Reducing assembly complexity of microbial genomes with single-molecule sequencing

Table 2 Genome assembly continuity and correctness using hybrid and self-correction approaches

Organism	Corrected by	Assembly bp	Number of contigs (expected)	Number of contigs (actual)	N50 (expected)	N50 (actual)	LAP	Number of discordant bases	QV
E. coli K12	Reference	4,639,675		1	4,639,675	NA	-9.65E + 07	4	>60
	MiSeq 100×	4,647,253	1	2		2,367,319	-9.64E + 07	3	>60
	454 50×	4,649,004	1	1		4,649,004	-9.64E + 07	3	>60
	CCS 25X	4,653,267	1	1		4,653,267	-9.64E + 07	3	>60
	Self	4,653,486	1	1		4,653,486	-9.64E + 07	3	>60
E .coli O157:H7	Near neighbor	5,594,477		3	3,776,951	NA	-3.82E + 07	1,282	36.40
	MiSeq 100×	5,624,394	10	10		3,089,011	-3.66E + 07	4	>60
	454 40×	5,613,057	10	12		927,294	-3.67E + 07	13	56.35
	Self	5,611,389	10	9		4,324,437	-3.66E + 07	0	>60
B. trehalosi	MiSeq 100×	2,402,545		6		1,603,511	-3.28E + 07	1	>60
	454 50×	2,413,761		4		1,051,672	-3.27E + 07	2	>60
	CCS 25X	2,411,501		1		2,411,501	-3.27E + 07	0	>60
	Self	2,411,068		1		2,411,068	-3.27E + 07	0	>60
M. haemolytica	MiSeq 100×	2,712,467		1		2,712,467	-3.31E + 07	0	>60
	CCS 25X	2,739,949		2		2,686,992	-3.31E + 07	0	>60
	Self	2,736,037		1		2,736,037	-3.31E + 07	0	>60
F. tularensis	Near neighbor	1,895,727		1	965,253	NA	-1.33E + 07	113	42.25
	MiSeq 100×	1,879,071	3	10		357,518	-1.33E + 07	0	>60
	454 50×	1,863,947	3	15		201,203	-1.33E + 07	0	>60
	Self	1,828,135	3	8		401,731	-1.33E + 07	0	>60
	Self (300×)	1,877,407	3	3		573,021	-1.33E + 07	0	>60
S. enterica Newport	Near neighbor	5,007,719		2	4,827,641	NA	-2.26E + 07	20	53.99
	MiSeq 56X	5,027,784	4	2		4,918,796	-2.24E + 07	2	>60
	454 25X	5,034,500	4	3		4,095,943	-2.24E + 07	2	>60
	CCS 22X	5,030,885	4	2		4,921,886	-2.24E + 07	2	>60
	Self	5,029,197	4	2		4,919,684	-2.24E + 07	2	>60

Organism: the genome being assembled. Corrected by: the short-read data used for correction. Assembly bp: the total number of base pairs in all contigs (only contigs containing at least 100 reads are included in all results). Number of contigs (expected): predicted number of contigs for a known reference (or near-neighbor). Number of contigs (actual): the number of contigs comprising the assembly. N50: N such that 50% of the genome is contained in contigs of length ≥N. LAP: the assembly likelihood score. A score closer to zero indicates a better assembly. Number of discordant bases: the number of SNPs and indels identified by mapping MiSeq sequences back to the assembly and recording discrepancies. Each incorrect base is counted (that is, an indel that is a deletion of two bases from the assembly counts as two in this column). QV: estimated from the number of discordant bases as ${log}_{10} (\frac{assembly length}{# incorrect bases}) * 10$ . The QV can be converted to an error probability P=10^(-QV/10). Assemblies were generated by Celera Assembler [31] followed by post-processing with Quiver [32]. NA, not available.

ISSN: 1474-760X