The point is to look at genomic variations across the entirety of the human genome, not so much to match genomic variation to phenotype. What you're saying makes sense, but I think that is more a next step.