PLINK: Compute LD from list of SNPs using 1000 Genomes

Question

[8chocinco]

I have a list of SNPs. I'd like compute the r-squared values to remove any SNPs that are in linkage disequilibrium, similar to what SNAP does -- SNAP does not allow large lists of SNPs.

So, I downloaded the 1000 Genomes data from the LocusZoom website. Then I sought to use PLINK to compute the r-squared values by first combining files across chromosomes:

plink --bfile locuszoom/data/1000G/genotypes/2012-03/EUR/chr1 \
      --merge-list geno_list.txt \
      --make-bed \
      --out locuszoom/data/1000G/genotypes/2012-03/EUR/chr_all

Then I ran the LD command using my SNP list ("snp_list.txt"), but I get the following error:

PLINK v1.90b3.36 64-bit (31 Mar 2016)
https://www.cog-genomics.org/plink2 (C) 2005-2016 Shaun Purcell, Christopher Chang GNU General Public License v3 Logging to plink.log. Options in effect: --bfile locuszoom/data/1000G/genotypes/2012-03/EUR/chr_all --ld-snp-list snp_list.txt --r2

1033897 MB RAM detected; reserving 516948 MB for main workspace. 16504864 variants loaded from .bim file. 379 people (0 males, 0 females, 379 ambiguous) loaded from .fam. Ambiguous sex IDs written to plink.nosex. Using up to 127 threads (change this with --threads). Before main variant filters, 379 founders and 0 nonfounders present. Calculating allele frequencies... done. 16504864 variants and 379 people pass filters and QC. Note: No phenotypes present. Error: No valid variants specified by --ld-snp/--ld-snps/--ld-snp-list.

Similar error when I try just two SNPs:

PLINK v1.90b3.36 64-bit (31 Mar 2016)
https://www.cog-genomics.org/plink2 (C) 2005-2016 Shaun Purcell, Christopher Chang GNU General Public License v3 Logging to plink.log. Options in effect: --bfile locuszoom/data/1000G/genotypes/2012-03/EUR/chr_all --ld rs2840528 rs754594 --r2

1033897 MB RAM detected; reserving 516948 MB for main workspace. 16504864 variants loaded from .bim file. 379 people (0 males, 0 females, 379 ambiguous) loaded from .fam. Ambiguous sex IDs written to plink.nosex . Using up to 127 threads (change this with --threads). Before main variant filters, 379 founders and 0 nonfounders present. Calculating allele frequencies... done. 16504864 variants and 379 people pass filters and QC. Note: No phenotypes present. Error: --ld variant name not found.

Suggestions?

Answer 1

[Floris Brenk]

plink2 --bfile INPUT --ld SNP1 SNP2 --r2

works for me fine, so guess something wrong with your snps e.g. not in your map/bim file