Our knowledge about the biochemical reaction taking place in a living
cell has increased to a point where we are able to map large parts of
their metabolic network. Moreover, advanced simulation techniques
enable us to simulate the growth of a cell on the metabolic level. To
grasp the topology of the metabolic network and to interpret the
fluxes through its reactions a visualization of the simulation results
is indispensable. However, traditional graph visualizations are unable
to cope with the complexity of metabolic networks and often very
cluttered graphs. In contrast, AMEBA reduces the number of nodes
displayed while retaining as much information as possible by
identifying branch points in the metabolic network. Although the
algorithm does not rely on extrinsic information, it can be guided
additionally by expert knowledge to produce camera-ready figures.
Furthermore, the run-time of less than a second -- even for large
metabolic networks -- enables an interactive usage. This way, a
researcher can track the fate of compounds through the whole network
and thus is able to get the big picture. The development of AMEBA was
tremendously facilitated by the availability of FLOSS. In fact, the
first prototype was created within a week and subsequently refined to
a stable analysis tool, which is now available under the terms of the
GPL (http://metano.tu-bs.de/ameba). The talk will cover the details of
the algorithm and will highlight the steps at which FLOSS was crucial
for the success of this piece of scientific software.