What does the number of unique haplotypes within a project say about the variety of the underlying target population's DNA? Or, more precisely, the ratio of the number tested to unique haplotypes. How, would this affect potential match rates? FTDNA's new GAP allows us to get some data from http://gap.familytreedna.com/project-statistics.aspx and http://gap.familytreedna.com/unique-haplotypes.aspx. (Secured sites -- log onto your GAP and go to the "Project Statistics" & "Unique Y-DNA Haplotypes" links.) We can combine these statistics to get inter-project relevant comparisons. For our Taylor project on 31 Dec 2010: Tests Number Haplotypes* Ratio Y-12 361 232 1.56 Y-25 272 234 1.16 Y-37 245 229 1.08 Y-67 139 135 1.03 *"Haplotypes" means unique haplotypes (also called "unique strings"). A difference on just one marker means a separate haplotype. Remember that everyone tested at a higher level is also tested at the lower levels, so gets included in the number tested and haplotypes. Naturally, there are fewer unique haplotypes at the lower levels, as there are fewer degrees of freedom. At the 12-marker comparison level, there are slightly more than 3 haplotypes for each 2 tested; at the 67-marker level, the ratio is slightly more than 1:1. A ratio of 1.03 67-marker tests (individuals) per haplotype would seem to indicate a wide variety in the target population's DNA; almost every man tested yields a different haplotype. Of course, many of the unique haplotypes will match another; matches allow some differences in haplotypes. Still, this seems to be a rough measure of underlying DNA variety within the project's population. Does anyone have any thoughts? How does your project's data compare? -ralpht_/)
Ralph For what it's worth here are my figures for the Cruwys/Cruse project: Y-DNA12 53 21 2.52 Y-DNA25 53 32 1.65 Y-DNA37 53 39 1.35 Y-DNA67 16 16 1 I think I'm getting more diversity for my rarer surnames because of the higher proportion of project members from the UK who are less likely to be related to each other than their American counterparts, and also because I'm targeting specific lines for testing. Debbie Kennett http://www.familytreedna.com/public/CruwysDNA