"Race" is more of a sociopolitical idea with a particularly noxious history, but human groups with differentiable shared recent ancestry - while fuzzy - do exis

"Race" is more of a sociopolitical idea with a particularly noxious history, but human groups with differentiable shared recent ancestry - while fuzzy - do exist. Because "race" can be described as a continuous distribution, it helps to think about it probabilistically: there are no "definitely Caucasian" SNPs, but we do have suites of what we call Ancestry-Informative Markers that can be use to infer the likelihood that someone belongs to a particular ancestry group. Perhaps unsurprisingly, many of the polymorphisms associated with pigmentation are informative in this regard. The fuzziness of the continuum means that it is quite possible to define an arbitrary number of "races", depending on your chosen criteria. And admixture at multiple points in history further complicates the picture. But it is possible to discern biogeographical mappings within a generous error margin. As the parent poster points out, these mappings sometimes conflict with the those that we've constructed out of ideological convenience.