851826 rs11671664 0.8899 0.7134 C,G chr19:50852809 rs11671664 0.8990 0.7340 A,G chr19:50853145 rs11671664 0.8990 0.7340 G,G rs4375772 rs11671664 0.8899 0.7134 C,G rs11671664 chr19:50865055 1.0000 0.6138 G,G

but unfortunately I haven't been able to find that data on the 1000 Genomes FTP server. it would definitely be very interesting to know where exactly is this data available.

[Michael]

yep, exactly, I was desperately looking for this

Answer 2

[Mary]

Just bumping this with a new tidbit: the 1000 Genomes paper is out today at Nature

A map of human genome variation from population-scale sequencing

I'm just starting it, but it may have some guidance on things people are interested in.

Answer 3

[Ryan D]

The 1000G files from the website are best approached with something that keeps the phasing information in-tact--one of the benefits of 1000G data. For that, haploxt is recommended: genome.sph.umich.edu/wiki/Haploxt It can be used on the files downloaded from here: sph.umich.edu/csg/abecasis/MACH/download/… If you want to do some PLINK calculation, I can try to help by offering a script for extraction of such data from the phased file. Working on that at the moment.

[Ryan D]

It looks like SNAP has updated with the 1000 Genomes pilot 1 data. So if you don't mind being a bit out-of-date, you can use that for quick calculations: http://www.broadinstitute.org/mpg/snap/ldsearch.php

Under SNP data set choose "1000 Genomes pilot 1".

Now we're using a ruby script which pulls the region of interest from the 1000G data and calculates LD within PLINK. Sub-optimal, but better than pilot 1 data.