Monday, October 01, 2007

A picture is worth a thousand words   posted by p-ter @ 10/01/2007 07:10:00 PM

Below is a graph of the derived allele frequency spectrum in the three HapMap populations (from here). Many papers have been written about this spectrum, statistics have been derived around it, arguments have gone on about its interpretation, and now...we can simply look at it. Data is a wonderful thing.

ADDENDUM: I should probably give a little explanation of the graph. Razib has added the names to the population labels, and the dotted line gives the expected allele frequency spectrum (given their two-chromosome ascertainment strategy, in case you're wondering why it's a straight line and not theta/i) under the assumption of a constant population size. The shift towards rare alleles in the Yoruba indicates a population expansion in recent history, and the shift towards common variants indicates a population bottleneck (or a longer-term small population size) in the European and Asian populations.