Comparison strategy used to construct the MOSAIC database. When at least two strains of a species are sequenced, genomes are first extracted from the GR database [step (1)]. Systematic intra-species pairwise genome alignments are then performed with MAUVE [step (2)]. A test for the presence of rearrangements in the pair of aligned genomes is then applied using the number and size of defined MAUVE LCBs (Locally Collinear Blocks) in step (3). The LCB analysis permits genomes to be designated either as collinear or rearranged. The collinear genomes are realigned with MGA (at least pairwise, and possibly with maximal multiple alignment if sequences of more than two strains are available). Rearranged genomes are aligned with MAUVE (maximal multiple alignments). Finally, MGA and MAUVE alignments are post-processed and genomes are segmented into backbone regions and variable segments in step (4), and integrated in the MOSAIC database together with annotations in step (5).