sequence not available

Paul Gilna pgil at TEMIN.LANL.GOV
Mon Apr 5 15:59:02 EST 1993


Terry Farrah writes

> In response to questions about why database sequences are not always
> released immediately after publication, Paul Gilna comments
> (regarding GenBank):
> > ... we have conferred upon the submitting author, the
> > responsibility for updating and maintaining his or her data.
> 
> This is sensible for every aspect of the data except publication status.
> When scientists ask that their data not be released prior to publication,
> it is often because they are in a competitive situation and are motivated
> to restrict access to their data.  It is not surprising, then, that
> they don't make the effort to notify sequence databases when their data
> has been published.
> 
> I like Renee Lippens' suggestion:
> >   ... why not ask to the Journals or Publishers to notify the databanks when
> >   a publication containing an Accession Number is issued.
> 
> -- 
> Terry Farrah  (farrah at immunex.com)
> 
> 

Though with some journals it is stated more explicitly than with others
(e.g., Cell), there is a form of agreement between the journal and the
publishing scientist that expects that all data and reagents associated
with the publication are to be shared with the scientific community.
Deliberate delay in having the data released could be considered to be
in ethical abbrogration of this understanding.

But that is taking a stick to the problem.

Let me give a simple example as to why it is in a scientist`s interests
NOT to have the data held confidential. Except perhaps in a legal court
of law, the act of submission has little bearing on whether a scientist
has proprietary publishing rights (i.e., proof that group A got there
first).  On the other hand release of the data to the public domain has
a very real bearing on these issues. nce out there, your data have
associated with them, both a date of release and your name. Therefore
the scientist who is worried about being scooped from competitive
access can very easily be scooped by publication from those same
competitors simply because his or her data were stuck in confidential
status sitting around for a journal to get its publishing act together
while the competitors data sailed gaily and publically by.

I would suggest that given the rate and volume of data
appearing in the public domain (approx 2 x 10**6 bp per week) today and
the increasing acceptance of the databases as an electronic data
publication forum, justification for confidential data by fear of
competitor's access is becomming increasingly moot. Therefore it is
more in one's interests to get the data out there than it is to hide
them.  This can certainly be evidenced by the fact that as a percentage
of total submissions, we find that the fraction of confidential
submissions is showing a definite downward trend.

Finally, to bring into perspective that last point, much has been made
in this issue of authors (or databases) inability to release
confidential data in time to coincide with publication. Lest readers be
left with the perception that all submittors are recalcitrant, let me
reassure you that the trend is such as to suggest the exact
opposite--more and more submittors are taking it upon themselves to
inform us of changes to the publication status of their data (and much
more besides)--many with our help!. Further, we find we are handling
considerably less "I couldn't retrieve the following sequence" queries,
which once represented the *only* class of messages coming to update when we
started it. While we are by no means at 100% compliance, we are
definitely headed in that direction.


> 
> I like Renee Lippens' suggestion:
> >   ... why not ask to the Journals or Publishers to notify the databanks when
> >   a publication containing an Accession Number is issued.

A number of journals, particularly Science, PNAS and Appl. Env. Micro.
do this already for us, and we are grateful to them for that, yet
increasingly the submitter is getting to us first.

Cheers,


--paul



More information about the Embl-db mailing list