500 years

[MarcuccioV]

Does the 500-year window apply to all DNA testing companies, or only to Ancestry.com?

Some companies advertise "ancient" DNA, but it's mostly populations like "hunter gatherers", etc & not specific ethnicities. For that, I think 500 years is the accepted threshold. Beyond that, things get so diluted down that it would be next to impossible to get any kind of accurate data, I would think...

[darkerhorse]

Before considering ancient DNA, my priority is to pick up where the paper trail leaves off in the 1700s.

[MarcuccioV]

Before considering ancient DNA, my priority is to pick up where the paper trail leaves off in the 1700s.

The "ancient" DNA, IMHO, doesn't tell you much unless you're are a deep-dive researcher or genealogist. I, like you, am more interested in recent generations and ethnicities...

[darkerhorse]

If you assume 30 years per generation, then 500 years implies about 17 generations.

I think you have 131,072 14x great-grandparents back 17 generations?

On my father's side that's 65,536, 32,768 from each of his parents.

So, there are 65,536 possible paternal candidates (living c.15th century) for donors to my admixture DNA (assuming no duplications)?

[MarcuccioV]

If you assume 30 years per generation, then 500 years implies about 17 generations.

I think you have 131,072 14x great-grandparents back 17 generations?

On my father's side that's 65,536, 32,768 from each of his parents.

So, there are 65,536 possible paternal candidates (living c.15th century) for donors to my admixture DNA (assuming no duplications)?

True. However, due to random replication, you rarely get ANY ethnicities beyond about 6 or 7 generations -- unless they are constantly interbreeding with exactly the same ethnicity without any interlopers.

I have a few ancestors, for example, on my father's side which were 100% French. But because they are so far back in the tree, the French ancestry has been completely diluted out.

Ancient ancestry might be good to see where & how your forebears migrated, etc, but for ethnicity purposes it's practically useless...

[darkerhorse]

[

True. However, due to random replication, you rarely get ANY ethnicities beyond about 6 or 7 generations -- unless they are constantly interbreeding with exactly the same ethnicity without any interlopers.

I have a few ancestors, for example, on my father's side which were 100% French. But because they are so far back in the tree, the French ancestry has been completely diluted out.

Ancient ancestry might be good to see where & how your forebears migrated, etc, but for ethnicity purposes it's practically useless...
[/quote]

If they can only capture 6-7 generations back why say that admixture goes back 500 years? Doesn't that exceed 7 generations?

At 7 generations back (1700s?), which is what I have mostly complete by paper trail, you'd only have 128 forebearers. What a difference from 131,000! In my case 64 likely Sicilian.

Interesting point about your French dying out. They'd still be on your paper trail though. So, are you part French or not? I'd say certainly yes. Would you say DNA ethnicity is a subset of your total ethnicity, or that if it isn't in your DNA then you are not of that ethnicity? I opt for the former.

Maybe in the future they'll come up with a way to discern all forebearers in your DNA or elsewhere.

How your forebearers migrated would tell you something about where they came from, which is often related to ethnicity.

[MarcuccioV]

I'm basically neutral on the subject. If the French was diluted out & now only exists by paper trail, I suppose depending on your own preferences, it can go either way. In my case since (on my maternal side) I currently show Italian & Greek, then my supposition is those ethnicities must be fairly recent or have been reinforced by the same ethnicities for many generations. Anything older that diluted out may still be detectable as ancient DNA, but I doubt would have all that much sway in my current ethnic breakdown...

[darkerhorse]

"my supposition is those ethnicities must be fairly recent or have been reinforced by the same ethnicities for many generations."

That's a salient point. Isn't that exactly what admixture captures?

[darkerhorse]

And the more distinct or isolated an ethnicity, and the more it represents a race, the more likely it can be detected. Examples are African-American, Asian, Hebrew, etc. as opposed to European, say.

If you had the same percentage of German and African, I bet the African would be more likely to be identified or detected in DNA testing.

[MarcuccioV]

"my supposition is those ethnicities must be fairly recent or have been reinforced by the same ethnicities for many generations."

That's a salient point. Isn't that exactly what admixture captures?

Pretty much, yes. But there are complexities.