The Genomic Repairman is currently a Ph.D. student who escaped from the deep south, and studies DNA damage and repair through biochemical and genetic approaches. He intends to use pine away about his scientific interests and rant about the things (and there are lots of them) that annoy him.
My posts are presented as opinion and commentary and do not represent the views of LabSpaces Productions, LLC, my employer, or my educational institution.
Please wait while my tweets load
So the other day I was going through a field specific journal and found an article from a European group that was sort of interesting to me but outside of my field. I perused their data and was interested in their siRNA knockdown data and was reading further and decided to pull up a previous paper that they published back in 2008. Low and behold the siRNA data is their too! And its the same exact western blot looking at protein levels and loading control. I pulled the images and redacted the data to let you guys take a look at this.
They don't make mention that this is from the 2008 paper and try to pass it off as original data in the 2010 paper published in a different journal. I think that is the kicker too, I'm sure the original journal would be none to pleased that this EuroGroup was doing this. So what do you guys think, is this misconduct, is it okay? My own opinion is that this shit is pretty shady.
This post has been viewed: 1657 time(s)
My understanding is, this is a control, just to show their KD worked. Is it possible that they did this KD, plated some of the cells for an experiment that ended up in the 2008 paper and some for another experiment that ended up in the 2010 paper, and harvested some for the Western blot? Or is this out of the question, just on the grounds that they are a "EuroLab"?
When I do experiments with knockdown cells, I take a sample at that time of the experiment for westerns to show proteins levels. Also this was siRNA knockdowns so its not like it was a stable line and you do the western for every time you transfect, not just the first time. They were using knockdowns to study modification of another protein in cells, so my thought process is, every time you do the experiment, you do the western.
"Or is this out of the question, just on the grounds that they are a "EuroLab"?"
And I have no problem with them being a EuroLab, I have collaborators that I work with in Europe and elsewhere in the world. So its kind of tiny-minded that you think I apply different standards to scientists based upon where they are geographically located.
Just to clarify, my original suggestion was that the authors knocked down a gene in a batch of cells, performed two experiments (perhaps more) using cells from that batch and verified the knockdown once for the entire batch. One of said experiments ended up in a 2008 paper and another in a 2010 paper. Seems perfectly acceptable to me. Please explain what I'm missing.
Unless you tell us that the figures are labeled as different cell lines or siRNA sequences in the two papers, I say there is no reason to assume misconduct.
You're missing the fact that the gels are the SAME. If they did what you said, no one would have noticed a difference because bands never resolve EXACTLY the same way. They do here so you know for a fact that data has been re-used. I'm all for doing things in batch, but since this is science, I also put a high value on unique experimental replicates...
Brian Krueger, PhD said:
You're missing the fact that the gels are the SAME.
I am missing nothing. I'm saying it's the same picture of the same gel. You seem not to understand what the role of this figure is in the experiment. I'll try to explain. This figure is not the experiment. Experiments such as this are performed to explore genetic dependencies of biological phenomena. This genetic approach consists in preventing a certain gene from being expressed and observing the phenotype. Whatever does not/cannot happen when a gene product is not present is what that gene product does when it is present. This is genetics.
Now, preventing a gene from being expressed can be achieved in a number of ways, one of which is to knock it down using siRNA. I won't go into the details of the method, I'll just mention that it's temporary and reversible.
Now, to the figure: These guys performed an experiment (or experiments), about which we know very little (DNA Repairman knows more), that required knocking down a gene. They show this figure not as the result of their experiment, but rather as evidence that the gene was indeed knocked down, and that their experiment(s) was/were performed in the absence of the product of that gene (or rather in the presence of very little of it, as there is a faint band visible in the KD lane).
Given this background, I'll repeat my explanation: Experimenter decides to perform two or more experiments in a given genetic background. To achieve the desired background, they introduce siRNA of a certain sequence into a batch of cells. The knockdown status of all the cells in the batch is identical, agreed? Now the experimenter seeds some of the cells from the batch onto some plates and performs experiment A. They seed other cells of the same batch onto other plates and perform experiment B. Finally, they harvest some cells *of the same batch* and do a Western blot (that's the figure under discussion) to see whether the gene product is present in the cells or not. Same Western blot, two experiments. I hope this is clearer now, sorry I had to be so verbose to (hopefully) get the point across.
If I'm pulling a batch sample out of freezer to study modification of a protein in that particular genetic background, I'm going to check my knockdown again if I'm trying to interpret a new piece of data from the sample. But that is a moot point, the point is they reused a figure. They could have referenced back to their previous journal but instead they actually committed copyright infringement by lifting a portion of a figure from one journal and putting it into their new journal.
@Unamerican, So you don't try to replicate your experiments? You think doing something just once and copy and pasting that into every paper you publish is rigourous science? I'm just trying to understand why you're defending this level of asshattery.
I never said anything about pulling a sample out of a freezer. I'm not going to repeat the same scenario over and over again. It's not that complicated, and it's here in writing for anyone to read and judge for themselves.
There may be an issue of copyright infringement here, but that has nothing to do with academic misconduct.
Brian Krueger, PhD said:
@Unamerican, So you don't try to replicate your experiments? You think doing something just once and copy and pasting that into every paper you publish is rigourous science?
What's replicating gotta do with anything? They may have repeated the entire procedure three times for each experiment, as is the norm, but they'd still only show one of the three figures in the paper. I don't see why it's important to show one repeat of the supporting figure in one paper and a different repeat of the same in another paper, if the same supporting figure applies to both experiments, as outlined above.
I think DNA Repairman should go ahead and complain to the Journal. I'm fairly confident the group can provide scans of lab notes showing that what happened is precisely what I described above.
If I do a western blot to look at protein levels of something else, I'm always going to check my knockdown. Every time. I mean that is what good rigorous science is. If you disagree, then by all means do. I'll hold myself to my standards and you can hold yourself to yours. And I'm pretty sure this may qualify for the tag of FWDAOTI when we can apply those to our blogs here.
Genomic Repairman said:
If I do a western blot to look at protein levels of something else, I'm always going to check my knockdown. Every time. I mean that is what good rigorous science is.
But this isn't what good, rigorous reading comprehension is. If I write again that these experiments (in my scenario) were performed side by side, as in AT THE SAME TIME, will you finally get it? Otherwise, I give up. I really don't know how else to state this.
unamerican: I understand your point, but, the papers are two years apart so hopefully in two years, they would have repeated the control experiment and would of had a new figure. Two years later, they'll be working with new lots of siRNAs, new cells, and antibodies.
This control should be done whenever new reagents are used.
What we don't know is if in this new paper, they used the same cell lines or siRNAs and are trying to pass off this knock down experiment as original.
They could have just referred to the previous paper if they were not trying to pass it off as something else.
The authors of a paper I was reviewing did exactly this same thing. They re-used a figure from their last paper, which was a cloning experiment, and tried to pass it off as the results of a new cloning experiment. Exact same figure, but different genes.
They probably submitted their paper somewhere else where they won't be caught.
Anyway- back to your point, it is possible that all the experiments for the 2008 and 2010 paper were all actually done in 2007 and then they published some in 2008 and then the rest in 2010, but this seems unlikely. As mentioned, why not just reference the previous paper if the experiment is the same? Why pass it off as new?
GR- in the two papers- are they using the same siRNA and same cells and same antibody for both figures or are they saying they are different?
Jade Ed, they report the siRNA knockdown in the same cell line as before, but don't make mention to it being a previous sample. Also they do not fully describe or reference the antibody used in the paper (which to me is a big no no but the journal let them slide on that). The siRNA sequences are referenced to the old paper. The description of the work is pretty vague so I can't tell if its from the old batch of cells or its a new batch. Either way if I'm saying that if I knockdown the amount of this protein and it affects modification of this protein. I'm not only going to run the western for the protein modification but for the knockdown as well too rather than relying on a two year old film.
We can debate this rubbish until the cows come home, but its still reuse of a figure, which I was fairly certain is frowned upon by journals. Is it out and out fabrication of data? No. But I don't agree with doing it. Reference back to the old work and give credit to the authors and journal that the original work is published in.
Jade Ed said:
Anyway- back to your point, it is possible that all the experiments for the 2008 and 2010 paper were all actually done in 2007 and then they published some in 2008 and then the rest in 2010, but this seems unlikely.
Finally someone who can read. The rest doesn't seem to matter anymore. What a relief. Thank you, Jade Ed.
GR- it sounds like the reviewers just let this paper through without reviewing the paper.
I review many papers where my company's products are used and I see this all the time. I've seen papers with experiments with no controls, figures labeled wrong, and conclusions made on "data not shown".
In one of these situations, I looked at where the authors were from, and then looked at the editorial board and saw that a member of the board and the authors both were from the same small liberal arts college. I contacted the editor-in-chief about the paper and his response was "peer review works a lot better than you think." And he invited me to write an official response that would get linked to the paper in pubmed. I don't have time for that nonsense.
In another case, I contacted the authors of a paper to ask about their results and why they did something the way they did, and they wrote back and said, "I really can't remember- we did that work in 2005." Seriously! More than once have authors written back to me to say they did the work 4-5 years ago and don't recall why they didn't use this or that control or why they didn't look at 260/230 readings, etc.
So then the question is, why doesn't a reviewer go "this paper is missing critical experiments to support their conclusions or is missing information that would allow someone else to repeat the experiment" if I do? Does anyone even read it? Could there be a lot of "old boys club" mentality going on and you need to have friends in the right places (editorial boards) to get your papers through without question? Maybe.
These journals that publish unedited rubbish? It reflects in their citation index scores.
You could use their paper for your next journal club to demonstrate to other students how not to write a paper.
Unamerican Scientist, when is infringement of copyright a morally acceptable behavior?
I still don't understand the ridiculous point you're trying to make. So what if they did the control for two experiments 7 years ago. That doesn't make it OK to reuse the data lanes in another paper. Don't be lazy and either re-run the gel (which is ethically not advised) or redo the knockdown (because this is science and replicates are GOOD) and check your results again.