First post will include 10 cases of duplication of data detected in 40 randomly selected published articles in International Journal of Oncology.
Please note that posting the detected irregularities here is absolutely not accusations of misconduct.

Case 1.
Tatsuta T, Hosono M, Miura Y, Sugawara S, Kariya Y, Hakomori S, Nitta K.
Int J Oncol. 2013 Dec;43(6):1799-808. doi: 10.3892/ijo.2013.2128. Epub 2013 Oct 4.
PMID: 24100413

Fig. 4C and 5B: duplication of flow cytometry data:

http://i.imgur.com/baPyS9w.png

~25% of the random sample? That's absolutely amazing. Brilliant idea.

Could you also post the list of the articles examined and detail how they were selected? Further examination could be crowd-sourced.

A full list of articles examined will be published soon. All figures were analyzed manually for irregularities using zoomed figures, using contrast/brightness adjustments in Photoshop.

Sigh. this simply looks like the panels you highlighted were repeated as part of a larger comparison panel (note the 3rd panel in the first image is also present, but below the first 2 rather than in-line). The legend/description is the same in both figures, etc. What is the problem here exactly ?

You should never duplicate data in a scientific publication. Duplication of data casts doubts on how the experiments have been performed. Publication of scientific data assumes that you have performed the experiment at least three times. If you need to duplicate data it indicates that you have not performed the necessary number of experiments.

Unfortunately this is all too common in scientific literature today, and probably an important reason why research findings are difficult to reproduce.

Not sure I even understand what you're saying, and it casts a shadow over your experiment and comments. They showed the same data in 2 places in the publication. They indicated it was the same data, based on all the available evidence. It is absolutely OK to show the same data on a different page in a larger context, for readability reasons or simply to make a related argument. This has nothing to do with reproducibility, etc. Again, I have no idea why you would consider this to be a problem.

I totally disagree with you. Data should never be duplicated. You simply should have enough data to avoid duplication! I think the publishers by being more strict would do science a favor and increase the reproducibility score, which now is depressingly low.

If publishers want to introduce specific guidelines to this effect, fine. It is completely unreasonable to state, arbitrarily, that the same image in 2 places right now is a problem, but you are entitled to your opinion. What if the authors simply left out the first comparison panel ? What would be the difference ? Asking for an experiment to be performed k times (and having the data available for the publishers ) is different from the "problem" you're pointing out here. Unless you planned to ask for 3 versions of the data in every other experiment in all other papers.

"Data should never be duplicated" is a statement that is so general and underspecified as to be almost meaningless. But since I am not involved in this field, I'll let others debate the issue further.

Indeed. Many of the cases posted here can hopefully be clarified by presenting the raw data for additional experiments.

(Different Unreg here.)

Dear Unreg @ February 4th, 2015 5:02am UTC (and February 2nd, 2015 11:00pm UTC):
Your comments are absolutely correct and right on point. There is no "problem" in most of these cases.

And you are right that, in essence, Peer 1 is suggesting that multiple versions of the data for every experiment be shown in all papers, which is an unreasonable expectation. (As you say, this is quite different from requiring that experiments are reproducible; it is demanding that data from multiple experiments needs to be shown, with no overlap. In this case, one could well imagine that other readers raise the reciprocal objection: "you can't show blots for two different proteins from two different experiments, they have to come from the same experiment".

There are some legitimate problems revealed in this thread. However, quite a few of items pointed out here are not legitimate problems. It seems that a "zero tolerance" filter has been applied, with no analysis about the actual impact on the conclusions.

I'd say it's not even "Zero tolerance" (as, scientifically speaking, there should be no tolerance towards any kind of misrepresenting data), it is just one person's opinion (Peer 1) imposed as what is misrepresentation of data and what is not. The tone Peer 1 uses, and the broad generalizations made ("duplication is never acceptable...") and the unwillingness to engage in discussion with other peers about cases clearly being not problematic gives off the impression that Peer 1 is on a mission, and does not want to engage in healthy PPPR. If PPPR wants to be taken seriously (as it should) and shake off the bad reputation as "witch hunts" that it sometimes gets, approaches like this need to be properly disseminated and dismissed if appropriate.

I wholeheartedly agree with Unreg from February 4th, 2015 11:46pm UTC, that the reciprocal objection is even stronger: "Why on earth was not the same standard used?". In my field (solid state chemistry), it would be frowned upon of series would NOT be done on the same sample, and thus the baseline was reproduced in different figures.
Essentially, reusing data (properly marked as such) is nothing more than a design choice, to increase accessibility. It is convenient not to have to flip back three pages to be able to compare two figures, because someone yelled "duplication".

All this diverts the attention from the cases where there ARE problems, so this is actually a disservice to the PPPR community.

A brief discussion on twitter last year yielded a similar estimate of 25% of papers in JCB having inappropriate image manipulation being cited by Richard Server from CSH Press: https://twitter.com/cshperspectives/status/435872738235150336

It is not at all clear where this "25%" estimate comes from. As far as I can tell, it is entirely unsubstantiated. It certainly does NOT emanate from the 2004 JCB piece by Rossner and Yamada mentioned in the tweet linked above, which describes only a "vague recollection".

If this estimate comes from an actual analysis, perhaps someone can cite the real source. Otherwise, to avoid propagation of unsubstantiated claims and rumors, probably it's best not to cite someone's vague recollection posted on Twitter.

http://www.nature.com/news/image-search-triggers-italian-police-probe-1.14295

"Now midway through the analysis, he estimates that around one-quarter of the thousands of papers featuring gels that he has analysed so far potentially breached widely accepted guidelines on reproducing gel images. And around 10% seem to include very obvious breaches, such as cutting and pasting of gel bands."

Regarding this issue, please note the following:

1. The story linked by Peer 7 says nothing about the journal JCB, as initially stated by Peer 6.

2. In the linked news story, the "one-quarter" and "10%" numbers are NOT reflective of papers in general. Instead, they are derived from analysis of a very specific subset of papers. Namely, they are papers from Italian scientists, and are further restricted to those "who scored highly in his automated check and collaboration network". This "collaboration network" refers to "scientists who had been co-authors at least three times with authors of the retractions". So, there may be serious issues among those papers, but it is extremely important to bear in mind that this is not a random sample, and instead is a select sub-group of Italian scientists who have collaborated with other authors with suspect papers.

3. So far, the analyses leading to these "estimates" are not yet published, and so they currently remain in the realm of unsubstantiated rumor. It seems prudent to refer to them at most as "tentative". Especially on this forum, we should reserve judgement until the data is available, and make sure the appropriate conclusion is being drawn, no?

http://www.the-scientist.com/?articles.view/articleNo/23749/title/How-to-Guard-Against-Image-Fraud/

"Twenty-five percent of all our accepted manuscripts have at least one figure that must be remade because we detect "inappropriate" manipulation..."

Finally, an accurate link that reveals the source of this claim. Thanks.

That you can't duplicate data (even if stated that the data are duplicated) makes absolutely no sense! Should we re-cristallize a protein every time we show the structure?

The lowermost beta-actin bands of Fig 5 and 6 look just as similar to me.

Er, you're calling them out for clearly marking the limits of contiguous sections of membranes in their blots?

I encourage the authors to present the uncut raw data. Should be easy to retrieve and end of discussion. This is a study to check all the duplications and to cross-check with the specific authors!

As the other Unreg explained (@February 5th, 2015 5:23am UTC), the authors have been abundantly clear and have done nothing wrong. And yet, Peer 1 seems to have some complaint, thought without ever explaining what the problem is. Instead, suggests the authors "present uncut raw data" to "end the discussion". But there is no discussion; there is just a vague insinuation on the part of Peer 1.

Peer 0, please take note: Peer 1 is out of control.

A really don't understand your mission. Why are you unreg? Is there a reason to be completely anonymous?

This is an example of rather suboptimal science. Authors should not have used the same controls, even if they have indicated this in Fig 2D. Same for splicing in Fig 3C, this is very inappropriate today, even if declared. For all we know, control may be on a different membrane in both cases, which doesn't allow comparison. Apparently, peer review failed here.

Maybe a spot in the lens?

Couldn't this one be due to a speck of dust on the camera sensor?

That could be a possibility. The authors have been contacted and hopefully this issue can be clarified.

Also saturated white background to blots: bad practice.

A curiosity: Kim DH and Kim ND are co-authors of as many as 6 articles published in Int. J. Oncology in 2014, and totally 11 articles in the same journal since 2012! Impressive!

You need to extend that rectangle of yours. HIF-1a section is also the same with the figure

Here is the list of articles from International Journal of Oncology that were included in the study. Articles were picked from a random journal volume and number (Vol 43, number 6), and 40 original articles were collected backward in time and analyzed for duplication of data.

http://i.imgur.com/fMURazZ.png

Maybe you could post the text for the list (here) and indicate which correspond to cases with comments here?

Posting text information in image form really makes it difficult for anyone to check your work and assess the quality of your selection. Personally, I'm not going to parse these by hand and look for patterns. Also a better explanation of your random selection method will be helpful. Did you use a script, or just closed your eyes and pressed random article links?

also
1. you make a very strong claim
2. you publish it in the most disorganized and hard to track manner possible.

It would be better if this was organised as a blog post at this stage, or you could have placed it as a single large post here.

Thank you for your message. First of all I do not make any claims other than presenting duplications of data detected in 40 randomly picked articles from each journal (randomly means I picked the newest articles that were available from each journal and continued backward in time till I reach 40). I posted the irregularities here simply because I wanted an open discussion and reply from the authors. The plan was to make a publication based on the reply from the authors.

I have serious doubts about this until we get to see the list of articles that were examined and how they were "randomly" chosen.

Of course I did read the text of all the articles where I found duplications. That is not an issue here. One should never duplicate data. Here the authors should have at least two additional set of identical immunoblots from independent experiments to include in the figure. Re-use of data casts unnecessary doubts about the experimental practice and if the data has been reproduced at all.

That paper was written by Alison Abbott, and not by Ivan Oransky. Please correct the statement.

Oransky tweeted a link to that article - indeed by Alison Abbott - and recalled Bucci's conclusion. The point was what Bucci found, not who reported it, but apologies for the confusion.

Natali and Bagnato are hardly newcomers to PubPeer:

https://pubpeer.com/search?q=natali
https://pubpeer.com/search?q=bagnato

They are same cells and same treatment. It is possible that data presented in Fig 4B and 5C are obtained from the same set of samples (and could be reprobing of one blot), but assembled in the separate figures for some reason.

Please don't misrepresent what I and at least one other poster said. All I meant is that a duplicated figure with the same conditions/legend/etc. in more than one place in a paper is something different from the same data represented as two different experiments, etc. One is a presentation issue, the other one is either an error (which, unfortunately, can and does happen, especially if figure seems exactly the same) or misconduct.
Until multiple domain experts look at the images presented here, it'd be great if people didn't tweet out stats about "problems! in cancer literature!" - you know, actually try some rigor in PPPR as well ? or don't, it's a free Internet.

It is a huge problem that most of the published results from cancer research are not possible to reproduce (Prinz, Schlange & Asadullah, 2011; Begley & Ellis, 2012). If no problem, the authors should present the raw data (whole membrane of the presented data) here.

I am not referring to stats about irreproducibility problems, I am referring to tweets and coverage of this thread under the heading of "% of cancer papers have data duplication problems!". Again, please do not conflate irreproducibility, which is an established but also complicated problem, and surface duplications which may or may not be problematic..

Does this study address irreproducibility ? it doesn't seem like it does. So please don't equate the two, obviously there are other studies not flagged here which are probably not easily reproducible and conversely we have no idea what the reproducibility of the work flagged here is.

Obviously, the authors can comment themselves if they were made aware of this thread. Perhaps creating entries for the specific papers at PubPeer would be more useful or readable though.

The way I see is that there are indeed several arguable cases here (e.g. papers where the only "issue" was WB loading control reuse), but there was an astonishing number of problem-papers as well, which will be much harder to explain. Massive misuse of microscopy images (verbatim duplication or partial FOV overlaps) to represent different samples and experimental conditions is completely unacceptable, and IMO greatly influences the reproducibility of the study. I personally would not even bother trying to repeat such an experiment.

I agree however that creating separate entries for each individual paper would help things a lot.

Yes, of course. All I was saying is that reproducibility is an overlapping but separate issue and not addressed by current study. Agreeing with you on the fact that there are, to my eye, clear problems alongside what I consider either not a problem or a minor problem.

The Zeb1 blots of 6D and 5B also have some similarities: http://imgur.com/2cbvJFm

The most amazing thing about this analysis is that the papers are all from 2013. One would have thought that by then web sites such as Science-Fraud (now closed) and Retraction-watch would have raised awareness that 'irregularities' do not always go undetected. Could the authors repeat their analysis on a series of papers that go back a little further, say 2006-2008?

Actually, I thought Peer's 1 focus on more recent papers was very commendable. Data from a paper published in the last 5 years or less is much more likely to be available (for the cases where the duplications are honest mistakes, which we have seen happen in other places on PubPeer). So it gives the authors a fighting chance and conversely, it leads to a much more clear conclusion when there is a problem, because the data should be around still. Sure, go ahead and repeat the analysis on other papers, but again, the focus on the more recent papers can lead to an actual discussion.

I am afraid that the proportion of irregularities like duplication of data are even higher for articles published earlier, before Pubpeer, Retraction Watch and other scientific blogs were up and running.

International Journal of Oncology, Oncogene and Cancer Cell were contacted many weeks ago with a presentation of the study (the duplications presented here are just a part of the detected irregularities). I kindly asked them to collaborate to study the irregularities more in details by examining the raw data. So far none of the journals have replied...

Please note that approximately 50% of the cases reported here are duplications representing different experimental setup, which is worrying.

I think that scientific publishing need a zero tolerance policy for irregularities like duplication of data and discontinuities. In addition a a huge cleanup of the published literature is needed.

"So far none of the journals have replied..."

Dealing with that many cases could ruin a successful business. It would take time, cost money, create bad publicity and antagonise a large fraction of your authors and referees. Only one sensible course of action: ignore the problem and see if it will go away.

Well, while I have my personal opinions about some of the cases, I do think Peer 1's focus on 2013 (and maybe 2012?) articles, again, was commendable and therefore the journals could likely enquire about the raw data, given that the studies are fairly recent. In other words, unlike some other comments I see in other places on the site, the intent and time range for this study seem fair and should likely result in some response, both from the journals and from the authors.

Peer 1, just wanted to say that I personally appreciate that you tried to make this a collaboration with the journals and a conversation with the authors by focusing on work where, frankly, the data is likely still around. Let's hope people will respond.

I think it's expected that authors may try to make maximum use of the anonymity offered by the site. If the negative comments are from the authors (for what it's worth, I don't think all of them are), we can suspect that they don't wish to respond openly.

In reply to: Unregistered Submission: ( February 5th, 2015 5:00pm UTC )

The number of such irregularities would increase considerably if we would go back in time. It is true that the rather ephemeral Science Fraud site was already around in 2013, but it never managed to gain the same popularity that is enjoyed by PubPeer today. Retraction Watch on the other hand has a slightly different focus and usually only reports the "hard endpoints" (a.k.a. retractions) of PPPR.

The number of visible(!) irregularities will most likely gradually decrease in the future. I fear that PPPR sites gaining widespread popularity are already starting to have an influence on the way people conduct questionable research. Researchers inclined to engage in improper conduct will unlikely to change their habits, however they will rapidly learn that it is much more difficult to catch input-level misconduct (e.g. altering the experimental conditions, samples etc. to ensure the preferred results) than doctored images and other visuals (output-level). The possibility of such a runaway reaction (pubpeer-effect?) seems to be quite high.

Thank you for your very interesting work.

This is offering some very useful insight into life science research and publishing. Could do better...

This is indeed sad but enlightening as well. I have to ask however: did you contact the journals and authors anonymously or not? This matters a lot!

The journals were contacted through my regular job e-mail in October. Authors were contacted through the Pubpeer system, that means anonymously (link from the post).

Noticed today the data highlighted here was cited by the PubPeer admins in this article: http://www.vox.com/2015/3/14/8203595/pubpeer
as having shown that 25% of cancer research has image problems, which they call "a shocking statistic".

I am not in the cancer field, but some of highlighted problems are disputed on this thread (e.g. as presentation choices, etc.)- has the community decided that 25% of cancer research papers actually have problems ?

Asking because that statistic is already being shared on Twitter without any context, etc.

I agree with unregistered (March 14th 7:22 pm). Peer one has pointed out some cases of irregularities that look legit but this is now becoming a contest to sling mud everywhere and lets see what sticks. I randomly checked some papers that peer1 has highlighted here. For example :
Peer 1: ( February 1st, 2015 9:54pm UTC )

Case 2:
Zitzmann K, Ailer G, Vlotides G, Spoettl G, Maurer J, Göke B, Beuschlein F, Auernhammer CJ.
Int J Oncol. 2013 Dec;43(6):1824-32. doi: 10.3892/ijo.2013.2130. Epub 2013 Oct 4.
PMID: 24100469

Fig. 5 and 6: duplication of Western immunoblots:

http://i.imgur.com/K0QL2Lbl.png

There is duplication but the panels are the same so the loading control is obviously the same. For presentation purposes the figure is split into 2 panels. There are more examples like this. I think this is the same peer who pointed out that since some data from one group is suspect he/she does not trust any other papers from that group. What if we take to same stand on Peer1 analysis.
Some restraint should be shown when commenting on other people's work especially when you put it under the heading
" STUDY: Incidence of data irregularities in a randomly selected pool of life science publications"

I am a cancer researcher, and I think that the assertion by peer 1 that all experiments have to be repeated 3 times to be valid is nuts.

For instance, if you detect a mutation by NGS, do you run the identical Sanger sequencing sample 3 times to validate the mutation? Orthogonal correlation of results is far preferable, to my mind.

Other experiments are simply too time, labor, and $$$ intensive to repeat 3 times. For instance, if you generate a knock-out mouse, and follow offspring for 2 years to look for tumor development, do you repeat the experiment from scratch 3 times? Again, corroboration of results by an alternate method (rescue of the knock-out, for instance) is a far better use of time and resources.

So, given that many of the problems identified are due to re-use of control panels, I'm not buying the 25% figure. PPPR is hugely useful, and I find this site interesting, but "extraordinary claims require extraordinary proof". Peer 1 isnt even providing "ordinary" proof.

As an aside, Peer 1 clearly owes Shah and colleagues a public apology:
https://pubpeer.com/publications/23832664

Completely agree 'Cancer Researcher'. Also in cases of western blots where it is possible to do the exp. three times and the authors clearly state in the manuscript that one representative exp. is shown, peer1 just makes the rules on the fly that the authors should show different runs in different figures. Why? who says that's best practice? ofcourse if the exp are published in different journals then yes (copyright issue, ma, but if the exp is split into 2 figures in the same article then as a reader I would much rather see one full representative experiment than parts of three different repeats.
Now someone mentioned that statistics based on this so called study is doing rounds on twitter. is this true?? This will cause unnecessary harm to cancer research field.
Why is no one moderating these comments?

"This will cause unnecessary harm to cancer research field. Why is no one moderating these comments?"

Why are almost none of the authors defending their work? The kind of issue you mention is easy to explain. The papers are recent, so it should be a trivial matter to post original data, if they do exonerate the authors.

I reply to the three previous comments from unregistered.
Please note that when conducting a study where the aim is to investigate the proportion of duplications in life science publications you will need to include ALL identified duplications.
Some of the cases could hopefully be solved by the authors with a presentation of supporting raw data, however only one author has so far responded.
I will still argue that it is of huge importance that all research data presented in articles should be based on several experiments. The knockout mouse example you refeer to is obvious: of course you will not need to make a knockout mouse three times, but you will need to VERIFY the knockout many time!
I do not understand why you mention another Pubpeer post where they discuss a paper by Shah et al. This post has nothing to do with my study and I am not the Peer 1 in this post!
I notice that you are all unregistered. Any particular reason why you need to hide your identity?

Indeed, Unreg (March 19th, 2015 9:24pm UTC ), why are you trying to discredit this study by dragging in an utterly unrelated case (Shah et al)? I initially fell for your insidious trick, until Peer1 has clarified!
As for other issue (Zitzmann et al): showing the same western over and over is not misconduct, but also not quite OK: it implies there was only one western blot ever done. The referees and readers should have been told that, to allow them to draw their own conclusions.

ok let me clarify and it is not a personal attack. Any study (scientific or a critique) should be open to discussion especially in this forum. There are two (atleast) unreg here. My comments are (March 19th, 2015 8:17pm UTC ) and ( March 20th, 2015 12:18am UTC ). My argument is two fold:
Firstly, "showing the same western over and over is not misconduct, but also not quite OK: it implies there was only one western blot ever done." is not correct. Anyone can argue like I mentioned that there is value is seeing one whole representative exp than parts of 3 repeats. It might be one of the repeats was the prettiest of the lot or had samples loaded in the correct way for the paper etc etc. Unless journal give more space (in supplementary maybe) and ask authors to show all three repeats, there is no way to show all repeats that are ever done. even showing different repeats in different figures does not guarantee that everything was repeated thrice.
Second and main concern is that the statistics of this study is being taken without context and posted on twitter (some other unreg pointed it out).
Lastly Peer1 comment about anonymity of ureg submission on this site is actually a little funny since that is the main aim of this site. How many times have the authors whose studies are commented here rue the same fact :-)

(Note, brand new Unreg here; call me "Fred".)

In response to Peer 1 @ March 20th, 2015 8:02am UTC ):
You say: "I notice that you are all unregistered. Any particular reason why you need to hide your identity?"

This statement does not make any sense. Your identity is just as hidden as the Unregistered commenters, isn't it? My impression is that everyone commenting here is equally anonymous, but if that is not correct then that seems important to clarify.

First, please explain how the identity of a "Peer" is less hidden than that of an "Unreg".

Second, please explain how the anonymity affects the legitimacy of the Unregs comments. (My opinion: it doesn't.)



In response to Peer 6 @ March 20th, 2015 9:11am UTC :
Your description of another Unreg's aside as an "insidious trick" strikes me as an exaggeration. Clearly, the Unreg assumed the two Peer 1's were the same (I admit I had also), and so was pointing out the apparent irony between the strident intolerance for imperfection displayed by the Peer 1 here with the unwillingness of another Peer 1 to acknowledge his/her mistaken interpretation. So, if they were the same Peer, then the hypocrisy would be clear. (Did you really not see that link?)

Repeated presentation of the same western blot must be clearly declared to referees and readers, so they can judge for themselves, can we agree on this?
To Unreg "Fred": there are hundreds or thousands papers being discussed on PubPeer, but I daresay most of us should be aware that NOT ALL of the "Peer 1"s there are the same person throughout. So what was again the point of discussing "Shah et al", if not to create confusion and discredit?
As for anonymity of "Unreg" as opposed to that of "Peer": the latter is always registered and internally identifiable with the real name, like I am now. Who "unregs" really are, only they know.

I am one of the unregs above. I find it funny that anyone on this site would say things like "why do you hide your identity ?" when so much being written about this site highlights why anonymity is good. I "hide my identity" because the whole point of the site is predicated on the content of the comments being enough. If the site ends the ability to comment anonymously and everyone has to register, I am happy to do so.
On a different note, while of course "Peer 1" is just a local id and so "Peer 1"s are not the same person across the site, it is obvious to anyone reading the comments that many of the comments for specific topics/journals, etc. come from the same people.
Regardless of all the above, people should stop quoting statistics based on this study, conducted by Peer 1 in this thread, until the community actually agrees that the identified duplications are problematic. Since some of them clearly are not, the default initial statistics are misleading. Peer 1 has not published the study (which is commendable, he/she could have put in on Arxiv and social media would have run with it!), but is basically working through it right now. The cancer community (to which I don't belong) should decide in some fashion what is and is not problematic and only then circulate the stats.

"Repeated presentation of the same western blot must be clearly declared to referees and readers, so they can judge for themselves, can we agree on this?"
yes but even then the papers are marked as duplication. For example see here:
"Protein kinase C α is a central signaling node and therapeutic target for breast cancer stem cells"
This paper was marked by a peer1 (same or different I don't know) when the legend states
E) Western blots of PDGFRa, PDGFRb, PLCg1, and PLCg2, along with protein phosphorylation, in NAMECs and EMT-TF-induced cells are shown. Sample loading controls (GAPDH) were the same as
Figure 4 A. Samples were loaded and analyzed on the same blot."
More worrisome like I stated before that statistics based on such analysis are doing rounds without the underlying context.
Do u see a problem here?

Still one the "unregs" from before: to be perfectly honest, the reason I don't log in is that when I do, my name will be in a database and I am frankly afraid that the site may at some point in the future be hacked (happens all the time). Other info is much harder to connect with my real life identity, but yes, I worry about real-life consequences of commenting on this site.

link to a good reason for maintaining anonymity........

http://news.sciencemag.org/people-events/2015/03/michigan-judge-asks-pubpeer-turn-over-anonymous-user-information

doesnt matter if what one posts is true regarding sketchy data....
no one likes being hassled by lawyers.

First, you forget the numerous other papers flagged in this study which are indeed problematic. So even if the number of 25% may be somewhat lower after all, the study does have its value. Is 20% or 15% or even 10% below toleration threshold for data manipulations in cancer research? What if the real number is much much bigger, since Peer 1 lacks proper sophisticated analysis software to detect all of the data tampering?
As for the Weinberg paper you discuss here: Peer 1 also raised a concern of excessively bright WB images, which is a very important discussion point. There is no valid scientific or technical reason to make WB images so bright. I know the defensive arguments (short exposure, etc), but they don't convince me, and I did enough WBs in my time. Finally, Weinberg had to retract or excessively correct several papers recently (see RetractionWatch), so it is not entirely unreasonable to look carefully now.

Re: Peer 6 comment @March 20th, 2015 7:39pm UTC )

Yes, we do realize that "NOT ALL" Peer 1s are the same. But regular browsing suggests some dominant and hyper-frequent commenters, based on similar style of comment or argumentation, and these similar styles can naturally lead to assumptions that the Peer 1 in two distinct threads might be the same. And it is clear from the other Unreg's comment that he DID ASSUME they were the same in the two posts. Once you recognize that, it is clear the point of the comment was NOT to "create confusion", as you imply, but to draw attention to the apparent double standard in the two threads. The point is clear if you realize what the assumption was. It turns out the assumption was a mistake. Pretty simple. Your suggestion that it was an intentional is simply not credible, in my opinion.

Second, about your description of Peers vs Unregs: "As for anonymity of "Unreg" as opposed to that of "Peer": the latter is always registered and internally identifiable with the real name, like I am now." I assume that by "internally identifiable", you mean that the PubPeer administrators know your identity. If so, that fits my understanding. But it is irrelevant to the point raised. The identities of all commenters, Peers and Unregs, are equally "hidden" to EACH OTHER, right? So, how does the "registration" affect the validity of anyone's argument? Peer 1 was clearly trying to delegitimize another commenter by questioning their anonymity ("Any particular reason why you need to hide your identity?"), and it is not the first time on this thread (e.g., see above @ February 5th, 2015 9:16pm UTC "A really don't understand your mission. Why are you unreg? Is there a reason to be completely anonymous?"). This is not a valid basis for debate.

-Fred

Fred, you argument makes no sense, especially with this Shah et al case. How on earth does this comment resemble the comments on this topic by Peer 1:
Peer 1: ( March 10th, 2015 5:16pm UTC )
Figure 1.
http://i.imgur.com/aU79FAS.jpg

Really, this is very illogical and inappropriate.

Peer 6, it's clear your not understanding something, but I'm not sure what it is. And I've tried explaining several times. You seem unwilling to accept a simple explanation. So, there seems little point in continuing. I will sign off now.
-Fred

The post by unregistered regarding a totally unrelated discussion (Shah et al.) may be a way of discredit my study, I don't see any other reason why this is put forward here in the way it is presented. Assumptions that "the Peer 1 in two distinct threads might be the same" is also very unserious and such nonsense should be avoided here. Statements like these may be the reason why some individuals needs to be unregistered and to hide more than others.

Regarding Zitzmann et al:
Apporox protein weights: ErbB3: 200kD, EGFR: 200kD ERK: 44kD, AKT: 70kD, p70s6-K: 70kD, 4EBP1: 20kD, , IGF-IR: 100kD, beta-actin: 40kD
Difficult to imagine all this was probed on the same gel, stripping or no stripping. M&M makes no reference to gradient gels, but even then it wouldn't have worked out.

These actin blots dont represnt same exp. conditions either. Also can anyone else confirm that theright 3 bands are spliced in or not? It looks like I can see a splice line but not sure

Same unreg here. The blots do represnt the same exp. Condition. My bad. But still not sure about yhe splice.

IKKa resolves at about 90kDa, pLAT resolves at <40kDa. It would not be unreasonable to cut the membrane into strips or strip and repropbe these blots for multiple proteins; therefore the duplicated actin controls may not be in error if the experimental conditions are indeed the same.

RE: Unreg @ June 26 7:14pm:

I fully agree with you. But your logical explanation misses the point of this example, which is this:

The author responsible for this duplication example is the exact same person who vehemently objects to such duplication. This is evident in comments from Peer 1 on this thread, and from the recent publication by Oksvold (in Sci Eng Ethics), where he says of such "Category I duplications" the following: "Duplication of data reduces the trust of a scientific publication and brings unnecessary uncertainty as to whether experiments have been performed in a scientifically acceptable way (e.g. with proper controls and a sufficient number of true replications)".

I (and apparently you) disagree that such "category I duplications" are inherently problematic, and instead think that if data in two panels emanate from the same blot, then it makes sense that the loading control should be the same.

But the author here is being self-contradictory: he argues that this practice is unacceptable while simultaneously doing that very practice. The hypocrisy is overwhelming.