followed by primary data analysis (version 1.1.1 chemistry and analysis software) on the Pacific Biosciences RS instrument following the manufacturer's recommendations. Secondary analyses, including read filtering, were performed by SMRT Analysis versions 1.3.1 (E. coli and P. falciparum) or 1.3.0 (R. sphaeroides). Because Pacific Bioscience's BLASR aligner does not currently support random placement of ambiguously aligned reads, alignment was performed using the BWA-SW long-read aligner [47] version 0.6.2 with parameters '-b5 -q2 -r1 -z20 -M -w200'. BWA-SW parameters were based on the software's suggested defaults for Pacific Biosciences reads, adding the '-z20' parameter for greater accuracy (validated in [48]), the '-M' parameter to generate only one primary alignment per read (the analysis ignores secondary alignments), and '-w200' to encourage the aligner to generate only one alignment per read. The aligner input files were the 'filtered_subreads.fastq' files produced by the standard resequencing protocol.
