Razib from Gene Expression has a frankly heroic post up dissecting a recent paper on genetic variation around the FUT2 gene, which encodes a cell surface protein involved in the pathway that produces the ABO blood group antigens.
The story illustrates some of the difficulties associated with extracting information about human evolution from genetic sequence data. There is fairly compelling a priori evidence for natural selection having a strong effect on this gene during recent human evolution, but even with data on genetic variation from a large sample of worldwide populations it is difficult to extract a clear picture of exactly what has gone on.
Razib does a great job of conveying the complexity of the resulting data, and of putting the results in a broader context - check it out.
Incidentally, Razib has four posts nominated for the 3 Quarks Daily Prize in Science - if you like his work, you should go vote for it right now. Other genomics entries include Derek Lowe's piece on the genomics frenzy of the 90s, and my own post on the value of a genome sequence.