What If Authors Do Not Share Their Published Data ?

Question

[biorepine]

Dear biostars,

I am wondering what do you do if any author do not share his/her published data with you ? Usually I send an email request after checking their supplemental and public database (ex: GEO) links or others in their paper carefully. Most of the time they do respond to my requests. However sometimes I never get any response from either first or corresponding author regrading sharing their data even after multiple requests. It is really frustrating as the guidelines of that journal (Cell, Nature) clearly state that the data should be public. Do you have any similar experience ? If you write any complaint to the editor, does it work ?

Thanx in advance
Sorry that it is not really a bioinformatics question

[Pierre Lindenbaum]

yep, that's not really a bioinformatics question. You'd better ask http://academia.stackexchange.com/

[biorepine]

thanx but do you think reposting this question there would make you admin angry :) ?

[Pierre Lindenbaum]

not at all :-)

Answer 1

[Zev.Kronenberg]

Step 0: Check to make sure that you did not overlook the published data. Done it several times ... embarrassing!

Step 1: Write the corresponding author ... busy people

Step 2: Write the first author ... graduate student / post doc should be easier to get a hold of.

Step 3: Write the corresponding author again ... try again

Step 4: Have your professor write the email or letter.

Step 5: Write the editor of the journal.

Step 6: Have your professor write the email or letter to the editor of the journal.

Step 7: Tweet about how awful the lab is (Joke).

[biorepine]

haha! I did until Step3 and decided to ignore the dataset from my analysis because of limited time.

[Pappu]

I once did that to obtain cryoEM map several time. My professor also wrote. They turned deaf ears. Then I complained to the joural editor. Then he contacted the corresponding authors but still they did not reply. When when I wrote to the editor again, he said that he is not able to help anymore :(

Answer 2

[Mary]

I recently had an issue of this, but it was somewhat unusual. Of 4 data sets that were supposed to be available in the SRA, only 1 was found in the database with the accession number from the paper. But that one was not yet public.

Here's what I was told:

Dear Mary,

I have released SRA[XXXXXX], but the user did not finish submitting their data. They only submitted 1 out of 4 data sets. I have contacted the submitter and I hope they will be able to finish the submission.

You will be able to locate SRA[XXXXXX] tomorrow, as Entrez needs to index the released submissions.

....[very nice support person at SRA].

Without me asking, the SRA folks chased down the submitter to get them to submit and release the other 3 data sets that were supposed to be in there. They appeared a few days later.

I'm not saying this is a strategy I would always use, but it was nice to have the SRA get on this for me. So if you are dealing with some data that should be in a public repository those folks might also be able to weigh in.

[Zev.Kronenberg]

SRA people are great!

Answer 3

[cdsouthan]

This is a crucial question and becoming ever more critical for the biosciences (and especialy bioinformatics). Zev's answer is useful but I also suggest the following. 1) You, your colleagues and your PI (if they have not already done so) set up a lab/team/department blog where you describe in detail, not only past research achievments, publications and maybe some unpublished data sets as figshare links (i.e. show willing in the reciprocal sense), but also what you plan to do. OK this exposes you to a certain extent but then you can't have it both ways. This would include specific links and citations of data sets you need and why, with author and affilation names. You then make clear, in the nicest possible way of course, how much you would appreciate an answer to whatever (Zev stage) your enquiry got to, and, importantly, express your gratitude to those who have replied, say you'll cite them wherever possible, and add how good it mght be for them (i.e. enlightened self interest) to drop the same data sets into the appropriate public repository and/or figshare for others to pick up, perhaps even with more detailed metadata than they had in the paper. This strategy could be called name-and-incentivise rather than name-and-shame but the idea is obvious. Note also if you write in Blogger this will be Google-indexed in seconds. This is not only increases the likelyhood of getting the investigator(s) attention and eliciting a positive response but also means you should stay polite (as I am sure you would anyway) and be careful to spell the author names correctly (not sure about non-eglish accents and Google indexing)