the argument 50 is the size of the sliding window that will roll over the contigs in search of LD, 5 is the step size for when the sliding window moves, and 0.5 is the correlation coefficient cutoff for pruning SNPs. So if two SNPs are correlated with a 0.6 r-squared value, that SNP would be pruned. We could think about increasing that value, too. Doing so would likely decrease the number of SNPs eliminated. Maybe we try pruning to 0.8?
For the steps detailed here ,in the command
the argument 50 is the size of the sliding window that will roll over the contigs in search of LD, 5 is the step size for when the sliding window moves, and 0.5 is the correlation coefficient cutoff for pruning SNPs. So if two SNPs are correlated with a 0.6 r-squared value, that SNP would be pruned. We could think about increasing that value, too. Doing so would likely decrease the number of SNPs eliminated. Maybe we try pruning to 0.8?
More on these commands from the plink manual here