Over the past decade or so much of the reconstruction of the human genetic past has occurred through inferences generated from variation of extant human beings. In more plain English the patterns of genetic variation of modern populations have been used to map out the patterns of the past. There are serious difficulties with these sorts of inferences. For example you generate a huge number of potential phylogenetic trees and zero in on the “most probable tree” (or, the distribution of trees). But at the end of the day these inferences are only as good as your assumptions.
Ancient DNA changes all that. You can calibrate and refine your trees (or graphs if you include migration edges) with the real patterns of variation from the past. Because of the reality of the decay and degradation of DNA over time we may never have many ancient samples, just as we don’t have many ancient fossils (or as many as we’d like!). But I am hopeful that we will have enough to prune the possibilities greatly when it comes to the human genetic past.
A new paper in PNAS, not online yet, does just this for some important questions. You can read the press release. But for me the big question that this answers is how literally to take inferences that East Asians and Europeans diverged ~20-30,000 years ago. If these data and results are straightforward then the answer to that question is that the inferences had some problematic assumptions, because East Asians and Europeans began to diverge > 40,000 years B.P. (one possibility is that there has been gene flow between the two diverged populations not accounted for by previous analyses)
Another question which I think will be answered by future results from Europe is going to be the divergence time of Eurasian H. sapiens sapiens lineages. The range of divergence times seems to go from ~100,000 years ago to 50,000 years. If the latter is correct then we will see that European and East Asian individuals from 40,000 years ago should be very close indeed genetically. In contrast if it is 100,000 years ago then they may already be rather diverged.
Finally, my understanding is that the next few years is going to see a revolution in our understanding of the human past because of the large number of ancient DNA studies which will come online. I’m hoping this paper is an appetizer.