Fig. 4

Improved sequence fidelity using updated methods. Homopolymeric regions that were compressed in the original MinKNOW-called sequence (top) were expanded when the base caller Albacore included a transducer mode that incorporates signal length into the called sequence (bottom). The consensus sequence from the final assembly (from contig tig00023109) is shown together with the same set of individual reads, called with MinKNOW (top) or Albacore I (bottom). Sanger sequencing of PCR amplicons confirmed the accuracy of the newer base-calling, specifically the stretch of nine A’s, marked by the black bar, which was not supported by a single read in the MinKNOW data. Some of the sequence variation here may reflect genuine genotypic variation and not simply sequencing errors