I am using the R package rehh to find SNPs that are undergoing positive selection in genomic regions of interest using the HapMap data. However I am having issues with the consistency of the results. If I have two overlapping regions: Chr1:151.8Mb-153.3Mb and Chr1:152.3-154.4Mb, the IHH, iES and iHS values are considerably different for the overlapping SNPs between the two regions. I would understand iHS being different since it is standardizing the results based on the SNPs that are in the same frequency bin. So the iHS values could potentially be different for a sliding region. But how are the IHH or the iES values different? Those calculations are based on the SNPs in a relatively close neighborhood of a central SNP, so the raw data is the same for the two regions. Did anyone else encounter this problem? Can anyone explain the reason for the discrepancies?