Assortative Mating

I recall hearing about this paper a few years ago, and there may have been similar others in addition to this one. Assortative Mating at loci under recent natural selection in humans.  Nicholas Christakis is one of the authors, which I take as a recommendation. We tend to choose partners who are equivalent to fourth-cousins genetically. Not that they are about fourth cousins, but they have the same amount of genetic similarity. That is in some ways not very much (0.025%) but is also more than the average person you run into in a city (variable, but less than 0.001%) at the loci that are current natural selection.  That means, not the 99% of genetic similarity that we have to a banana, because we are both living and have cells, but the places where evolution is selecting most heavily because of changing environments. Gravity isn't changing.  The amount of sunlight and oxygen isn't changing - there isn't much selection underway at those points.  We've got that covered.

This makes entire sense when I look at myself, my siblings and parents, and my two oldest children. All of us come from that lake that is the North Sea, where genes were in constant exchange, so that even the different varieties are similar. But when I look at my Transylvanian son who married a woman from Manila, the pattern breaks pretty thoroughly. Those populations have not interacted for tens of thousands of years. There is likely a lower threshold beyond which it makes little difference. 

The idea of "similar but not too similar" makes intuitive sense for our default programming. The default is easily overwhelmed by circumstances, such as migration to a different culture (or even a somewhat-different culture), but exerts some influence. The genetic pool we are choosing from is something of a kaleidoscope, not an infinite branching. See also Pedigree Collapse.