How does plink 1.9 prune SNP files for LD

[Katrina Catalano]

Hello,

I'm confused about how plink prunes out loci in linkage disequilibrium with SNP data generated from ddRADseq. Do the contigs need to be mapped to a reference genome? I have a SNPs from a non-model organism, so there is no reference genome to map to. 

How does the program prune LD loci when there's no information on chromosomal position? My SNPs are called in VCF format.

I've tried to find this answer elsewhere, so any clarification is helpful.

Thanks,

Kat

[Christopher Chang]

plink's LD pruning functions always skip "chromosome 0".  However, variants which have been mapped to a contig are treated normally.

If you're using a variant-count window (e.g. "--indep-pairwise 500 50 0.5"), chromosomal position is irrelevant; only the order of SNPs in your file matters.  If you're using a kilobase window (e.g. "--indep-pairwise 500kb 1 0.5"), and there's no positional information for a contig (all POS values are 0), the window will contain the entire contig.