Biomedical Engineering Reference
In-Depth Information
In many cases, open source software is used for drug discovery and
product development. Examples of such software include general data
analytic tools as well as industry-specifi c bioinformatic and
cheminformatic and pharmacogenomic tools (examples include KNIME
[12], see Chapter 6, JAS3 [13] and RDKit [14], Bioconductor [15],
BioPerl [16], BioJava [17], Bioclipse [18], EMBOSS [19], Taverna
workbench [20], and UGENE [21], to name a few among many).
Software applications used in drug discovery require no such formal
validation, which is just as well as the support from different development
communities can vary signifi cantly.
However, in some other cases the level of support is very good and
allows such open source software to be used for applications where
validation is required. In one example of good practice, the R Foundation
[22] not only provides good documentation for their open source
statistical programming application, but has also worked in conjunction
with compliance and validation experts and the US FDA to:
■
produce white papers providing guidance on how to leverage
community documentation and software tools to help validate their
software;
■
provide guidance on how to control their software in an operational
environment and maintain the validated state;
■
address questions on the applicable scope of Electronic Records and
Signatures (US 21CFR Part 11 [23]).
Although not solely intended for use in the life sciences industry, these
issues were of common concern to a signifi cant number of users within
the open source community, allowing people to work together to provide
the necessary processes and guidance to support the validation of the
software. This level of support has allowed the use of this open source
software to move from non-validated use in drug discovery, into validated
areas such as clinical trials and manufacturing. In other cases the open
source community may only be loosely organised and may have no
signifi cant interest in supporting users in the life sciences industry.
Although this does not rule out the validation and use of such software,
regulated companies should realise that they will incur a signifi cantly
higher cost of validation when compared to more organised and
supportive communities.
The cost of such validation support needs to be evaluated and this is
best achieved by identifying the 'gaps' left by the community and
estimating the activities, documentation and costs required to initially
Search WWH ::
Custom Search