Leading cancer researcher retracts 2003 paper for “inappropriate presentation”

cancer cellOne of the world’s leading cancer researchers, MIT’s Robert Weinberg, has retracted a decade-old paper after finding out it contained errors.

Here’s the notice for “Ras Modulates Myc Activity to Repress Thrombospondin-1 Expression and Increase Tumor Angiogenesis,” a paper originally published in Cancer Cell in 2003:

We recently were made aware of a number of errors that were made in the processing/compiling of data figures in this paper. Although we stand by the conclusions of the paper and intend to provide support for them in future publications, the inappropriate presentation did not and does not fall within the bounds of acceptable scientific practice. Accordingly, we are retracting the paper in its entirety. We apologize for any inconvenience this may have caused.

One of the original authors, Yi-Ning Cheng, could not be reached regarding this Retraction.

The paper has been cited 233 times, according to Thomson Scientific’s Web of Knowledge. The study was funded by the U.S. National Cancer Institute.

Weinberg tells Retraction Watch the errors were due to “inadvertent sloppiness,” not misconduct:

These errors occurred during the (electronic) assembly of the figures, with internal loading controls being misassembled with the incorrect experimental channels. While they do not and did not affect the major conclusions of the paper, the journal nevertheless requested that we retract this paper because there were multiple errors in assembling these figures.  We stand by these conclusions and we are about to republish these data (with corrected loading controls) plus additional newer data that extend and solidify the original findings.  There was not and is not any indication of scientific misconduct, simply inadvertent sloppiness in the last minute assembly of these figures.

Weinberg is one of the most-cited cancer researchers in the world, having discovered the Ras and Rb genes. Twenty of his papers have been cited at least 1,000 times, and is the subject of Natalie Angier’s 1988 book, Natural Obsessions.

61 thoughts on “Leading cancer researcher retracts 2003 paper for “inappropriate presentation””

  1. “There was not and is not any indication of scientific misconduct, simply inadvertent sloppiness in the last minute assembly of these figures.”
    “One of the original authors, Yi-Ning Cheng, could not be reached regarding this Retraction.”

    Had Yi-Ning Cheng been the assembler of these blots? If he had but could not be contacted regarding the retraction, how can they claim with any level of confidence the errors were due to sloppiness?

    1. Yi-Ning Cheng appears to only had one publication. A search of the name and MIT pops up some entries regarding tennis and an undergrad award (elected to Phi Beta Kappa in 2002).

  2. 1) This might be a little bit os a sea change for Cell journals in that a mega-correction was not allowed.

    2) Note how speculation was immediate regarding nefariousness in comments here, when there is an obvious non-nefarious interpretation for an author not being reachable. I do think these statements could give more information to anticipate the types of questions, but I also think the authors would be reticent to explicitly state who did what for the paper out of embarrassment and out of the desire not to have increased speculation about their motives or the inevitable hobby of someone trolling through their output for the last 10 years.

    1. The retracted paper probably has a footnote describing who did what but I do not have access to it. Yi-Ning Cheng is a second author on the paper, so he was probably some ephemeral undergrad in the lab. I have read many retraction notices on RW and authors typically want readers to know who, in their opinion, is the responsible party. Hence the phrases like “X did not sign the retraction notice”.
      If the first author of the retracted paper was sloppy with the blots, it is not impossible that his sloppiness went beyond this particular paper. After all, it has taken 10 years for the authors to figure out this paper contained sloppy artwork. Conversely, it is only recently that whoever assembled these blots has learned that he was going about this the wrong way.

      1. Or the errors have been known for a long time and the paper is just getting retracted now. There are many possibilities. It is interesting to know who the responsible party is, but the reality is that some commenters at RW would, seemingly no matter what, use a translation of any explanation of “what it really means”. This is why people are reticent to explain further. In the current case, the close wording of the note suggests that the reason for retraction is the figures are messed up but that the data are sound. I find it more interesting that the Cell journal editors did not allow a correction, suggesting that their policy may now relate not to the scientific outcome necessarily (as in previous mega-corrections) but in the actual data as published (not allowing new experiments into a correction, for example).

        1. “Or the errors have been known for a long time and the paper is just getting retracted now” – that is not what the retraction notice says.
          “[T]he figures are messed up but that the data are sound” – figures ARE data, so it is the data that are messed up.

          1. Figures are representations of the data. “The errors occurred during the (electronic) assembly of the figures…” If loading controls are misaligned for a qualitative analysis (representative experiment) this does not mean that the samples were unequally loaded and the experiment is invalid. It means that the presentation is invalid and that the results cannot be interpreted as presented. It does not mean the underlying data are “messed up”.

    2. Now you mention it Pinko, there does appear to be some splicing of westerns in this paper by the first author of the 2003 paper:
      Prosaposin inhibits tumor metastasis via paracrine and endocrine stimulation of stromal p53 and Tsp-1.

      Kang SY, Halvorsen OJ, Gravdal K, Bhattacharya N, Lee JM, Liu NW, Johnston BT, Johnston AB, Haukaas SA, Aamodt K, Yoo S, Akslen LA, Watnick RS.

    3. I was trying to be fair to the one author who did not appear. One publication and an undergrad suggests their interests did not extend to research, and they moved on. The first author:


      And by the way, Yi-Ning Chen is a she. Unless the one I have found information on is a different Yi-Ning Chen, which is possible.

      That said, I think she’s an MD/MPH now. Probably why they couldn’t find her. At this point, I will refocus on the Retraction-Watch stuff at hand.

      What does everyone think about this retraction? Is it good that he is going back and fixing his record? Anyone remember the comment from a previous post where a commenter mentioned that a faculty member found an error related to a new version of a program, and buried it?

  3. Oh no … not Bob Weinberg!
    I took classes with him and interviewed him when I was a Knight fellow at MIT a few years ago.

    1. I am confused. What is your concern? Here we have a guy who found an issue with his own paper, is retracting it, and will republish (contingent upon acceptance) the corrected MS. What’s not to like?

      1. The concern, which I share, is over the admitted “sloppiness in the last minute assembly of the figures” (if this is happening in the Weinberg lab, where ISN’T it happening?) and also the obfuscation within the notice. What does “internal loading controls being misassembled with the incorrect experimental channels” mean exactly? Were loading control blots flipped, used from the wrong experiment, duplicated?

        1. He has 13 postdocs working in his lab (http://weinberglab.wi.mit.edu/members.html) in an environment that is probably quite competitive (living up to the pedigree of the lab with its numerous high impact publications) – these are exactly the places where junior researchers might feel pushed to come up with exciting data with sometimes dubious methods…

  4. On paper, this looks like how retraction should be handled. Better late than never. That said, we could easily employ scientists for a lifetime peer reviewing and rechecking research with respect to contemporary technology, provided they wanted to work for free and had access to a TV show’s worth of laboratory equipment.

    1. “On paper, this looks like how retraction should be handled.”
      If you mean that the journal stood up to what appears to have been an attempt by the PI to, if not bully, at least bluster the journal into backing down, then I agree with you. Kudos to the journal for standing their ground against someone who wields a lot of clout.

      However I would part company if you meant something like this: “Here we have a guy who found an issue with his own paper, is retracting it”
      Since the PI is very clear that he felt retraction was the wrong path to take, it is difficult to see any other scenario than someone reported it to the journal, who reported to the PI.

  5. There is no evidence that Dr. Weinberg tried to bully the journal. One of two things happened, either someone told the journal there was a problem with the paper, or someone in Dr. Weinberg’s lab notified Dr. Weinberg who contacted the journal. In either case, Dr. Weinberg, who is likely to be privy to more data on the central conclusions of the paper, felt the paper was correct and he might be able to correct the record in a future publication. The journal may have requested that the correct data be submitted for the paper where there had been errors in order to allow a correction, but data that had been collected at the time of the initial paper. However after 10 years, finding the proper information with the absence of the author involved in preparation of the figures was probably impossible. The journal then decided that given the questions and the inability to provide the proper data to correct the paper, they requested a retraction. This is my best guess of the train of events.

    I think it is admirable that Dr. Weinberg did not throw his post-doc or student under the bus. If he knows there was no misconduct, and he is likely to be the best judge of that issue, and stood up for her even in the face of a retraction, then he took the more courageous path and put himself in line for criticism or even implications of covering up a problem. I tend to take him at his word. I think he did the right thing and I am certain he will provide a follow-up paper with the corrected experiments, and if the data support the original conclusions, Cancer Cell should publish it.

    1. just a curious question. What happens to the citations of the 2003 paper – will be added to the new paper?

      1. From a theoretical point of view it would be interesting if somebody else jumped in and attempted to publish a paper announcing that “Ras modulates Myc activity to repress thrombospondin-1 expression and increase tumor angiogenesis”. After all, the scientific record as of today lacks this discovery. While its acceptable to reserve a parking space, it is probably not OK to reserve a future scientific discovery for oneself. Alternatively, because the retracted paper has been cited so frequently, it is very likely that others, as part of their own research, have already either repeated the original study or performed something very similar. This would render the future resurrection of the retracted paper impossible, as it would duplicate what has already been done by others – and these people are within their rights to claim the scientific priority. Weinberg, as any other co-author of the retracted paper, is not a suitable person to act as a judge as to whether the numerous figures in the manuscript have been misrepresented due to sloppiness or misconduct. There is an obvious conflict of interests here. Consequently, as a matter of precaution, the local research integrity officer should look into the publication record of the person who provided the incorrect figures for the manuscript because the intent is the only feature that set sloppiness and misconduct apart – the net results of both are the same. If the raw data are no longer available, it is probably not an extenuating circumstance. Surely, the conclusions of the retracted paper are rock solid as evidence by its numerous citations but it quite possible to arrive at correct conclusions using either insufficient or incorrect data.

        1. well, if someone reproduces a similar study, then where he/she can publish the paper. Definitely not in Cell family – because it is not novel anymore. If there are no reproduction of the study, should we conclude that it can not be reproduced. What a dilemma? There might be plenty of novel discoveries which are published in Cell family, Nature family and Science which were never reproduced and community has accepted them as novel and major discoveries. What you all think? this is not directed to any one, really.

          1. “Definitely not in Cell family – because it is not novel anymore.” – this is an inaccurate statement. If it were correct, then Weinberg an co-workers could not declare their intention to re-publish their study with legit figures because it would not be, well, novel. As far as I can tell there is no record of people proving that Ras modulates Myc activity to repress thrombospondin-1 expression and increase tumor angiogenesis. If there is then I probably do not understand what a retraction entails.

    2. ” or someone in Dr. Weinberg’s lab notified Dr. Weinberg who contacted the journal….This is my best guess of the train of events.”

      Now you mention it that does seem highly likely.

      1. It is unlikely that Weinberg had anything to do with generating any sort of data for this paper. It is probably the first author, Randolph S. Watnick, who did the bulk of it.
        The way to distinguish sloppiness from misconduct is by either the frequency or attempted concealment. In this particular paper the data seem to have been misrepresented extensively. “[A] suspiciously white background” might indicate an attempted concealment of data manipulation. This would not square with being sloppy.

        1. We will never know, but it may have been another researcher from the lab who alerted the authorities who then alerted the authors who then alerted the journal.

          In my opinion it is highly unlikely that a PI would contact any journal and say he/she has discovered an error in oublished work and would like to correct it.

          Though there have been such cases in times gone by, i think there may have been pressure from elsewhere to get this done.

        1. In response to Fernandos post, take a look at these.

          Figure 3 panel A mNBS fibroblasts (left columns).
          90 degree flip of image representing two conditions: RAS LTK1 (top panel) and Ras LT-WT (bottom panel)

          Figure 2, Middle Panel A RAS left: LTK1 column 3rd band from left: Splice on left edge.
          Figure 2, Middle Panel A RAS left: V- column 2nd band from left: Splice on left edge.
          Figure 2, Middle Panel A RAS Right: V- column band on left: pixel shading (rubbed out image)
          Figure 2, Panel B. B-actin row flipped vertically.

          Red arrows would help

          Anyone see anything else?

          Maybe these were genuine errors with innocent explanations, but it may be worth looking into.

          1. Fernando is on fire at the moment. And yes there is more.

            Download the high quality images from the html version at Cell. It is to Cell’s credit that the figure images are often very good resolution, allowing detailed criticism of papers that are 9 years old and more.

            Figure 4A

            Beta-actin in top panel, lanes 1-2 are same as lanes 4-5 in the lowest panel.
            Beta-actin in top panel, lanes 3-5 are same as lanes 1-3 in the lowest panel. However they have been inverted in the lower panel.

            In the lower panel, the bands have been spliced into the large grey box. Splice edges are visible before band 1, after band 3 and after band 5.

            It would be very interesting to hear what the innocent explanation for this is.

            Figure 5C

            Beta-actin strip has splices after lanes 2,4,5

            Figure 5D

            p-Akt strip has splices after lanes 2,4

            Figure 5E

            Ral strip has splices after lanes 1,3,4. The bands have been cut into a grey box as indicated by the splices before lane 1 and after lane 5.

            Figure 2 splicing is even worse than Stewart has reported

            Figure 2A

            Left Ras panel has splicing after bands 1,2
            Right LT panel has slicing after bands 1,5
            Right Ras panel has splicing after bands 1,2, possibly after 4 and something strange is going on below 4-5.

            Figure 2C

            Ras splicing after 1,3

            Figure 2D

            Ink4a splicing after 3
            Beta actin splicing after 3

            Did we miss much Fernando?

  6. Cancer Cell. 2004 Aug;6(2):171-83.
    Figure 3A. Top row of panels. I think that the v panel (1st, left-most panel) and the Ras + panel (4th panel) are the same. Rotate one by 180 degrees to get the other.

    1. So it is! And still there’s more:

      Fig. 2B 2nd Left image column: Right half of LTK1 is the same as left half of LTdelta434-444. Not easy to spot but once you see it, there is no doubt.

      Fig. 5.
      5B image S35+C40 is the same image as 5G DN-Ral though the latter is cropped and rotated.
      5B image S35+G37 is the same image as 5G LY294002 though the latter is cropped and rotated.

      This does seem to be one of those papers that rewards close study?

      1. Reply to Scrutineer January 19, 2013 at 12:23

        When images are only partially overlapping, or rotated, or flipped, they can be difficult to spot.
        On first sight of figure 2 in Cancer Cell. 2004 Aug;6(2):171-83 I thought that something like that was going on, but decided to pick the lower hanging fruit first. The hairs on the back of you neck do not stand up, but there is that odd feeling. You know that something is the same,but cannot spot it.

        Much of modern physiology/neuroscience is akin to the discredited social sciences, but some might be salvageable. The neurological reactions to image manipulation warrants study, if only for its utility. EEG readouts of different types of image manipulation seems something easy to arrange. If any were found to be reliable they could be put to good use in high-throughput screens for scientific fraud.

        I believe that in Building 149 of the Old Navy Yard, Charlestown, Massachusetts, part of the MGH, there is such equipment and expertise.



  7. And more…
    Figure 4B mNBS-SV40ER C40 (bottom) is the same photo as 3A Bottom right. They are purported to be different experiments with different ras alleles. This is now the second Cancer Cell paper in question with one common author Rangarajan. This paper has a lot of problems. Difficult to explain by sloppiness.

    1. Errrr. two common authors, actually. Shall I mention who the other one is?

      It would be extremely unusual for someone listed as third author to have much to do with assembly figures for the paper that has been retracted.

      1. actually, i think, two cancer cell papers got the third author (2003 paper) and first author (2004 paper) got this person a “very” good job in a premier institution…if one is retracted and other one has irregularities – any further consequences?

        1. I expect there won’t be. All I am saying is it is absolutely poor methodology to say because Rangarajan is the first author of a paper with irregularities he is responsible for the irregularities for the paper he is third author. Did Watnick say to Rangarajan, “I am just heading out for a coffee, can you fix all these figures up for me while I am out.”? More likely it points to a lax culture in the lab at the point in time.

          The primary responsibility for each paper should be borne by the 1st author. There should also be a reverse onus on the last author to prove he did not know what was going on

          1. Agreed. the primary responsibility should be the first author – I did not check the paper properly – sometimes it is possible that some figures are provided by a co-worker. When the paper gets published in a good journal – all take credit. however, when something like this happens – only the first author gets the blame. What I mean to say is that all authors should take responsibility – not usually happens.

          2. Ressci, I have already pointed this out, but there appears to me to be clear splicing of western blots in this recent paper, last author Watnick
            Prosaposin inhibits tumor metastasis via paracrine and endocrine stimulation of stromal p53 and Tsp-1.

            Kang SY, Halvorsen OJ, Gravdal K, Bhattacharya N, Lee JM, Liu NW, Johnston BT, Johnston AB, Haukaas SA, Aamodt K, Yoo S, Akslen LA, Watnick RS.

            It is only of the loading controls, but it means almost certainly that no loading controls were actually probed for in the case of these experiments.
            At least panel B has a splice in the bottom pane but not the top pane. There is probably splicing in panel H, but it could be in both panes.

      2. Yes, you are quite right, 2 common authors although I doubt Dr. Weinberg had anything to do with figure prep. I also agree that it may not be that the same author was responsible for the problems in both papers. I was unable to figure out what the problems were with the retracted 2033 Cancer Cell paper to see if it might have been the 3rd authors contributions, i.e. a minor fraction of the data, but the 2004 paper above has a minimum of 8 examples of photographs being manipulated and reused as something else. This appears to be whole scale fabrication and does not even go into all of the obvious gel splicings, which are suspicious but not definitive. I have never seen anything quite like this.

    1. It is not possible to be certain who is responsible for anomalous images without a proper investigation that has access to the primary data. While circumstantial evidence, such as similar anomalies in other papers, can point in the right direction, sometimes it points all over the place. Fernando Pessoa has shown anomalous images in another paper in which A. Rangarajan was an author; littlegreyrabbit has pointed out another on which RS Watnick was an author, and here is another, in which RA Weinberg is an author but with a different set of co-authors: Proc Natl Acad Sci U S A. 2007 June 12; 104(24): 10069–10074.
      In the upper part of Fig. 1D, the first 3 lanes look the same as the 5th, 6th, 7th lanes, yet they are labelled differently, and in Fig. 4F the vector b-actin band does not appear to be continuous with the other bands.

      1. In reply to michaelbriggs January 20, 2013 at 12:24

        “a proper investigation that has access to the primary data” is what is needed.

    2. Generally I agree with what Michaelbriggs says, but with the extent of splicing and reuse in the retracted paper it seems to beggar belief that this could happen without the knowledge of the first author. I am always suspicious of scapegoating – first it was the MD, now the scientist returned to India. There (obviously) needs to be investigation, but the first author should a reverse onus of proof to show it wasn’t him and the last author should have a reverse onus to demonstrate he was unaware.

      I would say this one has a few issues with it
      Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis.
      Yang J, Mani SA, Donaher JL, Ramaswamy S, Itzykson RA, Come C, Savagner P, Gitelman I, Richardson A, Weinberg RA.
      Figure 1 Panel E, the last actin lane has been spliced on
      Figure 5, Panel C, the actin and twist strips have been reused. Possibly this could be a feature and not a cheat, if it is possible to probe for that many proteins on the one western.
      Figure 6, Panel D, it appears if the empty Fibronectin lane was spliced on. It certainly appears that way, but image resolution is very poor and there is an outside chance it could be an artifact of image compression software. None the less the splice appears to go from top to bottom of that strip.

      I know Adam Marcus likes his punning headlines – if this one gets retracted how about
      “Well, shake it up, Bobbby, now, (shake it up, Bobby) Twist and shout. (twist and shout)”

      1. Ahh littlegreyrabbit! Next you’ll be telling me that in this paper: Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis, by Yang J et al., that in Figure 3A the same tumor is shown for mouse #352 as for mouse #358, just taken from a slightly different angle…

        1. No because I missed it, thought I did have a good stare at that figure, but I think I concentrated on the top row. Do you think they bloodied one of them up a bit before taking the 2nd photo or was it done in photoshop?

          Perhaps they didn’t like cutting up mice? Can’t say I blame them.

          1. No, I don’t think they photoshopped it. They probably took multiple pictures and then put them in the draw labelled “positive results”, and then when they were taking them out to make the figure and adding the mouse numbers, they accidentally got two pictures of the same tumor. This might be purely accidental, but it seems to happen a lot in Boston. A similar thing happened in this paper: Nature 475, 231–234 doi:10.1038/nature10167, which was corrected here: Nature 481, 534 doi:10.1038/nature10789. The authors wrote “We have also been unable to verify without doubt that the image in Supplementary Fig. 9b shows four different mice within the treated and untreated groups and therefore wish to replace this figure.” Perhaps the authors of the TWIST paper can use this wording if/when it is corrected.

    1. I expect they are perfectly aware, not least because they know all about this blog. However the ORI doesn’t investigate papers this old. The reason might be that they only look at recent papers so that the data trail will be complete. (And if the data trail is incomplete for a recent paper, that in itself would be a major failing.)

      1. One of the papers was 2007. What is the cut-off date?

        Having said that, most of it could just be symptomatic of people taking short cuts but believing their results would be repeatable in another lab.

        That doesn’t mean there shouldn’t be automatic retraction – or else where is the incentive not to take short-cuts? The recent University of Lisbon example should set the standard everyone follows. It is hard to believe there was an intent to create false positives in that image manipulation – particularly as the result seemed so negative. But retraction followed automatically once the manipulation was demonstrated, regardless of what lab records may or may not have shown.

        1. LGR

          I think it is five years. But I didn’t find anything with a quick check of the ORI web site so it would be good if someone could link to that or correct me.

          Of course, the stricken institutes are still obliged to conduct their own investigations and submit reports to the ORI. If I understand correctly, the ORI only actively oversees the investigation of a small number of more or less contemporary fraud cases. For sure, their staff is not large enough for them to take on all the current accusations that you would be familiar with. But let’s keep on figure sleuthing and if enough pressure is exerted perhaps their staff complement and investigative capabilities can be beefed up?

  8. What will happen to people who quoted this paper in their own papers? Will they have to report that the 2003 paper was RETRACTED? And what about those suspected papers also mentioned? Help!

  9. cathy s. states “In my opinion sloppiness is misconduct.”
    This is a massive claim to make, even if only an opinion. If your equation were true, then the vast majority (?) of papers should be retracted from the scientific literature, no doubt, because most contain some level of “sloppiness”. I would personally like to hear more about your views about this equation. Could you try and elaborate more, and quantify as much as possible, because the concept could be useful for all fields of science.

  10. First author, RS Watnick.

    Science. 1999 Dec 17;286(5448):2337-9.
    Binding of transcription termination protein nun to nascent RNA and template DNA.
    Watnick RS1, Gottesman ME.
    Author information

    1Department of Biochemistry and Molecular Biophysics and Institute of Cancer Research, Columbia University, New York, NY 10032, USA.

    Figure 3B.

  11. First author, RS Watnick, as last author.

    Proc Natl Acad Sci U S A. 2009 Jul 21;106(29):12115-20. doi: 10.1073/pnas.0903120106. Epub 2009 Jul 6.
    Prosaposin inhibits tumor metastasis via paracrine and endocrine stimulation of stromal p53 and Tsp-1.
    Kang SY1, Halvorsen OJ, Gravdal K, Bhattacharya N, Lee JM, Liu NW, Johnston BT, Johnston AB, Haukaas SA, Aamodt K, Yoo S, Akslen LA, Watnick RS.
    Author information

    1Vascular Biology Program, Department of Surgery, Children’s Hospital Boston, Boston, MA 02115, USA.
    A correction has been published for figure 5.
    “In addition, several panels in Fig. 5 were labeled incorrectly.”

    Figure 1B.

  12. Proc Natl Acad Sci U S A. 1998 Feb 17;95(4):1546-51.
    Escherichia coli NusA is required for efficient RNA binding by phage HK022 nun protein.
    Watnick RS1, Gottesman ME.
    Author information

    1Department of Biochemistry and Molecular Biophysics, College of Physicians and Surgeons, Columbia University, New York, NY 10032, USA.

    Figure 3A.

    Figure 3C.

    Figure 3B.

  13. Weinberg ” We stand by these conclusions and we are about to republish these data (with corrected loading controls) plus additional newer data that extend and solidify the original findings.”

    Has there been such a publication?

  14. I believe that this be the later publication.

    Oncogene. 2015 May 28;34(22):2823-35. doi: 10.1038/onc.2014.228. Epub 2014 Aug 11.
    Thrombospondin-1 repression is mediated via distinct mechanisms in fibroblasts and epithelial cells.
    Watnick RS1, Rodriguez RK2, Wang S3, Blois AL4, Rangarajan A5, Ince T5, Weinberg RA6.

    THe correction of which is covered here,
    and here.

    Also, figure 5.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.