Information Technology Reference
In-Depth Information
of “on-line experts” was compiled, i.e. a list of email addresses of scientists working with
specific protein families or domains, who agreed to review protein sequences in Swiss-Prot
relevant to their field of research. This list is regularly updated and the ~150 experts' email
addresses, grouped by fields of expertise, are listed in the document
http://www.expasy.org/cgi-bin/experts.
However, it does not seem clear to most users - who have grown accustomed to the
repository nature of the nucleotide sequence databases, where only the original authors are
allowed to correct and update existing entries - that Swiss-Prot is extremely different in that
respect, and that we do have an ongoing editorial policy. We do indeed highly value our
users' expertise, and we believe that it is only with the assistance of our user community
that we can do our job of being comprehensive and up-to-date. We are therefore actively
seeking any type of updates and/or corrections, whether they have been published or not,
and would like to be notified about annotations to be updated, e.g. if the function of a
protein has been clarified, or if new post-translational modification information has become
available. In order to increase the visibility of these aspects, and to encourage our users to
let us know about outdated protein entries or errors, we have implemented update forms on
the ExPASy server (see 3.). The forms, accessible from the bottom of every Swiss-Prot
entry, prompt users to provide their corrections and updates in any format. Update requests
are treated with a very high priority by annotators. We are currently receiving about 300
update requests for Swiss-Prot entries per year, a number that we would very much like to
see growing in the future!
On the other hand, annotators send newly annotated entries to the original authors of
reports cited in these entries so as to check the validity of the annotations. We generally get
useful feedback, but not as much as we would like!
Another point of interaction with users is sequence submission directly to Swiss-
Prot and TrEMBL. We accept submission of sequences that have been obtained only as
amino acid sequence. A web submission tool (SPIN) has just been made available, which
guides the submitter through the process, and prompts for all required pieces of
information. There are about 300 such sequence submissions per year. It is interesting to
note that 10% of the proteins originate from venomous animals. This is explained by the
fact that toxins can easily be purified in large quantity from venom and are generally quite
small, thus they are easily sequenced at protein level.
We have to admit that we are disappointed by the low level of input from users in
the updating of the database. We may have been insufficiently efficient in publicising our
willingness and eagerness to welcome any type of help. Yet, after years of discussions with
researchers, we believe that the root of the project is of a sociological nature. The career of
life scientists is driven by the famous 'publish or perish' injunction and submitting data to a
database does not get any credit points on a CV. So we have to rely on the altruism of some
individuals. We are indeed indebted to those persons who take the time to make sure that
we adequately represent the results of their research in our database. However we believe it
is time that the community as a whole addresses this issue and initiates a process of
responsibilisation toward the biomolecular databases.
2.4 Tools for annotation
2.4.1 The basic data organization, the editor and the syntax checker
The working copy of Swiss-Prot is arranged in flat files, grouping proteins by family or
other functional criteria. Although it was apparent from the beginning that the complexity
of protein relationships could not be simulated simply by grouping entries one-
