Biomedical Engineering Reference
In-Depth Information
contigs. AnnoJ is limited in the feature types it can display and is extremely
infl exible with regard to styles, no real documentation on styling the
features exists and AnnoJ is closed source, the actual JavaScript must also
be retrieved over the internet as a web service. The core library for AnnoJ
cannot be downloaded and is 'mini-fi ed', that is compressed in such a way
that download times are reduced but the code in it cannot be read by a
human. In truth, we actually began using AnnoJ with help from the
author, who was a collaborator of a collaborator, so got a head start on
setting it up. Sadly, nowadays, the author seems to have gone quiet on the
project. Despite these drawbacks, we still use it. In particular, it has found
utility as a quick browser option for our gene feature and assembly
versioning database, Gee Fu [21]; as a pure web service it does have that
fl exibility we see in no other browser (as Gee Fu is a Rails application, it
is automatically a web service and together the two provide a very quick
way to provide feature data over a network for many users in an easily
extended platform). JBrowse is superfi cially similar to AnnoJ, as it is also
a Javascript project that renders gene feature data in a fast and attractive
way. The main downside to JBrowse is that the data fi les (JSON fi les)
have to be pre-generated and are not stored in a database. This includes
images representing features, which is fi ne for multi-use elements like
gene models but coverage information represented by histograms is not
dynamically generated and PNG fi les for these must be pre-rendered too.
Documentation provided with web applications and tools does not
necessarily cover all installation problems that you may have in your
particular setup. The FireBug plug-in for Firefox (and the developer tools
in Chrome to a lesser extent) are extremely useful for de-bugging the
requests made between browsers, CGI scripts/databases and web servers.
Usability is a constant concern when it comes to the tools that we
present to the biologist, and as far as possible we stick to graphical tools.
Free and open source software that has a graphical user interface never
comes along with iPod-esque levels of ergonomic or aesthetic design
fi nesse, but we have not come across any whose main interface is so
terrible that no analysis can be done with it. Phylogenetic analysis
software seems to be among the worst culprits for poor usability. The
strength of free and open source software is actually in its limitations;
quite often the software does not try to do too many things and this
makes the workfl ow within the program straightforward, so most
packages are at least usable for their main task. Sadly, programs with
graphical interfaces are in the minority. We fi nd that in practice, after a
bit of training, the worst symptoms of 'command-line aversion syndrome'
can be fought and biologists are not put off the command-line interface,
Search WWH ::
Custom Search