Fly pointed out this paper in the new issue of Science. The "T allele" of the protein KIBRA is associated with performance on certain human episodic memory tasks first in a Swiss cohort and then in two other populations controlling for attention effects and stratification effects. The "good memory" allele frequency varies across populations: Asian (75%) > African (~50%) > European (25%). fMRI during a memory task showed higher activation in the hippocampus and other related memory structures in people with the "bad memory" allele, suggesting that the memory structures were having to work harder or more inefficiently to encode the information. So far KIBRA is known to interact with dendrin, protein kinase C zeta, and protein kinase m zeta. You may recognize the name PKM zeta from a science paper a few weeks ago implicating PKM zeta in memory maintenance. I don't know what dendrin does yet. The initial SNP discovered is in an intron and doesn't seem likely to affect function, but there are other polymorphisms really close by, in neigboring exons, that are also significantly associated with the memory phenotype. It will be interesting to see in which part of the KIBRA protein these differences lie.