It compares the observed frequencies of specific allele patterns to what would be expected under a simple tree-like model of population separation, and if the observed frequencies differ significantly from the expected frequencies, it indicates that something more complex than simple population branching has happened.

When this test was applied to datasets that included ancient South American genomes alongside modern genomes from populations around the world, it found something unexpected: some ancient South Americans, and some present-day Indigenous populations in South America, showed statistically significant excess allele sharing with populations from Australia, New Guinea, and the Andaman Islands.

The excess was small — not large enough to suggest that Australasians made up a major portion of the ancestry of these populations.

But it was statistically significant, meaning it was larger than could be explained by random variation.