Hehe well the reason is that these integer values are numbers of non-reference alleles at specific sites and and AoA will be created for each gene and I'm taking these combinations and doing a correlation on them to a case-control pattern so to see if adding any of the sites in a gene produces a fit to that model. Yeah it's all hypothetical so if you're a statistician I know this likely isn't an optimal approach to looking for mutations at many sites over a gene to identify a target gene for further testing...yeah I am sure you didn't want to know all that much anyways!