Question on "First: select the SNPs to study."

[anbai106]

Dear Chris,

I read at some place (but I forgot where) that for the suggestive minimum number of SNPs (nsnps) needed for coloc. I assume that we cannot include too few effective snps (after merging two datasets) because of the limited statistical power.

In your tutorial: "SNPs must * have summary data available in both studies * cover a single genomic region (perhaps defined by distance about some GWAS peak or delimited by recombination hotspots), and * represent a dense coverage of the region"

Do you have a suggestion on this?

Thanks