>If humans bred only with other humans, all these markers would create a neat phylogenetic tree, showing that human genetic diversity can be traced to a single population that existed in Africa in the last 100,000 years.

I'm not a statistician, but from what I've studied of genetics, that seems like an unwarranted leap. Phylogenetic trees, at the end of the day, are guesswork (though they probably reflect reality pretty closely.) If you have a neat phylogenetic tree, you've probably been massaging your data.