Abstract
Recent technological advances (high-throughput whole genome sequencing, megainformatics) have made it possible to confirm the results of pioneering studies relying on classical markers, with much more refinement. When neutral common genetic variation is considered, most parts of the genetic diversity is observed within rather than between human geographical populations. However, the fact that genetic markers are correlated and not independent makes it possible to evidence a few major continental groups (Africa, Europe, Asia, Oceania, the Americas), which are not strictly separated and exhibit various intermediary populations; the much higher resolution power of modern markers evidences fine stratifications within each of the main continental groups ('Russian doll pattern'); modern genomics shows that most genetic variants in human populations have low or rare frequencies, and tend to be geographically restricted. Many convergent data strongly suggest that human evolution has not stopped with the advent of modern humans some 40,000-50,000 years ago, contrary to what was classically believed. Evolution and adaptations have instead been abundant and geographically centered in recent times and may still be going on at present. Lastly, admixture with archaic humans (Neanderthals and Denisovans) seems to have played a major role in the local adaptation of non-African populations.