A few months ago I posted here talking genetic testing to see if Elsa and her siblings were born out of incest, particularly testing for regions of homozygosity and how it can show shared ancestry but can't be definitively rule incest in or out.
Turns out I was partially wrong. In certain cases, the %ROH (AKA the portion of the genome that is identical) is simply too high to account for the distant shared ancestry commonly seen in patients with ancestry from small/insular ethnic groups. But %ROH only gets so high to where incest is the only explanation in cases where the parents are first degree relatives (parent/child or full siblings), or second degree relatives if there's a generational history of incest (cousins, half-siblings, uncle/neice, aunt/nephew, grandparent/grandchild). So if Elsa's parents were cousins and there was no other inbreeding in the family, ROH testing would be uninformative. It's definitely not a foolproof way to test for incest unless the parents are VERY closely related.