Viral deep sequencing needs an adaptive approach: IRMA, the iterative refinement meta-assembler

BMC Genomics

Table 1 Average absolute error and standard deviation (parentheses) are shown for each group & parameterization.

				Allele type
Reference	Program	QC	Observations	Consensus	Minority
Flu gene segment consensuses	Bowtie2	-	1.73Gb	0.30 % (2.97 %)	0.15 % (2.22 %)
	Bowtie2	+	1.61Gb	0.28 % (2.86 %)	0.15 % (2.36 %)
	IRMA	-	1.99Gb	0.17 % (0.47 %)	0.08 % (0.29 %)
	IRMA	+	1.97Gb	0.12 % (0.42 %)	0.05 % (0.27 %)
Assembly controls: both unmixed genomes	IRMA^*	+	1.97Gb	0.12 % (0.40 %)	0.05 % (0.26 %)
Assembly controls: both unmixed genomes	Bowtie2	+	1.97Gb	0.12 % (0.40 %)	0.05 % (0.26 %)

Observations are the total number of assembled nucleotides for all genes and mixtures in triplicate. Allele type is per mixture-replicate-gene-position showing consensus (99.994% majority, 0.006% plurality) allele error and minority allele (non-consensus) error. Bowtie2 was set to very sensitive, local assembly.
*IRMA with a single round of read gathering iterations, final assembly using original references

ISSN: 1471-2164