Andes

The Andes are sequenced. Dr David Reich is a machine.

The research is saying that the highland Andean genome underwent drift starting 7000 BC. This implies first that they had the technology to farm there; second that nobody was letting them farm elsewhere. 3800 BC is when there starts to be north/south differences. That implies larger populations in stable-ish sites, as opposed to the nomads they had running up and down the coasts beforehand. By 0 BC, the geneflow between populations ground to a halt.

They'd gone full Neolithic. Like Mesoamerica.

This somewhat implies that the language families of the Andes should be recent, like Romance or Berber. Although it also implies that the families themselves aren't easily grouped into superfamilies.

There weren't many North Americans getting into this viscous jam of peoples post 3800 BC, with one exception: some Californians surfed down there 2200 BC, early in the consolidation process. I don't know their language; if we could trace their pottery maybe we could hazard a guess.