Slashdot Mirror

← Back to Stories (view on slashdot.org)

Over 30,000 Published Studies Could Be Wrong Due To Contaminated Cells (sciencealert.com)

Posted by BeauHD on from the flawed-data dept.

An anonymous reader quotes a report from Science Alert: Researchers warn that large parts of biomedical science could be invalid due to a cascading history of flawed data in a systemic failure going back decades. A new investigation reveals more than 30,000 published scientific studies could be compromised by their use of misidentified cell lines, owing to so-called immortal cells contaminating other research cultures in the lab. The problem is as serious as it is simple: researchers studying lung cancer publish a new paper, only it turns out the tissue they were actually using in the lab were liver cells. Or what they thought were human cells were mice cells, or vice versa, or something else entirely. If you think that sounds bad, you're right, as it means the findings of each piece of affected research may be flawed, and could even be completely unreliable.

Horback and fellow researcher Willem Halffman wanted to know how extensive the phenomenon of misidentified cell lines really was, so they searched for evidence of what they call "contaminated" scientific literature. Using the research database Web of Science, they looked for scientific articles based on any of the known misidentified cell lines as listed by the International Cell Line Authentication Committee's (ICLAC) Register of Misidentified Cell Lines.There are currently 451 cell lines on this list, and they're not what you think they are -- having been contaminated by other kinds of cells at some point in scientific history. Worse still, they've been unwittingly used in published laboratory research going as far back as the 1950s.

63 of 106 comments (clear)

  1. This is what happens by TheReaperD · · Score: 5, Funny

    This is what happens when you don't let scientist harvest live humans for research... sheesh; and they thought *I* was mad.

    --
    "Be particularly skeptical when presented with evidence confirming what you already believe." -
    1. Re:This is what happens by Z00L00K · · Score: 2

      I'm more curious about how many cases that have been contaminated with the HeLa strain which is one of the most known immortal strains out there. Fun fact is that the scientists have grown about 20 tons of that strain so far - that's a huge woman.

      --
      If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
    2. Re:This is what happens by msauve · · Score: 2

      If you were actually curious, you would have simply checked the linked-to register. Instead, you jumped in to show everyone you've read that book about Henrietta Lacks.

      --
      "National Security is the chief cause of national insecurity." - Celine's First Law
    3. Re:This is what happens by Anonymous Coward · · Score: 1

      Would not make a difference. They switched human lung and human liver cell as well as many other human cells.

    4. Re:This is what happens by LifesABeach · · Score: 1

      All this big ruckus for a simple editing error? From my orientation, Technical Writers,(TW), now have something to do. Update Humanity's knowledge, and tell others about this new discovery. TW's don't want to? It's this, or they can go home and wait for their next gig.

  2. Whew by Kazymyr · · Score: 1, Insightful

    The HL-60 line isn't listed there. That's what I used in my research back in the day.

    --
    I hadn't known there were so many idiots in the world until I started using the Internet -Stanislaw Lem
    1. Re:Whew by Zorpheus · · Score: 1

      Is this the one that contaminated everything? Would be smart to just plan your research on that line, althoughyour colleagues might not like to have it around

    2. Re:Whew by Ungrounded+Lightning · · Score: 2

      Is this [HL-60] the one that contaminated everything? /i?

      I am not a biiochemist, but... That seems to be a human leukemia line.

      Are you maybe thinking of HeLa - the very hardy immortalized cervical cancer cell line that was the first to be successfully grown in bulk?

      I hear there was a model for the progression of cancer that had to be scrapped, because it was really the result of HeLa cells, escaped into laboratory environments, eventually contaminating virtually any cancer cell culture experiment and replacing the intended cells.

      --
      Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
  3. New revolutionary cure for lung cancer discovered! by qbast · · Score: 4, Funny

    *) Effective only if you are a mouse with a liver cancer.

  4. 30000 out of how many by Anonymous Coward · · Score: 2, Interesting

    What percentage of published studies are affected? Sure, 30,000 seems like a ton, and if there is critical work in there it is certainly bad, but if this is 30,000 out of 300 million or some arbitrarily large number of studies, it isn't as catastrophic as the headline suggests.

    1. Re:30000 out of how many by Anonymous Coward · · Score: 5, Informative

      0,8% - the paper is open access if you want the details.
      http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0186281

    2. Re:30000 out of how many by geekmux · · Score: 2

      0,8% - the paper is open access if you want the details. http://journals.plos.org/ploso...

      What percentage of the population is affected by that seemingly small impact?

      How many billions (or trillions) in costs are associated with that seemingly small impact?

      Risk mitigation relies on asking the right questions. Unfortunately, TFA starts to answer my questions. Oncology is the field most contaminated by a large margin. I'd say it's pretty damn important to understand just how fucked our studies are related to one of mankinds most pervasive killers. Cancer affects a hell of a lot more than 0.8% of the population.

    3. Re:30000 out of how many by Gilgaron · · Score: 2

      It is important to note that immortalized cell lines are essentially cancer cells in culture. So while using the wrong line due to contamination hurts the direct usefulness of research, it doesn't render that research useless, because the data is still about whether or not the compound of interest killed cancer cells... just different ones than you thought. This sort of thing could be the reason that so many treatments that seem promising in academia fail to pan out when they move to industry with more funding for quality control.

    4. Re:30000 out of how many by jedidiah · · Score: 1

      You kind hit at the problem in your own answer. I don't give a f*ck if MY wonder drug cures some other cancer. I want it to cure MY particular cancer. If the basic lab research was botched then there will be some very disappointed (and desperate) cancer patients when it comes time for the human trials.

      ALL of that is going to cost a big pile of money and make everything more expensive. The successes will have to subsidize the failures (as they always do).

      You will end up having to pay a lot more, or your insurance company will, or your government health care entity will just say "fuck it" and leave you to die because the meds are too expensive.

      --
      A Pirate and a Puritan look the same on a balance sheet.
    5. Re:30000 out of how many by Gilgaron · · Score: 1

      Oh it'd go through a bunch of stuff after academia before it got put it you, unless it'd be on an experimental protocol where "oops, he died" isn't so bad an outcome. These sorts of mistakes are impacting basic research, not third party safety and efficacy testing or manufacturing. FDA GLP/GMP inspections are no joke, people make livings just being consultants for telling companies how not to get their asses handed to them. New cancer treatments don't come directly out of some pottering professor's lab where the hung over grad students throw something together in the centrifuge after sacrificing a few mice.

    6. Re:30000 out of how many by geekmux · · Score: 1

      FDA GLP/GMP inspections are no joke...

      When validating basic research turns into a joke creating 30,000+ contaminated documents, other vetting processes tend to become rather irrelevant.

      ...New cancer treatments don't come directly out of some pottering professor's lab where the hung over grad students throw something together in the centrifuge after sacrificing a few mice.

      After finding 30,000+ mistakes, they might as well have.

    7. Re:30000 out of how many by geekmux · · Score: 1

      You will end up having to pay a lot more, or your insurance company will, or your government health care entity will just say "fuck it" and leave you to die because the meds are too expensive.

      Looking for the answer in the Medical Industrial Complex fueled by relentless greed?

      D) All the above

    8. Re:30000 out of how many by Gilgaron · · Score: 1

      Basic research is the scientific equivalent of brainstorming, or throwing shit at the wall and seeing what sticks. This sort of contamination problem means you have a higher failure rate at the next step of development, not that vetting and validation is irrelevant. It is worth solving because we're missing out on potential solutions and spending more grant money than we should for the data we're getting out of it, but your views appear to lack perspective.

    9. Re:30000 out of how many by parkinglot777 · · Score: 2

      It is important to note that immortalized cell lines are essentially cancer cells in culture. So while using the wrong line due to contamination hurts the direct usefulness of research, it doesn't render that research useless, because the data is still about whether or not the compound of interest killed cancer cells... just different ones than you thought. This sort of thing could be the reason that so many treatments that seem promising in academia fail to pan out when they move to industry with more funding for quality control.

      I completely agree with your point. Even though the cells in researches were contaminated, it doesn't render those researches completely useless as long as they can identified what cells they were using in their research. If they could identify the cells they used in their research, then they just need to update the conclusion of their research.

      However, the nature of science researches is to cite/refer to previous research results. That said, some (if not most) contaminated cell researches could become less useful to useless if the research is citing/using other researches that conducted on contaminated cells. In other words, whatever they have done in their research may not need to be done again because it might have been known/done already which results in redundancy. It is also very difficult and could be costly to validate those researches. As a result, dropping those researches and start from scratch may be cheaper (or a better way).

    10. Re:30000 out of how many by Hognoxious · · Score: 1

      Before it was used clinically there would have been in vivo trials, no? I doubt anybody died from this.

      --
      Confucius say, "Find worm in apple - bad. Find half a worm - worse."
    11. Re:30000 out of how many by Gilgaron · · Score: 2

      Yes, exactly. Compromised basic research leads to failed follow up experiments and shrugs all around as they move on to the next promising compound. It is a waste of resources and increases the background noise in looking for useful things, but not a danger to the public.

    12. Re:30000 out of how many by evilRhino · · Score: 1

      Don't worry too much. The failure results will be obfuscated by a paywall and/or HIPAA protection so you don't have to trouble yourself with them. The pharma company gets paid for your chemo round whether it works or not.

  5. The coffee conundrum by Wholehawg · · Score: 2

    Is this why they report that coffee is bad for you one week and good for you the next?

    1. Re:The coffee conundrum by DontBeAMoran · · Score: 4, Funny

      No, that's just the centuries-old fight between coffee-drinking scientists and tea-drinking scientists. There's also a third party of insane scientists who advocate drinking dihydrogen monoxide but they're a minority so you rarely hear from them.

      --
      #DeleteFacebook
    2. Re:The coffee conundrum by Headw1nd · · Score: 2

      They probably died, as anyone who drinks dihydrogen monoxide inevitably does.

    3. Re:The coffee conundrum by Nite_Hawk · · Score: 1

      That's sort of the whole point of the article though. The author discovered evidence that (amonst other things) historical research findings on tea and coffee related fatalities may have used samples with dihydrogen monoxide contamination that previously went unnoticed all the way back to the 1950s. while correlation does not equal causation, there's a strong body of evidence that anyone who ingests large quantities of dihydrogen monoxide eventually dies.

      How much dihydrogen monoxide contamination did the samples have? Are there specific interactions (say with tea and coffee respectively) that have unique and unforeseen effects? Hopefully this new research will help us better refine and answer those questions.

    4. Re:The coffee conundrum by DontBeAMoran · · Score: 1

      Now don't go and quote me on this, but I've heard that even the temperature of said dihydrogen monoxide can skew the results, by a lot.

      --
      #DeleteFacebook
  6. Henrietta by 110010001000 · · Score: 4, Interesting

    This is Henrietta's revenge. That is what scientists get for stealing her cells in the first place!

  7. Time to get Silicon Valley involved by Applehu+Akbar · · Score: 1

    This is a problem that could be addressed by improved medical databases. Of all the ways in which SV could 'disrupt' medical research and practice, this would be the least controversial.

    1. Re:Time to get Silicon Valley involved by Bing+Tsher+E · · Score: 1

      Yes, but what is the reimbursement model for that!?! It doesn't involve microtransactions, nor push notification.

      I suppose it could.... hmmm....

    2. Re:Time to get Silicon Valley involved by DontBeAMoran · · Score: 2

      Improved medical databases? Blockchain!

      --
      #DeleteFacebook
    3. Re:Time to get Silicon Valley involved by Goldsmith · · Score: 1

      Here, they looked at "research" results. As soon as you have "medical" results, you're constrained by law, not technology. Disrupting laws sometimes works (Uber), but usually not (Theranos).

    4. Re:Time to get Silicon Valley involved by WrongMonkey · · Score: 1

      There are a lot of laboratory information management systems (LIMS) that purport to address issues such as this. But many (most?) research labs don't use them because they are too expensive. Either in terms of licensing proprietary software or managing open source solutions.

  8. I don't believe this for a second by SlaveToTheGrind · · Score: 1

    This is exactly the sort of thing that would have been caught during the rigorous, diligent, inherently skeptical peer review process.

    But seriously, no wonder most studies can't be replicated by others -- the odds are high that either the cells in the original study, the attempted follow-on, or both were screwed up.

    1. Re:I don't believe this for a second by SlaveToTheGrind · · Score: 1

      * whoosh *

    2. Re:I don't believe this for a second by serviscope_minor · · Score: 2

      This is exactly the sort of thing that would have been caught during the rigorous, diligent, inherently skeptical peer review process.

      erm... while you seem to be taking a dig at peer review, and I'll happily bend your ear for ages over the flaws of it, this isn't really one of them.

      Peer review even at its best is basically a test of "seems legit" before publishing to avoid obvious sloppy mistakes, bad stats, unwarranted conclusions. If someone says "we did X, tested hyothesis Y, got results Z and rejected the null hypothesis (p 0.5)", peer review will check if the reaults are plausible givne the experiements, if the boldness of the claims are consistent with the level of evidence and so on and so forth.

      Peer review is an indication that something is not obviously bogus. What hapens afer peer review (replication etc) is what eventually catches the stuff that (a) anyone cares about and (b) is subtly bogus. Something isn't generally considered a hard fact until it's been replicated. Scientists generally know this...

      --
      SJW n. One who posts facts.
    3. Re:I don't believe this for a second by SlaveToTheGrind · · Score: 1

      Peer review even at its best is basically a test of "seems legit" before publishing to avoid obvious sloppy mistakes, bad stats, unwarranted conclusions. . . . Peer review is an indication that something is not obviously bogus.

      I think you're talking about the idealized version of what peer review was originally intended to be. As we're all aware, over the past few decades the mantra "peer reviewed articles in highly regarded journals" has become one of the prominent measuring sticks used to present whatever topic is currently catching grant dollars as "settled" and try to shut down debate. That's pretty much the polar opposite of "not obviously bogus," and that's what I was taking a dig at.

  9. What a relief by rastos1 · · Score: 1, Offtopic

    TFA makes me much less worried about the crappy code I write. I mean, reading /. for years taught me to be ashamed about writing sloppy code, and ignoring best practices in SW development and security. It got me in trouble for raising a hell for using obsolete/EOL-ed development tools/languages/OS/, etc etc. But now ? I feel relief. There are people out there that are doing much more important work than me, they fuck it up and it still goes on for decades. I can sleep well. I have nothing to worry about anymore.

    1. Re:What a relief by Bing+Tsher+E · · Score: 1

      Aren't you jealous of their tenure track, though?

  10. Makes me happy... by VitrosChemistryAnaly · · Score: 1

    Makes me happy to know that we use primary cells in the research in my lab rather than cell lines.

    Still, I wonder how many primary cell lines become contaminated with immortalized in labs where both are cultured in the same space?

    --
    "It's a tarp!" -- Dyslexic Admiral Ackbar
    1. Re:Makes me happy... by Gilgaron · · Score: 2

      Academic labs aren't really known for following strict protocol, so I imagine quite a bit of cross contamination occurs. That'd be among the reasons to limit how many times you passage the cells before going back to the LN2 tank for more. Still, it isn't like human cells live very long on a stainless steel BSC surface so it wouldn't take too much effort to drop the contamination rate.

  11. Re:Is this why by DontBeAMoran · · Score: 1

    https://www.youtube.com/watch?...

    --
    #DeleteFacebook
  12. Looks like Firesign Theater was right: by Cornwallis · · Score: 1

    Everything You Know Is Wrong.

    1. Re:Looks like Firesign Theater was right: by The-Ixian · · Score: 1

      Black is white, up is down and short is long.

      --
      My eyes reflect the stars and a smile lights up my face.
    2. Re:Looks like Firesign Theater was right: by Talderas · · Score: 1

      His hibachi dealer told him.

      --
      "Lack of speed can be overcome. In the worst case by patience." --Znork
  13. This is beyond hard to believe by omfglearntoplay · · Score: 1

    Can someone who knows more about this process comment how this could possibly be? It smells so much like somebody is feeding us fake information. First off, if this has been a problem since the 1950s, why is it suddenly news that makes it sound like nobody knew of any problems till yesterday? Or is it more like every year we find some bad cells and we need to go back and redo those experiments... and someone decided to add up all problems for the last 100 years and make a write up about it?

    1. Re:This is beyond hard to believe by Gilgaron · · Score: 2

      It is easier to test now than it used to be. You wouldn't be able to tell them apart visually with a microscope, for example. It has been a known issue for a while now, the scope is still being sussed out.

    2. Re:This is beyond hard to believe by drunken_boxer777 · · Score: 4, Interesting

      This has been known for a while, but perhaps not broadly known, and not by "the general public". The link to the ICLAC in TFS is evidence of that.

      The actual paper will answer your questions, but briefly: people make mistakes in maintaining cell lines, and contamination is easier than you think, particularly in primary cell lines.

      I didn't see the authors mention if reproducibility sorts this out, if someone can't reproduce the results in another cell line or in an animal model, the original results are considered context dependent.

    3. Re:This is beyond hard to believe by pesho · · Score: 5, Informative

      It is a problem that has been known for a long time. Initially we lacked tests to validate the origin of the cell lines. Now we have an established panel of markers that can be used to cheaply and reliably confirm the identity of cell lines. The National Institute of Health, which is the major funding source for biomedical research in the US, requires all funding applications to have a plan for authenticating biological materials including cell lines.

      Most of the literature using misidentified cell lines is probably old, although there are still people doing research who are either oblivious to the issue or just don't care. I think the conclusion that all these studies are invalid is an overstatement. Many if not most of these works are likely investigating fundamental biological processes that would be the same regardless of the cell line. The studies that are questionable would be the ones relying on the cell line fatefully preserving the characteristics of the original cell, like studies trying to develop therapies for various diseases or investigating processes carried out by specialized cell types.

      There are many other problems that are associated with cell lines that I would think are more serious than the mis-authentication. For starters cells change when they are placed in the dish and loose many of the important characteristics of the originating cell. This means people need to be really careful when deciding if a particular experiment can be done on a cell line. Then there are examples of low level microbial contamination that goes unnoticed by the people growing the cells, but can clearly be detected in gene expression data if you look for it (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4239086/ if you look at figure 2 of the paper you can see that some labs are consistently sloppy). There is also a host of technical issues that can impact the conclusions of cell line studies ranging from the quality and source of reagents, to the experience of the staff and the techniques used to maintain the cells. These tend to vary a lot across labs and rarely documented in the publications.

    4. Re:This is beyond hard to believe by Guybrush_T · · Score: 1

      It's not that hard to believe. In every field, and computer science is one as well, many scientific papers are completely wrong with their analysis, and most of the time, it is pretty hard to detect unless you're an absolute expert in the field. Peer-review is supposed to detect bad papers, but finding experts is not always easy, and sometimes everything you read in the paper seems consistent, so unless you try to reproduce it, you can only say "that looks like good work".

      Now, that's not the end of the world. Real discoveries that matter will be replicated and used as a base to develop products that will be effective, whereas crappy stuff will just be ignored.

  14. Re:I call schenanigans by omnichad · · Score: 2

    On step #4, they may be using the same suppliers. How many suppliers are there for experimental cell lines?

  15. Classic. by Qbertino · · Score: 2

    This is classic.
    The bane of modern utopia.
    Efficiency and fragility are directly correlated.
    This goes for any system and society.

    If there is one thing that has a large chance of being modern societies demise, it is this.
    Scary, if you think about it.

    Just imagine: One replenishing bioculture that goes back some decades turns out to be labled wrong and all of a sudden countless biological studies are beyond worthless.

    Long story short: Do not over-optimize. And question the status-quo once in a while. Especially with systems that seem to run flawlessly indefinitely.

    --
    We suffer more in our imagination than in reality. - Seneca
  16. Re:I call schenanigans by Ungrounded+Lightning · · Score: 3, Insightful

    4. Other scientist recreate the process that proved or disproves the theory

    4.5. Other scientists buy cells from the same misidentified cell line to use in their "repeatability" study. ...

    6. Rinse and repeat.

    --
    Bantam Dominique roosters crow a four-note song. Once you've heard it as "Happy BIRTHday" you can't NOT hear it that way
  17. Re:I call schenanigans by pesho · · Score: 1

    2 or 3 that I can think of. In the US the major one is ATCC. There are known cases where the cell line was mixed up before it was deposited in ATCC. Consequently everyone who got it from there was working with the wrong cell line. It is also common practice for people to borrow cell lines from the lab next door, rather than buy them for a supplier. My guess is that most of the cell lines used for research were not directly purchased from a supplier. Even if you did, it is not hard for careless worker to mix up cell lines in the lab - mislabeled tubes, handling many cell lines simultaneously while being distracted, forgetting to swap the dirty pipette ...

  18. Wouldn't be a problem... by Archtech · · Score: 2

    ... if all studies were replicated at least twice by other teams in other institutions (and preferably funded by different sponsors).

    --
    I am sure that there are many other solipsists out there.
    1. Re:Wouldn't be a problem... by WrongMonkey · · Score: 2

      Every working scientist would love to see this happen. Who's going to pay for it?

    2. Re:Wouldn't be a problem... by SlaveToTheGrind · · Score: 1

      Every working scientist would love to see this happen.

      Even if it compromises (perhaps significantly) their publication rate / bonuses / career path? I'd love to see a survey or two on that.

    3. Re:Wouldn't be a problem... by Gilgaron · · Score: 1

      I think the scientists would like to see the publish or perish schema go away as well...

    4. Re:Wouldn't be a problem... by SlaveToTheGrind · · Score: 1

      I think the scientists would like to see the publish or perish schema go away as well...

      Helpless victims one and all, I'm sure.

    5. Re:Wouldn't be a problem... by WrongMonkey · · Score: 1

      Who cares about surveys? Fund some grants to reproduce results. I guarantee there will be takers.

    6. Re:Wouldn't be a problem... by iggymanz · · Score: 1

      and they both order their tissue cultures from the same place? or from different suppliers selling the same wrong one?

    7. Re:Wouldn't be a problem... by serviscope_minor · · Score: 1

      Who cares about surveys? Fund some grants to reproduce results. I guarantee there will be takers.

      Not that many. You will never get great publications merely replicating results. Sure it's a grant now, but at the penalty of your future career.

      --
      SJW n. One who posts facts.
  19. Re:GNAA - GAY NIGGERS ASSOCIATION OF AMERICA by Tesen · · Score: 1

    Bannon... is that you?!?!