Cross trait map

Author's note: that I'm currently experimenting, and trying to work out an alternate method for imposing a sort on the SNPs in the below graphic. I've put a little gray rectangle to the right of each of the phenotype descriptions, and if you click that rectangle then all of the SNPs ( across all 25 phenotypes) will be sorted by the magnitude of the chosen phenotype. I'm imagining that biologists might wish to easily identify a lead SNP, and to see if it tracks with other SNPs. The problem is that by sorting on the basis of P value, I have to ignore the default sort (which is based on each SNP's genomic location). Both methods of sorting are potentially useful, but going back and forth between them seems disruptive and confusing. What would be a better approach? Maybe using another graphics channel ( size, shape, color, direction)? or am I instead simply trying to take this graphic in an unnatural direction? Any ideas or comments are welcome, and could come through the feedback mechanisms on my bovinecontemplation.org website.