Twins get 'mystifying' DNA ancestry test results (Marketplace)

♪ - A pair of identical

twins

, five DNA

ancestry

test

s. That?! I have French and German. Oh, I don't have that. - Five different sets of

results

. We strongly believe that you and your sister should receive the same report. I don't want to be a rumor killer. I don't think it's entirely accurate. - Do you want to know more about your heritage? You cannot afford to miss this Market. - I'm on vacation in the Mediterranean. Tunisia I think it is like that. And Sicily is like that. I'm going back to my roots or what I think are my roots.

Giuseppe driving the boat. My family returns to Sicily. So here we are in beautiful Noto. It will be good. Off the tip of Italy. And I also have Polish and Ukrainian roots. Or at least that's what my family tells me. So how can I be sure? I thought I married an Italian. I found out that she was only 16% Italian. She was 34% Eastern European. -Judging by marketing like this, DNA

test

ing has the answer. When I found you in my DNA, I knew where my strength comes from. - Millions of people buy kits out of curiosity about their

ancestry

.

Can you point specifically? - Companies promise that their tests will reveal where you come from, who your people are. It's really nice to have some kind of connection to my experience. - But how accurate are they? We're testing five companies to find out. Ancestry DNA, 23andMe, MyHeritage DNA, FamilyTree DNA and Living DNA. Hello. Thank you. Oh, and there's a twist. Oh yeah. You have some slime. There you go. -My sister Carly is also involved in this. That's her son, Max. And no, you're not seeing double. We are identical

twins

. Oh. Alright. -Step one. - Fill the tube with saliva up to the black wavy line.

Carly and I are trying two things. Will companies have different breakdowns from each other? Do I have too many bubbles? And will they have different breakdowns for each of us? I feel like I'm on Law and Order. - Yes. CSI. It is the science of DNA. We are identical twins, so all

results

should be the same, right? Health. But don't touch it. - I check again to make sure we're not mixing our samples. This will reach Dublin, Ireland. The Netherlands. Concordia, Ont. And two go to Texas. I'm shocked. - These ancestry kits can change some people's sense of who they are. 11% Scandinavia, which I love.

Now, this is crazy because, like I said, it's a very large percentage. We are 53% Nigerians. Court! My mother was completely Chinese. My dad is half Chinese. Half of us thought she was Spanish but she is British. 5% of me is Ashkenazi Jewish. I don't even know what Ashkenazi means. It's D-day or rather DNA day for the Agro sisters. Alright, then let's get to it. - First, ancestral DNA, the biggest in the business. Well. - Ready? 1-2-3. That?! 30-- - Russia?! - 39% Eastern Europe and Russia. No way. - What does yours say? Did you get exactly the same thing? - Wow, I didn't expect so much from Eastern Europe.

And it gets very specific with Italy. We get Sicily. - Greece and the Balkans. - Greece and the Balkans. - Yeah! Oh great. Well. - Familytree DNA has a few surprises in store for us. Wow. I am 13% Middle Eastern. - And then there is Iberia. This is cool though. That's Spain, that's Portugal. - Yes, I would never have imagined it. And get this, according to MyHeritage... - ...we may not even be Italian. Holy. So... - 60.7% Balkan. - Okay, click on them. - 19% Greek. Oh my God. - Car, 3.4% Italian. Actually, it says you're 3.4% Italian. Are we then a little or a lot Italian?

Mostly from Eastern Europe or mostly from the Balkans? What's up with these results? I hope Simone Gravel can explain to me. Hello, Dr. Gravel. - Hello, welcome to the laboratory. - Charlie. - Nice to meet you. - Thank you. She is a population geneticist at McGill University in Montreal. The last time she saw your DNA, I guess it was in a little tube like this. I spit out quite a few of those, yeah. That's how it is. So once you send them to the companies, they will take your breath away. They clean it. They will break down the cells so they can release the DNA and then clean the DNA and break it into small pieces.

And then they put it on one of these chips here. - So if we look at this chip, will my saliva on this chip tell me that I'm definitely Italian or I'm definitely from wherever? No. There are many intermediate steps. A lot of interpretation and statistical modeling and, you know, what genome your genome is compared to. - Is that how it works. All of our DNA is made up of 3 billion parts. However, Ancestry companies analyze less than 1% of them. About 700,000 sections where they know that there are slight differences or variants between humans. They then compare the patterns of their differences with groups of people in their databases.

And with the help of mathematics, they try to find out which groups you are most likely to belong to. It's not one hundred percent safe. So the statistics tell us in our model, what is the best situation? It's a statistical assumption. - Yes. - It's an educated guess. But it is never one hundred percent certain. - Sure. When we grew up we were German. - Is not safe? Surely that's not the message here. The big surprise was that we are not German at all. 52% of my DNA comes from Scotland and Ireland. So I traded in my leather pants for a kilt. - So you were shaking your head.

This one drives me crazy. You know, and I see it all the time. Whether you want to wear leather pants or anything else is not tied to your DNA. - Well. Timothy Caulfield is a health policy professor who studies the business of ancestry. It's an exciting story, right? It's about you, right? I understand why they are doing it. But is that really what is happening? Can they really be that precise? I think they are selling something that is not really backed by science. With 23andme.com you can know your DNA percentages from 31 populations around the world. - What do you think of the messages that are being sent to consumers?

I think it's misleading. These companies are really trying to push the idea that this is scientific, right? They are using scientific language. They present it in a way that seems very scientific and precise. - Caulfield says what people really buy is entertainment. I think it's recreational science. You can have a little fun with these things. Don't take it too seriously. Know that you are only getting information that is an approximation of how your DNA compares to other people's. It's not about tracing your heritage. - Back in Montreal, Gravel agrees. Specific percentages should not be interpreted as definitive, as if here is your percentage of ancestry from this place.

That's not it. - No. The only one we can interpret like this is the continental level. - Hmm. Actually? The safest thing they can tell you is what continent are you from? And get this, in Living DNA in fine print, the company admits that some of its results are guesswork. We highlight the sources of his ancestry that are likely present using our best estimate of the exact source. The deeper I go, the more questions I have. And at 23andMe their guesses aren't evident until we stumble upon this. Select the confidence level. That? Why does it say 50%?

Does that mean you're only 50% sure what you told me is correct? Yes. That's what it says. - That is what it means? Yes, or at least 50%. That seems shocking. Just half safe. Yes. - Do you think most people know that? No no. I would like you to present your doubts more clearly. - That's how it is. 50% trust is 23andMe's default setting. See what happens when we change it to 90%, the safest setting. 23andMe tells us we're mostly from somewhere in Europe. I wouldn't take this as if science had talked about something like that. - Forget all my Italian traditions and embrace the Balkan in me. - That's how it is.

Exactly correct? I would not do that. - This is your Market. - DNA ancestry testing, have you heard of this stuff? - Yes. I have seen many offers at New Year or Christmas. Half discount. - I'm in town wondering what people know about the DNA ancestry kits that sell like crazy. I bought it for my mom for fun just because she is very proud of her Irish heritage. Health. But don't touch it. - My twin and I are testing the five best-selling ancestry kits. Oh! - And obtaining different results. That?! I think they are selling something that is not really backed by science. - So you would think that the test should be quite accurate?

It is not? I think they have to be pretty accurate or else they wouldn't get all the hype they get, at least I guess. I think they are true. I mean, I would trust them. We are engineering students. We think about science, we think about precision. Personally, I think they are accurate results. - I did five tests. My sister too. What would you expect the results to be? The same, identical. They should be the same. - The same. Identical. - They should be identical, right? It should be identical. - Now is the time to perform the twin test.

A little chubby. That's me. - Yes, because I think that's me. Ready? 1-2-3. - At 23andMe it is a promising start. Mine says 98.1% European and yours says 98.2%. We are pretty close. - Italian, 37.7. Oh, 36.8. But then we started to grow apart. In this case only 28% are Eastern Europeans. In this case I am only 24% Eastern European. This one actually says Poland. - Not mine. Mine Yes. - That? - Yes. Look. Poland, possible match. - Poland was not detected in mine. She specifically says not detected. And it gets stranger. Okay, this is really funny because, look, French and German.

That? Oh, I don't have that. - You don't have that at all? No. - We are identical twins. But you have to wonder with results like these. I am 3.8% generally European. I'm 12.7. - Wow, that's a big difference. Yes. - How is this possible? In search of answers, I travel to New England, to Yale University, and meet with Mark Gerstein and his international team of computational biologists. They study DNA and statistics. And they are surprised by our twin results. I must say that really surprised us. I mean, we expected two identical twins to have exactly the same ancestry and that's how they should be.

The fact that they present different results for you and your sister seems very disconcerting to me. We thought for sure that the differences had to do with the fact that when a person spit there was a contaminant in the sample. - But then you looked at the raw data and you say you verified it. - There's no difference. - And there is no difference? No difference. It's strikingly similar. It's terribly the same. - The Yale team downloaded and compared our DNA sections from all five companies. The agreement between you and your sister is 99.6%, 99.7% agreement. - 23andMe's own data says we are statistically identical.

The same goes for other companies. Why would Car be more Eastern European than me if our DNA is the same? She is not. I mean, there's nothing to say... You and your twin sister have the same genetic ancestry, after all. There is nothing to say. I mean, that's the truth. - How can we explain this result? This must be purely in terms of the analysis that 23andMe performs on the data. - 23andMe isn't the only company with funny twin differences. Oh great. You are now Scottish and Irish. All right, I'll give it to you. - Living DNA is also causing some problems.

Yours says southern Italy. Mine says Tuscany. I just don't understand how you can have a Tuscany marker and I don't. - I don't know. You have to come from a region of Italy. You can't come from Tuscany and Sicily. It just doesn't work from that. I mean, in this one, your sister comes from England. But you come from Scotland and then Ireland. - But not everyone has us so separated. MyHeritage, the company that says we barely have Italian ancestry, gives us the results for closest twins. One of the good things about this company is that it is very consistent. - Okay, can we go to ascendancy?

And Ancestry DNA and Family Tree are also pretty consistent. So if DNA technology says we're identical twins, why do some companies give us different ancestry breakdowns? We spend many hours trying to understand why they might be different, but we can't. I think the clearest thing is that we don't know how they did the calculations, but we strongly believe that you and your sister should receive the same report at the end of the day. - When we asked 23andMe and Living DNA about our twin results, they said that even small variations in our DNA can cause their formulas to give us different estimates because, remember, while all companies are based on the same DNA science, they have different formulas and people in your database to determine which population group you belong to.

And whatever your results... 11% Scandinavia, which I love. - ..don't get too attached because they can change. There have been some changes. There have been changes in my results. In my other test it was 11% Scandinavian. Where did that 11% go? - It happened to Carly and me too with the ancestry DNA. Look, look, look, look, though, look. This was my first result, but it was updated when the database and math changed. Look at the differences. I think it's crazy how much it changed. - The politics teacherhealthcare knows everything. I was once 100% Irish. I love Guinness very much. But now apparently I'm at like 80%. - Your results do not stay the same over time.

They often change as companies update their databases. The commercials don't mention that. What do you think about that? They don't mention that, do they? In fact, it evolves, which is a really telling point. It shows how imprecise it is and gives an idea of ​​how science really works. - People hear DNA and think science, legit, CSI, Law and Order, that everyone goes to jail or gets out of jail because of DNA. You're right. And they're taking advantage of that enthusiasm, the enthusiasm that comes from the genuine things of science, but also the enthusiasm of pop culture to give it this veneer of legitimacy.

And I think that's a little misleading. Charlsie: This is your market. ♪ ♪ - My twin and I have learned that ancestry tests are not really as advertised. Now I'm more confused than before I started. And I have more questions than answers about who I am, where I come from, and what it all means. Look, I don't want to be a grump. If people find this funny, that's fine. But don't put too much weight on it. - However, the industry is booming thanks to low prices, clever marketing and our own curiosity. You can connect more deeply with the places of your past. - All five companies in our test say they use rigorous science.

They also acknowledge that their ethnic breakdowns are statistical estimates. Family Tree DNA says that discovering ethnicity is unclear because it is both a science and an art. MyHeritage says its clients understand that the results are estimates. I'm going to show you a commercial. On the street, do people really know? Now that you've seen the ads and some of the promises, what do you think about what you're really getting, what can they tell you? There is... not the slightest doubt that what they are showing you is accurate. - It's an estimate. Actually? It's not exact? - Estimate.

That's what the experts tell us. Are you surprised? Yes, I wanted to do it. I'd like to assume there's a little more evidence behind it when they check the DNA. More science behind it. - Do you think there is good science behind this? I would like to believe, yes. - Well. I don't know. I'm a little sad right now. Make a statement about how accurate they are. Those ads where they have very small writing at the bottom of the screen, just write it there and say the results may vary depending on the test you run. I wish the messages were more scientifically accurate, right? - Mmm-hmm.

If you could recreate them for these companies, what should the message be? How should they look? Well, you know, I don't want to be a bully. I think my ads would probably be a little boring. Promote the idea that this is recreational science. You know, let's see how your DNA compares to other people's DNA. - Well. This won't be very exact or precise, but it will be fun. - Caulfield says there is a subtle message in some of the marketing that he finds worrying. Yes, there are biological variations between populations. But there are no such discrete biological boundaries.

It is a dangerous message if you think about this world where we are worried about nationalism and tribalism. They are saying that there are biological families. This is scientifically incorrect but also potentially socially problematic. - And here's some science that everyone agrees on. Humans share more than 99% of our DNA. We are much more similar than different. - Have you had a DNA test? What did you think of your results? Take our Marketplace survey on Facebook.