It was the result of comparing USR1 to the other genomes available in 2018 and finding the model that best fitted the observed pattern of genetic relationships.

The model was a hypothesis, not a certainty — it was the most parsimonious explanation of the available data, subject to revision as additional data became available.

And revision, in the years since the 2018 paper, is exactly what happened.

Not revision in the sense of the core findings being overturned — USR1’s position as basal to other sampled Native American populations and her classification as representing a previously unsampled ancient lineage have been robustly confirmed by subsequent research.

But revision in the sense of increasing complexity and increasing uncertainty about the specific demographic history that produced the observed genetic patterns.