Table 1.

Read error correction frequencies

Quality (Geometric Mean)	0.9 - 1.0	0.6 - 0.9
Error-free Input and Output	544 669	21 095
All Errors Retained	4 023	4 675
Input Errors Reduced	50 082	27 668
Errors Introduced	0	37
Total	598 774	53 475

Summary of error frequencies in assembled Illumina paired-end reads generated from sequenced V3-region amplicons of Methylococcus capsulatus strain Bath.

All error data were analyzed solely within the region of overlap, which was relevant to PANDAseq assembly. Low-abundance "contamination" was observed in the dataset (data not shown), possibly due to reagents used for PCR. These will contribute to the counts of sequences that had errors that were retained. This category will also contain sequences in which both reads contain low-quality bases with quality scores masked by CASAVA.