My lab is pretty ‘old school’. We do basic biochemical procedures– cloning/subcloning, PCR/RT-PCR/RT-PCR/RTRT-PCR, pulse-chases, western blots. My flow cytometry is about as crazy as we get.

We also do those ‘old school’ procedures the ‘old school’ way. We dont use Qiagen kits to extract our plasmids, we do cesium chloride gradients. We also make sure to use real film to visualize our radioactive pulse-chase experiments and our western blots. Oh we also do it the ‘new school’ way and take pictures electronically, but there are real benefits to keeping film in your lab notebook: You have a defense if someone accuses you of fraud.

Larry pointed out a case last week, where a scientist had falsified her data with Photoshop.

“As we looked at it, we realized the person had cut and pasted the exact same bands” over and over again, says Ushma S. Neill, the journal’s executive editor. In some cases a copied part of the image had been flipped or reversed to make it look like a new finding. “The closer we took a look, the more we were convinced that the data had been fabricated or manipulated in order to support the conclusions.”

Experts say that many young researchers may not even realize that tampering with their images is inappropriate. After all, people now commonly alter digital snapshots to take red out of eyes, so why not clean up a protein image in Photoshop to make it clearer?

“This is one of the dirty little secrets–that everybody massages the data like this,” says Mr. Farid. Yet changing some pixels for the sake of “clarity” can actually change an image’s scientific meaning.

Our lab uses Photoshop to crop and resize images of our gels. The gels I run are 18×16 cm– you cant fit them in a journal. But I sure as hell dont ever airbrush them, or cut/paste ‘pretty’ bands over ugly ones. I dont touch shit except the sizes.

But why should anyone believe me? “Everybody massages the data like this”, right?

Luckily, if anyone ever calls BS on my data (or challenges me to a dual), Ive got my ‘old school’, unPhotoshoppable film sitting in my lab book. Im never whining at Bossman for being ‘old school’ ever again.

Comments

  1. #1 PlausibleAccuracy
    June 23, 2008

    Anything can be faked. There will always be ethically challenged people.

    Still, I prefer a real pen & paper chart recorder over some locked-down proprietary bit of software. There is just something about watching the plotter pen move and scribbling notes on the paper as it comes off that feels like SCIENCE

  2. #2 Mike O'Risal
    June 23, 2008

    My lab is pretty old school, too. Naive me, though, I hadn’t even considered that people would do this. I also capture images and edit in Photoshop, but my editing is nothing more than resizing and bumping up contrast to make the images clearer. Oh, and I stick lane numbers on them with Photoshop, too. Examples are posted in my blog even now; they look exactly like the printed versions in my lab notebook.

    But my gels are so pretty to begin with that altering them in any substantial way could only ruin the whole esthetic!

    Seriously… I hadn’t even considered that people might be doing anything substantial like this. Such people should themselves be forced through mile-long stretches of 1% agarose.

  3. #3 Foo
    June 23, 2008

    You could always take pics using the RAW format of your digital camera and keep those as “evidence”. That’s as close to film as possible, I think, when it comes to digital shots.

  4. #4 minimalist
    June 23, 2008

    Useless without pictures. I always want to see pictures of these things, if only for the schadenfreude factor.

    A scientist working in the lab next door once pointed out to me where a researcher had recycled four entire figures from a previous publication. The dirtbag even had the temerity to list himself as sole author on the recycled paper, when there were at least three other names (including the senior author) on the paper he’s stolen the images from.

    To this day I have no idea how that got past peer review (who doesn’t at least GLANCE at the author’s past publications?) or if he was beaten to death by his irate former lab.

  5. #5 Coleran
    June 23, 2008

    There are abuses possible of almost any tool and what is required is smart and educated use of the tools available.

    It should be noted that Photoshop actually has a forensic history tracking function built into it, that I would have expected anyone who has to deal with authenticity and reliability of data would know about. At least you do now.

    There should be no problem with digital imagery as long as people take the extra step of authenticating and tracking that data properly.

  6. #6 ctenotrish
    June 23, 2008

    In my former lab, we A) kept the original, unretouched .tif files from each and every image we collected, and B) kept all the original stained slides (all thin tissue sections for RNA gene expression) with coordinates that matched the image files, and C) offered our raw data to all reviewers. No reviewers ever took us up on that, but I can go back to every image I used in my dissertation and the resulting publications and pull the original data. Similar method of keeping raw images for our gels, though we obviously didn’t keep the gels. Keep your raw data!

  7. #7 Jason Dick
    June 23, 2008

    Wow. Just wow. I work in cosmology, and I would never think to even suggest hand-cleaning any of our maps. Not only that, but we make sure to keep all of the original data around that was used to make the maps in the first place, as well as document all of our cleaning operations.

    It takes a special kind of moron to think it’s “okay” to just touch up a picture that is used for scientific results without recording exactly what you’re doing. And even recording your operations, it would be more than a little irresponsible to fail to keep the original data around.

  8. #8 JimNorth
    June 23, 2008

    You call that “old school”? Cloning, PCR, RT-PCR, Flow Cytometry, Western Blotting – old school?

    Piffle.

    When you have to synthesize oligos by hand (syringe, actually), now that’s old school. Heck, our lab didn’t even have running water – unless you emptied the bucket into a wash basin. And we didn’t have ceilings, for that matter.

    NPR just aired a piece about scientific fraud recently which makes me proud to be a scientist. The number of cases is relatively low.(http://www.npr.org/templates/story/story.php?storyId=91658116&ft=1&f=0)

  9. #9 PhysioProf
    June 23, 2008

    When you have to synthesize oligos by hand (syringe, actually), now that’s old school. Heck, our lab didn’t even have running water – unless you emptied the bucket into a wash basin. And we didn’t have ceilings, for that matter.

    When I was in grad school, our lab was in a fucking tent! And we had our own nuclear reactor to make radioisotopes for labeling nucleic acids! And we mined our own uranium to put in the reactor!

    LOLZ!!

  10. #10 Colin M
    June 23, 2008

    This is a great resource for “digital imaging ethics”:

    http://swehsc.pharmacy.arizona.edu/exppath/micro/digimage_ethics.html

    They have a 6-page PDF containing some guidelines and explanations intended for scientific journals, especially focused on biology. The authors point out that photographs were manipulated to misrepresent scientific data long before Photoshop was invented; however Photoshop sure made it easier.

  11. #11 efrique
    June 23, 2008

    Digital cameras already manipulate the information coming in through the lens. If you’re publishing the digital images from your camera, you’re publishing a manipulated image. Whose to say the results aren’t already biased in certain directions?

    As Foo points out above, the RAW image is as close as you can get to unmanipulated. Who makes that available?

  12. #12 efrique
    June 23, 2008

    gah. “…who’s to say…”. What a dill.

  13. #13 sparc
    June 23, 2008

    Digital cameras already manipulate the information coming in through the lens.

    Maybe, but data are already manipulated before any light gets into the camera.
    Conventional films also influence data since their linear range is quite narrow. E.g., you may have to pre flash X-ray films or vary exposure time to make things visible. And taking immunofluorescence pictures with a conventional camera at a microscope was a real pain and to a degree biased too since everyone choose those pictures that fitted best with what he wanted to show. I beleave this is fine as long as beside just keeping original data on films records of how these data have been generated (exposure time, filter sets used, camera sttings) are kept. Without these information it is rather unlikely that someone could independently redo the experiment.

  14. #14 The Chemist
    June 24, 2008

    Let’s face it, nothing will ever be better proof of honesty than an old-school permanently bound lab notebook with every line filled in pen and every mistake initialed.

    These will never last. Nevaahr! (Ok, well maybe they will, but I’ll refuse to acknowledge it.)

  15. #15 Brian X
    June 24, 2008

    Colin M:

    And for those who can’t afford Photoshop for their skullduggery, there’s GIMP and, for the oldschool, NIH Image…

    I must follow that WP:BEANS comment with the following joke:

    A man decides to fight a moving violation where a cop got him for running a stop sign. He goes to court and provides the judge with a series of photographs indicating that there is no stop sign at the corner where he was stopped.

    The judge says, “This is all very persuasive,” then looks at a form in front of him. “What did you say you did for a living?”

    The man mumbles… “Photo retoucher…”

    “That’ll be $200, please.”

  16. #16 Zarquon
    June 24, 2008

    Photo prints can be manipulated as well, so unless you’re using glass plates…

  17. #17 Confused
    June 24, 2008

    If you really want to fake a gel, you can fake a gel. I mean, you can’t tell from a film what was in the sample, or which antibody was used – all of that you have to rely on the persons record, which may or may not have been accurate. There’s also a whole bunch of ways you can manipulate your loading controls…

    Copy/paste manipulations are actually *really* easy to detect if you know how (push contrast and brightness in photoshop). It’s a little harder if you’re using some of the more advanced photoshop functions (although most people don’t), but there are ways and means to do that too. However, the only journal that does it systematically is JBC (or possibly JCB – I get those mixed up). They came and gave a talk on it.

    I spotted on manipulated gel in a paper I was doing as a journal club. I’ll see if I can dig it up.

    Coleran@5 – I’ll place money on that forensic history not being saved in the tiff files you send to publication.

  18. #18 Confused
    June 24, 2008

    Got it -

    as published…
    …but with a little contrast adjustment…

    From Carson et al 2006
    Developmental Dynamics
    Volume 235, Issue 6, June 2006, Pages 1678-1688

    Bit of a gel added, bit of a gel taken away. Also some gel splicing, which is technically okay (although some journals insist on using obvious lines to show the joins)

  19. #19 Joe
    June 24, 2008

    (or challenges me to a dual)

    While you might be making some really deep comment that has gone over my head, I think you meant duel there. Unless you are a pair of twins, I’m not sure how you could be challenged to a dual (or heck, maybe your cloning work has advanced a lot)…

  20. #20 Stephen
    June 24, 2008

    Digital cameras already manipulate the information coming in through the lens. If you’re publishing the digital images from your camera, you’re publishing a manipulated image.

    There’s an awfully big difference between automatic transformations which affect the whole frame equally and manual manipulations of a portion of the frame.

  21. #21 midwifetoad
    June 24, 2008

    The newer digital SLRs can save an image simultaneously as jpg and raw. It would be pretty difficult to apply touch ups to both in exactly the same way. At any rate, forensic investigators can detect changes from the original camera image. No editing program shades pixels exactly the way cameras do.

    If you are truly dishonest, you can edit a digital image and then convert it to film. Some of these hypotheticals are ridiculous unless the payoff is CIA important.

  22. #22 Jason Dick
    June 24, 2008

    There’s an awfully big difference between automatic transformations which affect the whole frame equally and manual manipulations of a portion of the frame.

    Definitely. Especially as you can, in principle, model any changes the camera makes to the image. This is the sort of stuff we have to do in astronomy all the time. Either way, deliberate manipulation of data without recording your exact procedures is just plain wrong.

  23. #23 MH
    June 25, 2008

    When I was in grad school, our lab was in a fucking tent! And we had our own nuclear reactor to make radioisotopes for labeling nucleic acids! And we mined our own uranium to put in the reactor!

    Luxury! Our lab was a hole in the ground, covered with a sheet of algae, and we had to get our uranium from filtering sea-water with our teeth.

    In case no-one has seen the Four Yorkshire-men sketch, this is the first and best version, IMO.

  24. #24 damitall
    June 27, 2008

    Primitives.

    We wander into our lab about 11, tell the voice control unit what experiments to do and what results are wanted. After a deal of clicking and whirring, the thoroughly manipulated data and images are ready for the paper-writing software. (The automatic quill pens are writing in the lab books, and adding coffee spills, along with traces of very nearly real blood, sweat and tears). By the second coffee of the afternoon, the first draft is ready.

    Complete with absent and misplaced apostrophes for verisimilitude.

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!