Flow diagram of the ISsaga pipeline. The figure shows how the different ISsaga functions are assembled. Following loading of the appropriate genome file, the system identifies ORFs using the ORF identification module. Module (a): if the file is pre-annotated, the protocol performs a BLASTP (filter off and e-value 1e-5) analysis followed by BLASTX (filter off and e-value 1e-5) to identify any ORFs that may have been overlooked. If the file is not annotated, an automatic Glimmer annotation is performed prior to BLASTP and BLASTX. Identified ORFs are included in a candidate ORF list. The replicon is then subject to BLASTN (filter off, word size 7 and e-value 1e-5) analysis, which yields an IS prediction and generates a web-based annotation table. If no ORFs are found, BLASTN is performed against the ISfinder database and any candidate ISs are fed into the IS prediction step. This step identifies partial ISs without ORFs. In a second module (b), ISs that have been identified and are already present in ISfinder are automatically fed into an IS report that must then be validated (module (c)). These modules are linked to the web interface (module (d)), which permits annotation management and provides tools for identifying and defining new ISs.