Figure 3.

Pipeline dataflow. Finished genomic sequence is deposited in Gadfly, and then fed to the pipeline database, which manages jobs, dispatching them to the compute farm via PBS. When a job finishes, the pipeline database stores the output. BOP filters this output and exports GAME XML to Gadfly. A cycle of annotation consists of curators loading GAME XML into Apollo, either directly from Gadfly or from a data directory. Modified annotations are then written to a directory and loaded into Gadfly.