BASE: a practical de novo assembler for large genomes using long NGS reads

BMC Genomics

Table 1 Contig assembly of deeply sequenced bacterial genomes

	Tools	Parameters	Correct N50	Misatch/Indel	Aligned rate	Coverage	Time(sec)
S.aureus MW2 (240X, 100 bp HiSeq)	SPAdes	51,63,85	299,305	134/6	99.79 %	100.00 %	1239
	SOAPdenovo2	87-95	82,495	0/0	99.84 %	99.27 %	25;16
	SGA	29;91	74,584	7/0	99.81 %	99.98 %	1228;1149
	BASE	4	92,706	0/0	100.00 %	99.97 %	161; 93
V.para (240X, 250 bp MiSeq)	SPAdes	33,55,65,75,85,99	169,978	118/45	99.97 %	99.97 %	4616
	SOAPdenovo2	125	88,858	23/30	99.98 %	99.98 %	110;1
	SGA	29;149	95,711	58/26	99.80 %	99.97 %	2478;2884
	BASE	4	159,715	29/29	100.00 %	99.75 %	676; 388

S.aureus MW2 has its real reference with length 2.8 Mb and V.para has its species’ reference with length 5.1 Mb and two chromosomes. Both of these two bacteria are sequenced up to 240X. GAGE validation pipeline was used to calculate the corrected contig N50, base errors, structural errors, contig aligned rate and reference coverage. Except BASE used single thread for contig assembly part, and other the assemblies were all performed with 24 threads. The time before semicolon is for index building and after semicolon is for assembly. For SGA, indexing time contains the time used in the indexing after error correction and filtering; assembly time contains the time used in the overlap and assembly

ISSN: 1471-2164