Some collaborators and I are building an ancestral inference tool, and we're having trouble obtaining reliable ground-truth data for samples of mixed ancestry. All the ground-truth DNA datasets we're aware of (Thousand Genomes, Simons Foundation) are from persons reported as unequivocally belonging to some population (East Asian, African/African-American, South Asian, or even something specific like Somali, Icelandic or Cambodian.)
But we can't find any samples that are attestedly mixed-ancestry. We've thought about just generating simulated data using VCFs from two persons of different known ancestries, but we're not sure how to account for the non-random aspects of recombination, which are described here: https://www.bio-itworld.com/news/2019/03/06/latest-genetic-map-finds-recombination-is-not-so-random And we don't know of any tools that could perform this task for us.
Does anyone have suggestions for obtaining/simulating DNA data for persons of mixed ancestry? This is a natural enough problem that we suspect others must have encountered it also.