Once again we are hearing claims that the "only valid" genealogical DNA results are those from actual ancestral remains. But is this true?
The answer is, "it depends".
When you have a thoroughly confirmed finding like Richard III, you know what you've got.But we now have at least three prominent cases in which "ancestral DNA" just didn't cut the mustard.
The oldest and best-known, is the "head of Henri IV/blood of Louis XVI" case that supposedly "proved" the Bourbon Y-DNA haplotype to have been G, as confidently predicted by Merovingianist theorizers. Unfortunately, later and far more detailed tests on three descendants of Louis XIII (two via Louis XIV and one via his brother Phillipe d'Orleans) gave completely different results (a specific R1b variant). And it eventually came out that neither the head nor the blood had been adequately authenticated (there were even problems with the chain of possession of the head).
Then there was the attempt to determine the Y-DNA haplotype of the dukes of Normandy - which was foiled because it was discovered that someone, at some time, for reasons unknown, had been playing shell-and-pea with the remains, shuffling in bodies that predated the founding of Normandy and things like that. So nobody was able to verify an *actual* duke's remains, and the investigations were quietly dropped.
Now we have the "toebone or Robert the Bruce", which may not be anything of the kind, due once again to problems with the chain of possession.
Your best hope, as with Richard III, is an intact and undisturbed finding, backed up by DNA matches with modern descendants (mtDNA in Richard's case, because it turns out there were *no* known surviving male Plantagenet descendants - not even the Somerset "cousins").
The farther you are from that, and the more problems there are with the chain of possession, the less accurate any results are likely to be.