Plague and anthrax - now you see them, now you don't! (on correcting published errors)

0
Lenny Teytelman Aug 06, 2015

A study that erroneously suggested presence of bubonic plague and anthrax in NYC subways has been corrected, as reported at Retraction Watch. This is a terrific demonstration of the power of post-publication peer review to catch errors in a way that pre-pub review simply can not (See Ed Yong's excellent There’s No Plague on the NYC Subway. No Platypuses Either). At the same time, the way the article has been corrected is alarmingly inappropriate.

The journal Cell Systems issued an erratum (good) and then replaced the original mansucript with a new version, with all traces of anthrax and mistakes gone (bad, bad, very bad). In the new PDF of this study, there is no mention of any corrections and no link to the erratum. On the full-text website view, there is also no indication of the changes, and only at the very bottom in "related articles", do you see the erratum as one of the three links.

The new manuscript is revised and corrected in many places. For example, the abstract changed as follows:

(This is a screenshot of document comparison between the original and new versions of: Afshinnekoo et al., Geospatial Resolution of Human and Bacterial Diversity with City-Scale Metagenomics. 2015. Cell Systems)

It seems self-evident that published research cannot just evaporate and be edited with no trace. Of course, we don't want to mislead the public and add to panic by displaying mistakes. However, given the technology we have today, it is straighforward to implement corrections in a proper manner. Wikipedia and Google Docs track all edits. F1000Research does this nicely (e.g. old version and new clearly marked). 

Our entire protocols.io startup exists because we belive that making it easier to correct published research is critical for accelerating science and making it more reproducible. But we have to do it in the right way: show what was wrong, leave the original intact, direct readers to the updated version.

--------------------------------

[Update 8/7/2015: Ravi Sachidanandam wrote to the journal regarding the disappearance of the original manuscript. Below is the journal's response to the inquiry.]

---------- Forwarded message ----------

From: Caputo, Joseph N. (ELS-CMA)

Date: Thu, Aug 6, 2015 at 1:55 PM

Subject: RE: original article

To: Ravi Sachidanandam 

Dear Ravi,

Thank you for your e-mail and sharing your concerns. I can respond on behalf of Cell Systems for you. We do only have the corrected version posted, and we take this approach for recently published papers based on community feedback that it is best to have the correct version be available for download, and that having more than one version available generates confusion. However, we are happy to send the original version if it is requested. I’ve attached a PDF of the original article here for your convenience.

 

Please let me know if I can help you with anything else.

 

Best,

Joe

Joseph Caputo :: Media Relations Manager :: Cell Press/Elsevier

@sciencemetro

 ---------------

From: Ravi Sachidanandam

Sent: Thursday, August 06, 2015 7:45 AM

To: CELL PRESS Press Office; Caputo, Joseph N. (ELS-CMA)

Subject: Fwd: original article

 

Dear Dr. Caputo,

 I sent the message below to the editors of cell-systems, but have not received a reply yet. I am sure they are very busy, but this impacts the integrity of the journal. I was wondering you could comment on this practice of removing the original articles, which do not allow tracking the changes that were made. The erratum does not clearly show the extravagant claims of the original paper, which got a lot of press.  

I believe a link to the original article should be preserved allowing readers to form their own opinions. It is probably embarrassing, but worth it, if only for the salutary effect it will have on future authors who might be tempted to exaggerate the implications of their work. 

 

thanks

ravi

 

---------- Forwarded message ----------

From: Ravi Sachidanandam <ravi.mssm@gmail.com>

Date: Tue, Aug 4, 2015 at 4:10 PM

Subject: original article 

Dear Editors,

 

The following article had made some erroneous claims that were later corrected

Geospatial Resolution of Human and Bacterial Diversity with City-Scale Metagenomics  Afshinnekoo et al.  Volume 1, Issue 1, p72–87, 29 July 2015

 

http://www.cell.com/cell-systems/abstract/S2405-4712%2815%2900002-2

 

When I check your website, the original article is nowhere to be found. Could the original article be also provided as a link  so that readers can see what happened ? Otherwise it is very misleading. 

 

thanks

ravi

Ravi  Sachidanandam

 http://katahdin.mssm.edu

------------------------//////////////---------------------

---------- Forwarded message ----------

From: Caputo, Joseph N. (ELS-CMA)

Date: Fri, Aug 7, 2015 at 6:11 AM

Subject: RE: original article

To: Ravi Sachidanandam 

Dear Ravi,

 

Yes, you can quote this as the policy.

 

Best,

Joe

Joseph Caputo :: Media Relations Manager :: Cell Press/Elsevier

----------------------------

From: Ravi Sachidanandam

Sent: Thursday, August 06, 2015 10:47 PM T

o: Caputo, Joseph N. (ELS-CMA)

Subject: Re: original article

 

Dear Joe, 

 I assume this is the policy of the journal ?  can I quote this as being the policy ? 

it seems to run counter to accepted practice in scientific publishing.

 

thanks

ravi

 

ALL COMMENTS (3)