Quake: quality-aware detection and correction of sequencing errors

Table 5 E. coli SNP calling

Method	Reads mapped	SNPs	Recall	Precision
Two mismatch uncorrected	3.39 M	79,748	0.746	0.987
Two mismatch corrected	3.51 M	80,796	0.755	0.987
Quality-aware uncorrected	3.56 M	85,071	0.793	0.984
Quality-aware corrected	3.55 M	85,589	0.798	0.984

We called SNPs in 35× coverage of 36 bp reads from E. coli K12 by aligning the reads to a close relative genome E. coli 536 with Bowtie using both a two mismatch and quality-aware alignment policy and calling SNPs with SAMtools pileup. SNPs were validated by comparing the E. coli K12 and E. coli 536 reference genomes directly. Under both alignment policies, correcting the reads with Quake helps find more true SNPs.

ISSN: 1474-760X