Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
×
Science

Identical Twins Test 5 DNA Ancestry Kits, Get Different Results On Each (www.cbc.ca) 258

Freshly Exhumed writes: Uh-oh, something is not right with the results of most popular DNA ancestry kits, as a pair of identical twins have found. Charlsie Agro and her twin sister, Carly, bought home kits from AncestryDNA, MyHeritage, 23andMe, FamilyTreeDNA and Living DNA, and mailed samples of their DNA to each company for analysis. Despite having virtually identical DNA, the twins did not receive matching results from any of the companies. "The fact that they present different results for you and your sister, I find very mystifying," said Dr. Mark Gerstein, a computational biologist at Yale University. Gerstein's team analyzed the results, and he asserts that any results the Agro twins received from the same DNA testing company should have been identical. The raw data collected from both sisters' DNA is nearly exactly the same. "It's shockingly similar," he said.
This discussion has been archived. No new comments can be posted.

Identical Twins Test 5 DNA Ancestry Kits, Get Different Results On Each

Comments Filter:
  • Now stop stealing my clothes and trying to trick my girlfriend, asshole!

    • Now stop stealing my clothes and trying to trick my girlfriend, asshole!

      Good use of commas there.

    • Scanning for reply by elrous1...

  • by mentil ( 1748130 ) on Friday January 18, 2019 @08:30PM (#57985048)

    Clearly, one of them was switched at birth. With their triplet.

    • by jrumney ( 197329 )
      Clearly not. The Yale biologist analysed them and found their DNA to be 99.6% the same. This places them somewhere on the scale between human-human (99.9% the same) and human-chimpanzee (98.8% the same). If I were to hazard a guess, I'd say that one of the twins is a neanderthal.
      • by kinko ( 82040 ) on Saturday January 19, 2019 @09:06AM (#57986636)

        99.8% between humans and chimpanzees refers to the entire genome (~3 billion base pairs).

        23 And Me and related companies only look at about 3 million positions - the positions that are often different between different human populations/races (these are known as Single Nucleotide Polymorphisms).

  • by aberglas ( 991072 ) on Friday January 18, 2019 @08:36PM (#57985066)

    They should be identical. Will not be due to normal error. Or may not be even close due to incompetence.

    • by mentil ( 1748130 ) on Friday January 18, 2019 @09:31PM (#57985304)

      Just curious, is it possible that people are normally chimeras to some degree? It may not be 'error' so much as 'unexpected true result'.

      • Well, women are all chimeras, but each cell should be the same as every other except for which X chromosomes are switched off, which appears random and is where the chimera comes from. And why all three color cats are female. https://www.google.com/search?... [google.com]

        It *is* true that blood tests of women who have been pregnant can pick up small amounts of the DNA of the fetus...but that's *SMALL* amounts. And that's blood tests, where the last time I read these tests used either spit or a biopsy from inside the cheek. (That was awhile ago, however, so check before you believe.)

        There's also a small amount of expected mutation during development, so identical twins aren't really exactly the same. Somewhat well over 99% however.

        • by Antique Geekmeister ( 740220 ) on Saturday January 19, 2019 @07:44AM (#57986478)

          There are apparently human chimeras who merged with fraternal twins in the womb to form one infant. One of the more bizarre but verified cases was Lydia Fairchild, who was found not to be genetically related to her children. The DNA of her cervical smear differed from that measured in other parts of her body, which helped establish her parentage of her own children.

      • It's now accepted that most people not born as twins are chimera and that identical twins are not absolutely genetically identical.

    • by Tony Isaac ( 1301187 ) on Friday January 18, 2019 @11:09PM (#57985698) Homepage

      Measuring DNA results is always an exercise in statistical analysis. Mitosis does not produce exact copies. Every cell division has changes from its "parent" cell. When labs test your DNA, they rely on a large sample size, and calculate averages. Your results at a specific location might be 65% AA and 35% TT. They are going to show a result of AA in this case.

      DNA results ARE meaningful, but it is necessary to understand what the results, and the algorithms, actually mean before making conclusions from them.

      I've been tested by 3 labs; several hundred (of 700,000+) of my results differed between the three labs. When this happens, we geneticists throw out the mismatches as errors. I've never seen this change any results in a meaningful way.

      • by arth1 ( 260657 )

        This seems like it can be mitigated quite substantially by running three tests on each submitted sample. Of course, that would increase the costs.

        • by Tony Isaac ( 1301187 ) on Friday January 18, 2019 @11:29PM (#57985796) Homepage

          The raw results are not in question. The story itself says that the raw data was nearly identical, as one would expect. It is only the extrapolation of that data to infer ethic lineage that didn't line up so well.

          • Re: (Score:3, Interesting)

            by tlhIngan ( 30335 )

            The raw results are not in question. The story itself says that the raw data was nearly identical, as one would expect. It is only the extrapolation of that data to infer ethic lineage that didn't line up so well.

            And that's the important point - the interpretation can differ. Yet if you look at all the ads, they imply an exact location. Spit in the tube and you'll find which neighbourhood your ancestors grew up in.

            Of course, your results are basically guesses by the site - if you ask it for your real locati

            • Correct. What the ethnicity predictions really say is that "you are statistically genetically similar to people who are known to be from region X or Y." They cannot tell where your ancestors actually came from.

      • by kriston ( 7886 )

        Mitosis does not produce exact copies

        I only had some college electives on genetics and biology and even I remember this decades later. Everything about genetics is about statistics.

        Of *course* DNS samples from two living twins are not identical anymore.

      • Mutations are not part of it, at the sites being sequenced. If mutations could have played a significant part in it, the whole thing would have made a little sense.

        At this level scientists are looking at shuffling of the same genetic cards that were the same for many years, not the markings on them left by professional hasslers.

    • The people of reddit have been posting all kinds of 23&me related stories from doing tests over Christmas. While some stories were from parents coming clean about the skeletons in their closets, others may have been spurious results like this. I've always been skeptical of their methods and the validity of the results they have obtained.
    • Twins aren't "identical".
      If you ever had seen two, you knew that.

      One is born bigger, one got a bit more of hormone X while in the womb, etc. p.p.

      Every human on the planet is far over 99% genetically identical to every other human ... perhaps you missed that fact in school.

      • by Luckyo ( 1726890 )

        Genetically identical individuals can have different expression of that genome based on environment. That doesn't make them any less genetically identical.

        If you ever looked at a biology book, you'd know that.

    • I have submitted my data twice to FamilyTreeDNA and once to 23&Me. Absolutely identical results,as you'd expect. I then submitted (with permission) a fairly close relative on the same paternal line. The results differed to the degree one would expect and nothing more.

      However, I only tested autosomal (notoriously problematic) for 23&Me. Near relatives show up on the list at the same distance my family tree shows them at, but that's only a crude measure. That only tells me they got things mostly right

  • is making the random numbers come out the same when they detect twins submitting dna.
  • Let's have the Hensel twins repeat this experiment. Hilarity ensues.

  • by onkelonkel ( 560274 ) on Friday January 18, 2019 @08:58PM (#57985160)

    What is the point of the identical twins (other than adding click-bait value)? Why not submit two samples from the same person under different names?

    • Re: (Score:2, Offtopic)

      by bobstreo ( 1320787 )

      What is the point of the identical twins (other than adding click-bait value)? Why not submit two samples from the same person under different names?

      They are reporters, identical twin reporters. You can't have only one of them getting credit.

    • by Solandri ( 704621 ) on Saturday January 19, 2019 @02:48AM (#57986142)
      e.g. Here's the relevant part of 23andMe's terms of service [23andme.com] (emphasis added)

      Furthermore you agree not to use the Services to: (1) [...]; (2) impersonate any person or entity, including, but not limited to, anyone affiliated with 23andMe, or falsely state or otherwise misrepresent your affiliation with a person or entity;

      I suspect if a reporter did submit samples under different names, a court would side with the press against their terms of service. Eventually.

      But by using identical twins, you sidesteps the possibility of wasting time, effort, and money because your report has been tied up by a gag order while a court mulls over what to do.

      • by DocJohn ( 81319 ) on Saturday January 19, 2019 @08:58AM (#57986616) Homepage

        No judge is going to issue a gag order after something has already been published. There's no reason to alert the companies ahead of publication with the exact details of your investigation. In fact, there's no need to even tell the companies ahead of publication. The only reason this is done to some degree is to get the BS public-relations comment that tries to explain away the discrepancy.

        Who is the reader going to believe? The DNA experts from Yale or some other university, or the for-profit company trying to defend their reputation?

  • by nehumanuscrede ( 624750 ) on Friday January 18, 2019 @09:06PM (#57985182)

    Somewhat begs the question of the accuracy of said tests doesn't it ?

    • No, it doesn't. It only emphasizes the need for education.

      The ethnicity results of the various labs are not empirical science, it's statistical analysis. In a sample size of 700,000+ results, you're going to see some variation.

  • Don't panic, TFA said the raw data was "statistically identical". You can still use tools like SNPedia and Promethease to explore your own raw data.

    • by kriston ( 7886 )

      Came here to emphasis this. I planned to use my raw data results with Promethease but thanks for letting me know about SNPedia, too!

  • by acroyear ( 5882 ) <jws-slashdot@javaclientcookbook.net> on Friday January 18, 2019 @09:20PM (#57985250) Homepage Journal

    ...are the professional forensic kits that law enforcement use as bad as this?

    • by Mashiki ( 184564 ) <mashiki@gmail.cBALDWINom minus author> on Friday January 18, 2019 @11:09PM (#57985702) Homepage

      ...are the professional forensic kits that law enforcement use as bad as this?

      Yes and no. There's been several labs both owned by police and private that have been caught using this same junk science shotgun approach in policing, leading to retesting 30 years back and people walking out the door. Back a few years ago when this was the hot shit, they only used 10 genetic markers, most have moved to 20-30 markers. Here's a case from NY State [timesunion.com] where multiple people were hit with fake tests, manufacturing DNA tests and so on. There was a huge push by justice dept's for DNA testing vs physical evidence because it was believed to be 100% perfect all the time.

      There's probably more people then you can think of out there these days who are innocent because they were hit with a "common match" because their family has lived in the same area for generations, or because the person doing the testing lied for whatever reason - and shit there are a lot of reasons people lie when they 'certify' a test.

    • by Tony Isaac ( 1301187 ) on Friday January 18, 2019 @11:14PM (#57985724) Homepage

      These are not the same things.

      The story was discussing the labs' reporting of ethnic heritage, not matching two DNA samples against each other to find out if they are the same person's DNA.

      The story itself noted that the RAW DATA was nearly identical, as one would expect with identical twins. It was only the calculation of ethnic background that is (somewhat) in doubt.

    • by jbn-o ( 555068 ) <mail@digitalcitizen.info> on Friday January 18, 2019 @11:20PM (#57985752) Homepage

      Spying is big business. For example, spying is Google's main theme for which services to keep and which to do away with. Perhaps spying is driving these ancestry services as well. We already know these ancestry services share client data with police (1 [rt.com], 2 [rt.com], 3 [ajc.com]). Perhaps this data sharing is listed in the terms of service, but either way the sharing helps authorities augment their database and helps them perform more surveillance on ordinary citizens (most citizens don't commit crimes and therefore should not face such treatment; I'm not convinced those who commit crimes deserve this treatment but the vast majority of the public absolutely don't).

    • by guruevi ( 827432 )

      Yes, but presumably they do further testing when a match is found. DNA tests have never been proven to give unique results, it's statistically unlikely within a certain geographical area, but possible that someone across the globe will have matches across the markers they test for in forensic labs.

    • ...are the professional forensic kits that law enforcement use as bad as this?

      Law enforcement mostly wants to know: Is the DNA that we found at a crime scene identical to the DNA we took from a suspect? That's a much much simpler question, and can be decided clearly. If there is very little DNA, then they can say "the DNA matches, but it will be matching other people as well".

      They can also reasonable well decide that the DNA found at a crime scene belongs to a relative of someone in the system, if there is an unusual high number of markers matching, with a good number of mismatch

  • Some people don't realize that none of these services sequence the full genome. They sequence a collection of different sites as best they can (from the spit that you send them in a tube). Some sites will be sequenced really well and some not at all; it is the random nature of the system. What happens if twin 1 is sequenced really well at site ABC123 and has some rare mutation there but twin 2 is not sequenced at that site at all? They will assume that twin does not have the mutation - they will sub in "wild type" sequence at that locus as they won't have any thing better to go on - and you'll end up identifying them as being different. Take this many times over thousands of gene loci that they sequence and pretty soon you see how two identical twins can end up looking very different.

    If you want to see how similar they really are at the DNA level, you need Whole Genome Sequencing (WGS) or at least you need to know what the coverage was at each locus for each twin. The former still costs thousands in most cases, the latter should be in the raw data (though they would need to convince the companies to release said raw data to them).
    • This. It's a toy, a game. I just wish they didn't advertise it as a scientific test because it isn't.

      Out of curiosity last year I did one of these too. I live in the Midwest so it probably wasn't that surprising that my results came back with a small but noted percentage of Native American DNA. Thing is; I'm British. Grew up in Northern Ireland. My family heritage is pretty well documented going back to the time of the Vikings thanks to my grandmother's work many years ago. Unless there was some undocumente

    • by dAzED1 ( 33635 )
      Sure, but it still means their methodology is crap. If they put personX as heritageA, and personY as heritageB, and personX and personY are identical twins, then it means they either don't take a large enough sample, or...something...to accurately determine anything. That's the real point of the story.
  • It's entirely possible that twins have different fathers. They should have DNA testing done in person at a well respected facility to know for sure.
    • >"It's entirely possible that twins have different fathers"

      But the article is about "identical" twins, not "fraternal" twins. Identical twins cannot have two fathers. It is pretty easy to identify identical twins because.... they look identical. That is not going to happen with fraternal twins.

    • The story doesn't imply this at all.

      Note that the story stated that the two sets of raw data were "nearly identical." It wasn't the data that was the issue, but rather, the calculation of the ethnic background FROM that data.

  • by oldgraybeard ( 2939809 ) on Friday January 18, 2019 @09:45PM (#57985372)
    about my DNA information, except maybe if it is something medical. Have always looked on these DNA services as useless scams.
    I can't help but think of Bill Murry in the movie Stripes! Something about we are Americans, were Mutts. Our forefathers got kicked out of every decent country in Europe. Well something l;ike that it was many years ago.
    But that memory has always stayed with me.

    On a side note one of these services has a commercial on TV that I think is idiotic. It has this smug woman talking about how she travels all over the world. And people are always asking what her nationality is? And she says, I used to say I was Latino, Now after getting my DNA checked I realize I am everything!
    I sit back in my chair and shake my head. They asked her what her nationality is! Not what her ethnicity is! Seems to me a non idiot would say I'm American. But hey what do I know.

    Just my 2 cents ;)
    • After she dies.

      https://en.wikipedia.org/wiki/... [wikipedia.org]

      "FamilySearch is a genealogy organization operated by The Church of Jesus Christ of Latter-day Saints (also known as the LDS Church). ...
      The resource is maintained to support the process of obtaining names and other genealogical information so that Latter-day Saints can perform temple ordinances for their kindred dead.[3] ...
      In February 2014, FamilySearch announced partnerships with Ancestry.com, findmypast and MyHeritage, which includes sharing massive amounts of their databases with those companies. They also have a standing relationship with BillionGraves, in which the photographed and indexed images of graves are both searchable on FamilySearch and are linked to individuals in the family tree.[6]"

      https://en.wikipedia.org/wiki/... [wikipedia.org]...

      "The LDS Church teaches that deceased persons who have not accepted, or had the opportunity to accept, the gospel of Christ in this life will have such opportunity in the afterlife. The belief is that as all must follow Jesus Christ, they must also receive all the ordinances that a living person is expected to receive, including baptism. For this reason, members of the LDS Church are encouraged to research their genealogy. This research is then used as the basis for church performing temple ordinances for as many deceased persons as possible. As a part of these efforts, Mormons have performed temple ordinances on behalf of a number of high-profile people, including the Founding Fathers of the United States,[47][48][49] U.S. Presidents,[47] Pope John Paul II,[50] John Wesley,[47] Christopher Columbus,[47] Adolf Hitler,[51] Joan of Arc,[51] Genghis Khan,[51] Joseph Stalin,[51] and Gautama Buddha.[51] ...
      In February 2012, the issue re-emerged after it was found that the parents of Holocaust survivor and Jewish rights advocate Simon Wiesenthal were added to the genealogical database.[74] Shortly afterward, news stories announced that Anne Frank had been baptized by proxy for the ninth time, at the Santo Domingo Dominican Republic Temple.[75]"

      Mormons are playing the long game of baptizing everyone ever straight into their "free planets for everygod" heaven.
      If that means buying up banks of genetic data one by one... so be it.

    • by dcw3 ( 649211 )

      Certainly that's your choice, and nobody should argue against it.

      In my old age, I've become more interested in my family genealogy, so doing an Ancestry kit about a year ago was very enlightening. My folks were divorced when I was still in diapers, so I knew next to nothing about my father's side since there was no contact. And, what I thought I knew about my mother's side was quickly tossed out the window. I knew her parents had German ancestry, but after getting my results back and spending just a few

  • by dohzer ( 867770 ) on Friday January 18, 2019 @10:21PM (#57985522)

    Did their insurance companies store all 10 results to deny both of them cover in the future?

    • Re: (Score:2, Troll)

      by AHuxley ( 892839 )
      Wonder what the nation wide DNA generational crime search results are like too :)
    • by dcw3 ( 649211 )

      Did their insurance companies store all 10 results to deny both of them cover in the future?

      While I won't attempt to minimize the possibility, has there been any evidence that such a thing has actually occurred?

  • by eatvegetables ( 914186 ) on Friday January 18, 2019 @11:59PM (#57985854)

    Hey, just just bought two nearly identical barometers from crap-o-mart for $5 each. I put them next to each other in the same room. Imagine my horror when I noticed that their respective readings differ by as much as 10%.

    Oh, BTW, 23andme terms of service are clear, at least.

    "The laboratory may not be able to process your sample, and the laboratory process may result in errors ... Even for processing that meets our high standards, a small, unknown fraction of the data generated during the laboratory process may be un-interpretable or incorrect (referred to as "Errors" ..."

    Inexpensive direct to consumer DNA testing companies to not provide nor claim to provide results with statistically insignificant error rates. Don't we all already know this?

    The results are good enough to do fun things like find previously unknown relatives. To date, I've found three second cousins using 23andme. My ancestry information was likely not perfect but was accurate enough for intended purposes.

  • by whoever57 ( 658626 ) on Saturday January 19, 2019 @12:02AM (#57985872) Journal

    The article makes much of one difference: one service shows Eastern European heritage, while the other Balkans.

    But there is not a clear definition of "Eastern Europe" and it may include the Balkans, so this may be not more than a difference in semantics.

  • by Lorens ( 597774 ) on Saturday January 19, 2019 @04:37AM (#57986232) Journal

    I'm reminded of the opposite story: someone forgets their password to the DNA site, and (instead of resetting the password) creates another account, sends in new DNA... and later calls their kid saying that it's incredible, wonderful, this DNA site has found that I have an identical twin somewhere!

  • Sad that Mark chimed in.

    Vast majority of differences are in one point, which is as we all know, naturally expected error if the number of digits presented is constructed correctly, to reflect that natural error.

    As for larger deviations it came from lack of statistics. They probably have a randomizer to present the data with low statistics so the distribution of results represents the fact that statistics is low. I would not do that, but it's a legit technic.

    My experience was laughable with one of the compa

  • What differed was ancestral breakdown, which is a flawed methodology based on correlating patterns with other patterns in a pseudoscientific way.

    The results were the same.

    Ancestral breakdown is not a result, it is a fable. You cannot determine ancestry by country or region through DNA, with the data set that currently exists, and it may not be possible at all. There's no reason to think it is.

  • I met him for the first time yesterday. It said we share 23.5% of our genes (not 25%) Other evidence corroborates it. It's 100% certain. Maybe this is an issue; maybe it isn't, but this "test" seems superficial to me.

No spitting on the Bus! Thank you, The Mgt.

Working...