Plink/VCFTools pairwise LD is always 1

Question

[pmc.sa]

Hi everyone,

I've been trying to calculate pairwise LD from my multisample VCF file. I tried with both Plink and VCFTools but in both cases the output was always (or almost always r2=1).

VCFTools:

vcftools --gzvcf file.vcf.gz --geno-r2 --ld-window-bp 10000 --out file_ld_10kb
CHR     POS1    POS2    N_INDV  R^2
1       827     4036    2       1
1       827     4285    2       1
1       827     4940    2       1
1       827     4988    2       1
1       827     6815    2       1
1       827     8515    2       1
1       827     9393    2       1
1       827     9422    2       1
1       3982    6384    2       1

Plink:

plink --vcf ganda_chr1.vcf.gz --r2

CHR_A         BP_A SNP_A  CHR_B         BP_B SNP_B           R2
     1          827    .      1         4036    .            1
     1          827    .      1         4285    .            1
     1         3982    .      1         6384    .            1
     1         4036    .      1         4285    .            1
     1         4036    .      1         4940    .            1
     1         4036    .      1         4988    .            1
     1         4285    .      1         4940    .            1
     1         4285    .      1         4988    .            1
     1         4285    .      1         6815    .            1

Because the values are so strange to me, maybe I'm doing something wrong. Appreciate some input on this matter.

Thanks anyway :)

Pedro

Answer 1

[chrchang523]

It looks like you only have 2 samples. You need at least ~50 to compute decent LD stats.

[pmc.sa]

Thanks. I actually have 4 samples but your answer still stands.