I have SNP data for approximately 600,000 SNPs that I'll be using for eQTL analysis.
I've been advised to derive a pruned set of SNPs that are in approximate linkage disequilibrium (LD).
I've used the SNPRelate::snpgdsLDpruning function in R to do this, using an LD threshold of 0.2.
Is this an appropriate threshold to use?
This takes our 600,000 snps down to ~60,000; using a threshold of 0.1 leaves ~20,000 snps.
This is quite a big difference so I'm wondering if 0.1-0.2 is too stringent.
Is there a more-or-less standard threshold that is used for LD pruning?