It would be very useful to be able to identify SNPs that have undergone evolutionary selection in the recent past, as this may indicate that they are important in resisting disease. One approach is to look at the haplotype structure within the genome. Generally haplotypes are broken down over time by recombination, so if a SNP allele is associated with only one haplotype or with a few haplotypes that are very similar to each other, this suggests that it has arisen recently. If the allele is also at high frequency, that suggests that it has undergone recent positive selection.

Haplosimilarity is an attempt to apply this approach across large genomic regions. It is a web-based algorithm whose input is SNP-based haplotypic data, and whose output is a metric of haplotype similarity for the minor allele of each SNP, estimated using a sliding window and normalised against the major allele. To give an idea of what a highly selected allele looks like, the sickle haemoglobin allele of the beta-globin gene has a haplosimilarity score of 56; and in a screen of over 4000 SNPs along chromosome 20, the highest score recorded was 20.

You can use the Haplosimilarity algorithm to analyse your own haplotype data, or you can browse results that we have already computed for a region of chromosome 20 that has been mapped in detail.

Analyse your own data

Browse chromosome 20 data

The plot below shows the computed haplosimilarity value drawn as an annotation on a MARKER map, for a test set of data.