Biomedical Engineering Reference
In-Depth Information
dentiality compared to data derived from a simple enzyme assay. These human
issues could extend to business agreements and intellectual property arrange-
ments that may underpin any major collaboration, but when deposited later
into public databases, it is unclear how these will be made transparent. For
example, depositing data in the public domain does not mean that there are
no constraints of licensing issues if the data are to be used for other purposes.
For instance, compounds with associated public malaria data, such as the
recently released Glaxo Smith Kline (GSK) data [13], may have been screened
against other targets and the owner may have patents or prior art on other
activities in which other scientists might be interested. There is thus a signifi cant
challenge as to how to make people aware of this. Should public depositors of
data be required to reveal all associated constraints simultaneously?
Some collaborations may be very narrow, focusing on a specifi c target or
molecule and requiring sharing of data on one project for only a defi ned time.
There may be certain boundary conditions that could inhibit further collabora-
tions. For example, if the collaboration was to be extended in new directions,
then there may be challenges regarding whether any software would be avail-
able to integrate and share data with additional systems outside of the original
collaboration. Software used for collaborations may not be integrated between
two or more parties so the process of connecting data between all the key
tools that may be used (e.g., chemistry and biology databases) can become an
issue. Computer-computer interactions may simplify or complicate the process
compared with human-computer interactions and hard-copy data sharing. The
lowest common denominator between people with their different types of data
before, during, and after a collaboration and their interactions with collabora-
tive software become very important issues. Chemical structures, experimental
data (both continuous and discrete), and computational models derived from
such data may all need to be shared. There are, as yet, no agreed standards for
quantitative structure-activity relationship (QSAR) model sharing while
there are many standards for sharing molecular structures [simplifi ed molecu-
lar input line entry specifi cation (SMILES), InChI, SDF, Mol, etc]. Despite the
fact that there are several sites that want to promote access to computational
models [e.g., Chembench (http://chembench.mml.unc.edu/), Ochem (http://
ochem.eu/ , http://ochem.eu/static/home.do ), and VCCLAB ( http://
www.vcclab.org/)], it is not yet clear whether any standards will emerge from
these sites. QSAR-ML has been recently proposed as an XML exchange
format for QSAR data, descriptors, software, and response data [14]. It will be
interesting to see if it is accepted by users of software beyond the open-source
Bioclipse workbench [15].
Even the manner in which data are uploaded into collaborative software
platforms presently could be standardized, and such simplifi cation combines
a required format and data organization and should be a catalyst for increased
collaboration by lowering the barrier to share data. Providing a simplifi ed data
upload standard would be a laudable fi rst goal of any CCTBR standards
development effort.
Search WWH ::




Custom Search