You may see large changes to your results. As we add new samples to our reference panels this is because a single individual that's a distant relative can be enough to dramatically change these assignments.
Most of the time these changes should make our results more accurate, but it’s also possible for them to cause results to diverge from your paper trail, we may even be taking you back further than your paper trail as we can give you an insight up to 1500 years.
Migration and human movement for many reasons such as famine, war, or marriage makes it harder to pin your ancestors to a particular location than you might think. Your results reflect not just the movement of your ancestors but also your ancestor’s descendants. You may be able to trace your ancestors back for many generations to specific regions but this doesn’t mean that their descendants did not live in other regions. Remember your path does not just stop, even if your historical research has hit a bit of a wall. If these migrations happened far enough in the past your distant relatives may end up in a different reference panel from what you would expect.
It’s difficult to determine the recency of admixture. For example, admixture between Germanic tribes and Romans made a significant genetic contribution to modern France. It’s therefore difficult to distinguish between people with French ancestry and people who have both Germanic and Italian ancestry. This means if you have some known French ancestry and have a percentage breakdown of German or Italian ancestry your french could be hiding there.
We have updated our panels. Our data set has expanded dramatically this means that In addition to being more accurate due to its size, this panel is more representative of different regional populations. We have been able to ‘tidy up’ the small percentages you may have had which may have been noise. This allows us to filter out populations that we are not confident in being true, creating a more focused representation of results. This may result in populations with small percentages (5% or less) that only appeared in the Compete view previously being absent from the updated results