snopes.com

snopes.com (http://message.snopes.com/index.php)
-   Science (http://message.snopes.com/forumdisplay.php?f=25)
-   -   Why determining ancestry is rarely accurate (http://message.snopes.com/showthread.php?t=96189)

ganzfeld 10 November 2017 02:21 AM

Why determining ancestry is rarely accurate
 
Full title:
These siblings took DNA tests and got different results. Why determining ancestry is rarely accurate.

http://www.philly.com/philly/health/...-20171012.html

Quote:

“It’s very difficult to accurately find your ancestry under any circumstances,” said Jonathan Marks, an anthropology professor at the University of North Carolina at Charlotte. “There has been genetic mixing for thousands of years. These tests are fun but rarely accurate — 10 percent Scandinavian could be no Scandinavian because the test could very easily be 10 or 15 percent off.”

The imprecisions are not limited to Ancestry. My sister’s son took a DNA test from 23andMe, another popular DNA testing company. His father is Italian and Irish, named Rettaliata. The test showed no Greek and just 1 percent Italian.

crocoduck_hunter 10 November 2017 05:26 AM

My mother had her ancestry tested from one of those sites. For $50 bucks, it told her roughly the same thing that she already knew based off of family history: she's Irish and British.

ganzfeld 10 November 2017 05:54 AM

The ones I've seen will say something like "2% Sub-Saharan African". While it's nice to know that people in my (entirely Western European) family share some genes with some people in Sub-Saharan Africa, I knew that % claim was bunk right away. And not because I think it's impossible we have some distant ancestor from Africa.

First, if we go back only a few thousand years - long long after people first spread out into the world - we all have African, and European and even Polynesian, genes. People mixed at the edges a lot more than we might imagine and the genes did spread. Second, Sub-Saharan Africa has more genetic diversity than the whole rest of the planet put together. I don't believe for one second that the folks over at whatever-and-me-gene-counters have any reliable way of telling what's from there and what's not. Yet they presented it as if it's a likely fact - the rest of the nonsense as well. It was sad to see people who should know better take it as if it has a good scientific basis.

I, too, noticed a difference in siblings that couldn't possibly exist. (Between two maybe could be a random chance but between three or more?? No, I don't think we won the mixed-up genome lottery of more than one in a trillion).

They found distant cousins and so forth. Well, so? We probably have far far far more in common with our next-door neighbours than these people we've never even met.

Seaboe Muffinchucker 10 November 2017 02:12 PM

Two friends (a brother & sister) had this done. The results came back more or less as expected for immediate ancestry, but the proportions were different. That actually made sense to me, because there are genetic differences between siblings.

Seaboe

Esprise Me 10 November 2017 04:20 PM

I'm extremely skeptical of the premise. What are these Irish or sub-Saharan African genes? How do we know that's where they're from? Any given genetic trait can evolve independently in multiple locations through random mutation and, if it's advantageous in the local environment, proliferate. Of course if you study the genetic profiles of groups of people in different geographical areas you'll see trends, but someone could still end up with a gene that's common in Ireland without having Irish ancestry.

One small part of my job involves asking my clients whether they have any Native American heritage. This is to comply with the Indian Child Welfare Act, which gives tribes the right to participate in juvenile dependency proceedings or request the case be transferred to the tribal court if the subject child qualifies for membership. I've been surprised by how many of my clients have already shelled out for one of these tests, or are eager to do so. I understand they're pretty expensive, and in any case, getting a result that you have Native American heritage doesn't mean your child is eligible for membership in a tribe. But in any event, I've yet to meet someone who took one of those tests and didn't get a positive result for Native American.

ETA: My inner conspiracy theorist worries these companies are really in the business of creating a DNA database they can sell or rent to the highest bidder, whether that be a totalitarian government, health insurance companies looking to exclude high-risk patients, or some other sinister as yet unforeseen commercial enterprise. Possibly involving clones. Evil ones, perhaps.

Lainie 10 November 2017 04:27 PM

Well, I didn't take one, but my brother did -- no Native-American ancestry reported. No sub-Sahran African, either. 100% European, 80 or 90% Northern/Western European, IIRC.

Alarm 10 November 2017 04:34 PM

I think it was in Patton Oswalt's new comedy special, he was talking about a certain percentage that was mid-eastern and when he called he was told that was because of Genghis Khan... because he was a...prolific... womanizer.

I think it's a good reminder that interpreting DNA is more of an art than a science, and that it should be taken with a grain of salt.

crocoduck_hunter 10 November 2017 04:41 PM

Quote:

Originally Posted by Esprise Me (Post 1963949)
I'm extremely skeptical of the premise. What are these Irish or sub-Saharan African genes? How do we know that's where they're from? Any given genetic trait can evolve independently in multiple locations through random mutation and, if it's advantageous in the local environment, proliferate. Of course if you study the genetic profiles of groups of people in different geographical areas you'll see trends, but someone could still end up with a gene that's common in Ireland without having Irish ancestry.

Generally, you won't find the exact same mutation arise independently in two completely isolated populations: the Sherpa people of the Himalayas, the people of the Ethiopian Highlands, and the Inca and related peoples of South America all have mutations that give them advantages for living at high altitudes, but while the effect was the same (increased oxygen transport ability) but the mutations that caused it are different in each group.

The big issue is, as Ganzfield already said, they're looking at current populations and assuming that their genetics haven't changed in the last 500 or 1000 years, when we know that humans like to get around in both senses of the term.

jimmy101_again 10 November 2017 04:46 PM

A common misunderstanding that confuses people when they try to understand their results, is that the DNA matching is to the genetic makeup of people in that region today. Your 2% subsaharan match does not mean your ancestors had subsaharan blood, it means the present day subsaharan people share that much DNA with you. The assumption is that old world peoples don't move around much, and that really isn't a very good assumption.

It has been proposed, but afaik nobody has pursued it, that DNA be collected from graves in various regions. That would allow comparison of modern DNA with the DNA prevalent in particular places at particular times in the past. Getting DNA from 1000 year old burials should be doable. For most people that would significantly extend how far back they can reasonably trace their ancestry.

ganzfeld 10 November 2017 04:51 PM

Quote:

Originally Posted by Seaboe Muffinchucker (Post 1963927)
Two friends (a brother & sister) had this done. The results came back more or less as expected for immediate ancestry, but the proportions were different. That actually made sense to me, because there are genetic differences between siblings.

No, sorry, that does not make sense. If the difference were noticeable on these markers in two people in the same generation with the same parents then the difference over four or five or whatever generations they are supposedly counting back would be too great to have the test at all. Once (g)you go through that exercise you realize they aren't really making any specific time frame, that means that they aren't talking about any specific places either because that's the whole point - genes moving from place to place between population during some time. If that time frame is too small then there simply isn't enough time for the genes to move at all. If it's too large then we all have the same sets of ancestors.

So what the test really means is that (g)you share some very specific genes with people in some other place now. By choosing which genes are used as markers, that percentage could say anything for anywhere because we share more than 99% of our genome with every other person on the planet. One might as well have a test to show which mole locations one shares in common with people from Ethiopia and call that a percent.

jimmy101_again 10 November 2017 04:53 PM

Quote:

Originally Posted by crocoduck_hunter (Post 1963957)
Generally, you won't find the exact same mutation arise independently in two completely isolated populations: the Sherpa people of the Himalayas, the people of the Ethiopian Highlands, and the Inca and related peoples of South America all have mutations that give them advantages for living at high altitudes, but while the effect was the same (increased oxygen transport ability) but the mutations that caused it are different in each group

More often than not though they don't look at "functional mutations", generally they are looking at variations that may not have any affect. Indeed the best evolutionary clock is changes that are non-functional since functional changes bias the propagation of the genetic change and therefore bias the clock.

Gene people can look at the roughly 90% of the genome that doesn't encode protein, or they can look at the third nucleotide of each codon which can be changed without changing the encoded amino acid, and therefore doesn't affect the viability of the individual.

ganzfeld 10 November 2017 05:06 PM

Quote:

Originally Posted by jimmy101_again (Post 1963961)
Getting DNA from 1000 year old burials should be doable. For most people that would significantly extend how far back they can reasonably trace their ancestry.

I don't think so. There are very few markers that do not exist to some small or large percentage everywhere in the world for the past 150,000 years. Those few that do would not allow any interesting or marketable difference. The vast majority of people would find out they're "most likely mostly European", not the interesting but nonsensical horoscopes they get now.

Beachlife! 10 November 2017 06:09 PM

Part of the issue could be that there are at least three different companies offering different things. One big difference is whether genotyping or sequencing is being used.

Die Capacitrix 10 November 2017 06:33 PM

Quote:

Originally Posted by Esprise Me (Post 1963949)
But in any event, I've yet to meet someone who took one of those tests and didn't get a positive result for Native American.

My parents both did DNA testing, but with two different companies. My dad's came back with Native American, my mother's not. It could be interesting if I were to do testing with both companies and see if there's any similarity.

Seaboe Muffinchucker 10 November 2017 06:43 PM

Quote:

Originally Posted by ganzfeld (Post 1963963)
No, sorry, that does not make sense.

Well, yes, now I know it doesn't make sense. At the time, however, it did.

Seaboe

jimmy101_again 11 November 2017 11:44 AM

Quote:

Originally Posted by ganzfeld (Post 1963968)
I don't think so. There are very few markers that do not exist to some small or large percentage everywhere in the world for the past 150,000 years. Those few that do would not allow any interesting or marketable difference. The vast majority of people would find out they're "most likely mostly European", not the interesting but nonsensical horoscopes they get now.

There are many markers that are much less than 150,000 years old. How do you think they get the current results placing people in, for example, the British isles? Genetic variation is fast enough to tell two identical twins apart if you try hard enough. Tracing lineage is even easier since it actually doesn't require any genetic drift, just normal genetic pairing that occurs with every offsrping.

ganzfeld 11 November 2017 10:56 PM

True. I didn't say there aren't markers younger than that.

The point is that the vast majority of them wouldn't tell a customer of these services anything marketable and they certainly wouldn't allow this kind of "You're 3% Italian" nonsense. They might be able to say "there is a good chance you had an ancestor from Scandinavia" but that would include pretty much everyone iwith roots from Northern and Eastern Europe. In a few rare cases they would be able to say "you have this unusual gene that only occurs in people from Iceland". People in Iceland might not pay to learn that. People outside Iceland might but for those rare markers I don't think there would be enough of them to make a marketable product. Such markers are useful for research in large samples for trying to piece together the history of the human species and in small samples when they relate to some specific events or physical presentation.

But I don't know; Maybe people would pay for that. I wouldn't but I'm not that interested in my 'genealogy'. I think it's pretty much all bunk.

Cervus 12 November 2017 12:01 AM

My mother has a gigantic book that traces her family back to Adam and Eve. (There are not enough eye-rolling smilies for that.) Some distant relative of hers put it together in the early 1990's and deduced that I was the 12th cousin of Princess Diana. For some reason this information was considered important enough to warrant a Post-it note on the first page when they sent the book to us. (But according to the book I am also related to everyone who ever existed, so...:confused:)

The story we were told is that my mother's family were Mormons and kept "extensive records", whatever that means. I'm not entirely sure how much of it my mom fell for, though she did truly believe, at least for a little while, that I was related to Princess Di and actually went around bragging about it to people. It was so embarrassing and weird that she'd latch on to this random thing. (I don't know if she took the Biblical lineage seriously. You never can tell with her.) Basically, I think what happened is that once her family tree ran out of hard evidence in the 1800's, someone tacked on the lineage of every royal family in Europe (with historical figures like Charlemagne and Christopher Columbus thrown in) and then added every Biblical figure in there just for the hell of it. At the end of the book it literally says "Adam (1)" and "Eve (2)" and that's where it ends. The book is about 1,000 pages long and it's ridiculously entertaining, but I hope she didn't shell out money for it.

GenYus234 12 November 2017 02:19 AM

So Jesus is your first cousin 71 times removed?

PS. You should add a Post-It note above Eve with "rib" and one above Adam with "dust".

BrianB 12 November 2017 08:10 AM

Neat article. Thanks for sharing it.
Quote:

Originally Posted by Cervus (Post 1964059)
My mother has a gigantic book that traces her family back to Adam and Eve. (There are not enough eye-rolling smilies for that.)

That's awesome. :D
Quote:

The story we were told is that my mother's family were Mormons and kept "extensive records", whatever that means.
The LDS (Mormon) church has very extensive genealogical records due to their belief in baptism of the dead.
Quote:

Basically, I think what happened is that once her family tree ran out of hard evidence in the 1800's, someone tacked on the lineage of every royal family in Europe (with historical figures like Charlemagne and Christopher Columbus thrown in) and then added every Biblical figure in there just for the hell of it.
You're spot on here. My degree is in history and when I was in college we studied what were and were not historical records and what was and was not history. Many of my fellow students were sadden to learn that genealogy isn't history and family trees aren't historical records.

In fact, one of the things I learned about was that in the 19th Century there was a cottage industry for creating family trees that tied you to famous historical figures. You could, for example, ask for a family tree that shows you're a direct descendant of England's King Richard I or George Washington. No matter that neither had any children -- you would get that family tree. So, you're probably right it was all tacked on in the 1800s.

And like you, I've also have had to deal with this although the claims weren't quite so dramatic. My mother asked me to research some places that were built by people who allegedly were our ancestors. I was happy to provide the info she wanted because she really enjoys genealogy. However, I felt a bit guilty that I bit my tongue when she brings up the subject of family trees.

Brian "direct descendant of William Adama" B


All times are GMT. The time now is 12:15 AM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.