The reference sequence for each genome was downloaded from NCBI, and converted to run-length encoded [19] (RLE) form (S. tokodaii - NC_003106.2, B. amyloliquefaciens NC_014551.1, and D. maricopensis NC_014958.1). Each read was also collapsed to its RLE form (i.e. homopolymers collapsed into a single base). This reduces the influence of numerous homopolymer errors on the alignment thresholds, as well as making it easier to transition between flow and base coordinates.
