Figure 3

Illustration of how initial preprocessing and chimera check did not remove ‘PacBio chimeras’. (a) Read length distribution from one representative look of Marshall et al.[4] before and after initial sequence preprocessing, where a ‘look’ is the field-of-view used in fluorescence data collection. This particular look was from the sample containing multiplexed bacterial and archaeal amplicons from Day 91 of a microbial electrosynthesis cell: barcode 1 was from supernatant and barcode 5 from granules (read prefex m120705_051145…s1/…/ccs). Initial preprocessing removed reads with any of the following: average quality score <25, ambiguous base calls (‘N’), >8 homopolymers, length >350 bp, >1 mismatch to each primer, >1 mismatch to a barcode, or a chimera. After initial preprocessing, 1.5% of cleaned reads were ≥700 bp. (b) Motifs from representative reads (names listed with motif) of unexpected sizes are shown to scale, with the accession number to best blastn (nr) match to each segment written above segments.