One of the strangest aspects of our understanding of evolutionary biology is the tendency to conflate a sprawling protean dynamic into a sliver of a phenomenon. Most prominently, evolution is often reduced to a process driven by natural selection, with an emphasis on the natural. When people think of populations evolving they imagine them being buffeted by inclement weather, meteors, or smooth geological shifts. These are all natural, physical phenomena, and they all apply potential selection pressures. But this is not the same as evolution; it’s just one part. A more subtle aspect of evolution is that much of the selection is due to competition between living organisms, not their relationship to exterior environmental conditions.
The question of what drives evolution is a longstanding one. Stephen Jay Gould famously emphasized of the role of randomness, while Richard Dawkins and others prioritize the shaping power of natural selection. More finely still, there is the distinction between those which emphasize competition across the species versus within species. And then there are the physical, non-biological forces.
Evolution as selection. Evolution as drift. Evolution as selection due to competition between individuals of the same species. Evolution as selection due to competition between individuals of different species. And so forth. There are numerous models, theories, and conjectures about what’s the prime engine of evolution. The evolutionary biologist Richard Lewontin famously observed that in the 20th century population geneticists constructed massively powerful analytic machines, but had very little data which they could throw into those machines. And so it is with theories of evolution. Until now.
Over the past 10 years in the domain of human genetics and evolution there has been a swell of information due to genomics. In many ways humans are now the “trial run” for our understanding of evolutionary process. Using theoretical models and vague inferences from difficult-to-interpret signals, our confidence in the assertions about the importance of any given dynamic have always been shaky at best. But now with genomics, researchers are testing the data against the models.
A recent paper is a case in point of the methodology. Using 500,000 markers, ~50 populations, and ~1,500 people, the authors tested a range of factors against their genomic data. The method is conceptually simple, though the technical details are rather abstruse. The ~1,500 individuals are from all around the globe, so the authors could construct a model where the markers varied as a function of space. As expected, most of the genetic variation across populations was predicted by the variation across space, which correlates with population demographic history; those populations adjacent to each other are likely to have common recent ancestors. But the authors also had some other variables in their system which varied as a function of space in a less gradual fashion: climate, diet, and pathogen loads. The key is to look for those genetic markers and populations where the expectation of differences being driven as a function of geography do not hold. Neighbors should be genetically like, but what if they’re not? Once you find a particular variant you can then see how it varies with the factors listed above.