Biomedical Engineering Reference
In-Depth Information
and although some in the computational biology group have such a
capability there is still, for the moment, at least, the need to licence
GeneSpring;
■
JBuilder [35] - JBuilder is a commercial IDE for software development
with Java. This was used initially until other IDEs which were FLOSS
became available;
■
Pathway Studio - the Ariadne Pathway Studio provides data on pathway
analyses allowing, for example, the opportunity to view a gene of
interest in its biological context and perhaps draw further insights.
Lastly, there have been instances when a suitable FLOSS or paid-for
option was not available, requiring new software to be commissioned.
An example of this occurred when there was a requirement to add
additional functionality to our Java-based CytoSure™ software [37]. A
clear user need had been identifi ed to provide a method that could take
data from comparative genome hybridisations (CGH) and make a call as
to the copy number of the regions being analysed. Thus providing a
means to translate noisy microarray genomic data into regions of equal
copy number. There was a pre-existing algorithm available that could do
this but no Java implementation had been created. At times such as this,
the bespoke approach is the best way to proceed as this provides the clear
advantage of having something tailored to an exact specifi cation.
Whereas the tools described above fall into the 'install and use'
category, there are also many instances when more specialised tools need
to be developed. In particular, BioPerl [38] modules for Perl, the equivalent
Java library, BioJava [39] provide a fi rm foundation for development in
the bioinformatics area. Similarly, the BioConductor [40] library is
invaluable for rapid development of R-based workfl ows. Although OGT
has custom software written in a variety of languages, Perl is perhaps the
most common. There are several reasons for this; fi rst, it is one of the
easier languages to learn; new colleagues and interns with little prior
programming experience can make substantial contributions in a
relatively short period of time. Second, the BioPerl set of computational
tools is the most mature, often providing much of the core requirements
for new scripts. As Perl is predominantly a scripting language, when there
is a requirement for a user interface then Java has proved more applicable.
Figure 10.4 shows a screenshot from internally developed software
designed to view data from ChIP-on-chip microarrays. Our ability to
create such software is greatly increased by the availability of free libraries
such as BioJava, which was used extensively in this example to process
the underlying data.
Search WWH ::
Custom Search