500 years

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 03:04

Can someone explain how is it that DNA admixture reflects ethnicity over the past 500 years?

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 03:11

As a follow-up, can you get separate admixtures for paternal and maternal sides?

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 03:21

darkerhorse wrote: ↑27 Jul 2022, 03:04 Can someone explain how is it that DNA admixture reflects ethnicity over the past 500 years?

With each successive generation (20-40 yrs), ethnicity gets weaker unless supported by the same ethnicity with regard to spousal contributions.

Occasionally a particular ethnicity gets "lucky" and is consistently replicated but I think that's more of an anomaly than the norm. Some ethnicities can literally be "phased out" within a couple of generations, especially if the prior gens were considerably mixed before, rather than purer breds of a particular ancestry.

If a particular ethnicity is "reinforced" by the same or similar ancestry later on in the gen cycle, I can see where it could hold out longer than an ethnicity that is not.

After 500 years of little or no reinforcement, I think any particular ethnicity would likely average 0-2% at best. Again, only if not reinforced by a later jolt of the same ethnicity at a later point.

JMHO. I'm not even sure science has that tight a grip on it all...

As to question # 2, Ancestry can now separate maternal & paternal admixtures. They are identified as parent 1 & parent 2. In your case, (as well as mine), you'll know which is which...

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 03:45

So here's an example:

Ancestry's matrix gives me a full 50% Italian/Med/W Asian & 50% NW Europe & UK.

As you can see All the Italian+ comes from parent 2 (my mother) & the rest from parent 1 (my father)...

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 03:49

Also their percentages are just "best estimates" & vary widely -- they say the 24% Northern Italy could be 20-40%, & the 22% Southern Italy could be 19-46%. So it's far from an exact science...

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 04:12

That's useful information. Which company did it come from?

Since siblings can very in their admixtures, I'm assuming that the percentages for parents reflect where your percentages come from, rather than what percentage the parent is.

For example, 27% of your 27% UK comes from parent 1, not that parent 1 was 27% UK. Parent 1 wasn't tested so we don't know their own percentages.

Moreover, If your sibling was 35% UK then parent 1 would be 35% UK in this table, yes?

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 04:19

Upon further thought, the breakdown by parent isn't all that useful in my case (or yours) since no overlap in ethnicities is expected between parents, especially over the past 500 years.

Of course, there might be a surprise, such as Nordic or Norman ancestry on both sides, but that's very unlikely to be detected over that rime period.

The paper trail suggests my paternal side is virtually thoroughbred over the past 300 years (and counting).

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 04:25

Detail on the paternal side, like your Northern Italy vs Southern Italy, would be of interest, especially if it had percentage Sicilian (which the paper trail suggests should be 100% over the past 300 years).

But, the relatively large confidence intervals could be problematic.

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 05:14

You are correct in your example that just because I have a certain percentage of a particular ethnicity, the parent can have a much higher percentage of that same ethnicity. Just a rough example:

If I get 45% combined Italian, 3% Anatolian/Caucasus and 2% Levant from my mom, then theoretically she COULD have 90% Italian, 6% Anatolian/Caucasus, and 4% Levant. Of course, there COULD be wide variations in those numbers. My (illegitimate) cousin Debbie doesn't show any Anatolian/Caucasus, but instead shows Cyprus. Her southern & northern percentages are flipped from mine. So, as you can see, it's even random for cousins.

Now an explanation for this is WHERE the genome is viewed -- EVERY testing company only looks at a microscopic portion of your total genome. AND each one looks at a DIFFERENT snippet.

This is apparent as I have the same DNA matches who uploaded tests to several matrices (as I did) and EACH time the results are different -- sometimes closer, sometimes more distant, all depends on the overlap of the genome being looked at.

On GEDMATCH, I uploaded BOTH my Ancestry & 23&Me tests. When I check for matches, both tests glean different lists, with only a percentage being on BOTH lists -- but even then, when I compare chromosomes, the results vary widely between the 2 tests because they are comparing different sections of genome (the overlap obviously varies, too, but there's no way to track that). Even if I compare the 2 tests of my own, I'm not a perfect 100% match as small parts of the genome are on one test & not the other.

And if your head hasn't already exploded, you're well on your way to finding out just how complicated & random this whole damn thing is...

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 05:20

BTW, the parent 1/parent 2 comparison is from Ancestry. It was available to me the first week of May of this year.

And your question about the sibling is yes. If they had 35% from parent 1, that parent would indicate 35%, but could have much more...

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 05:43

Just another thing to consider, using my own ancestry breakdowns as an example. I've gone back 150-200 years (in a small case longer) on my Italian side without ever leaving the area of origin. Just using my Ancestry results, I have a combined 5% Western Asian -- so that no doubt arrived in Italy (whether as an individual or as many mixed individuals/families) considerably prior to that. So I think that reinforces my argument that an ethnicity from 500 years ago would only show up as a very minor percentage (if at all)...

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 13:50

MarcuccioV wrote: ↑27 Jul 2022, 05:43 So I think that reinforces my argument that an ethnicity from 500 years ago would only show up as a very minor percentage (if at all)...

I assume you mean a unique ethnicity, one not re-introduced by subsequent descendants, would die out. The math makes sense, being watered down over time.

An interesting question would be, "what makes a person Italian"?

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 13:52

I have a cousin who has had family DNA tests, and concludes it's a "parlor game". That might be extreme, but there's some truth to it - being only part science.

MarcuccioV · Post by **MarcuccioV** » 27 Jul 2022, 14:14

darkerhorse wrote: ↑27 Jul 2022, 13:50
An interesting question would be, "what makes a person Italian"?

THAT is the 64,000 question.

There have been peoples from various places in Italy for thousands of years. There really IS no such thing as "pure" Italian. The mixtures for Sicily are even more recent.

I guess the mixtures sort of get 'averaged out' over the centuries once newer mixtures stop being introduced in large numbers. This applies to most nations, not only Europe or N America but worldwide...

darkerhorse · Post by **darkerhorse** » 27 Jul 2022, 19:05

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

500 years

500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years

Re: 500 years