NextDenovo: an efficient error correction and accurate assembly tool for noisy long reads

Table 2 Statistics of nonhuman assemblies

Sample	Software	Assembly size (Mb)	NG50 (Mb)/LG50	NGA50 (Mb)/LGA50	No. of misassemblies	QV	Gene completeness (%)	Wall clock time (hour)
A. thaliana (452X)	NextDenovo	128.37	15.18/5	15.18/5	19	33.25	99.20	6.83
	Necat	124.55	15.01/5	14.98/5	44	31.93	99.20	6.82
	Canu	138.29	9.31/5	9.31/6	430	25.09	99.20	312.13
	Flye	121.16	14.63/5	14.63/5	17	35.65	99.20	12.00
	Wtdbg2	157.75	2.68/14	1.87/19	326	19.78	94.80	2.10
D. melanogaster (62X)	NextDenovo	134.34	18.11/4	15.68/4	196	30.99	98.70	1.07
	Necat	144.01	19.55/4	15.90/4	1,200	25.86	98.70	2.45
	Canu	154.94	8.58/6	5.68/7	1,738	23.53	98.80	45.55
	Flye	135.82	18.89/4	17.32/4	335	29.97	98.80	1.58
	Wtdbg2	137.49	6.32/7	5.33/9	919	26.07	97.20	0.57
O. sativa (230X)	NextDenovo	392.56	30.55/6	18.00/9	81	26.45	98.60	13.05
	Necat	394.40	25.44/7	17.86/9	183	25.83	98.70	10.85
	Canu	395.23	11.57/13	9.41/15	204	24.94	98.70	728.78
	Flye	403.45	11.10/14	7.84/18	115	24.76	98.70	25.02
	Wtdbg2	488.33	0.96/88	0.81/95	553	17.90	94.10	5.85
Z. mays (51X)	NextDenovo	2,118.82	44.44/17	37.90/21	700	20.74	98.20	75.90
	Necat	2,171.54	22.76/32	17.71/38	3,307	20.41	98.20	87.87
	Canu	2,240.87	0.65/950	0.62/995	6,284	19.14	98.10	1,741.77
	Flye	2,122.73	2.87/222	2.59/242	863	20.63	98.20	-
	Wtdbg2	4,068.86	0.07/11298	0.05/13848	22,258	14.07	97.00	-

NG50 is the length N that 50% of the reference genome is covered in contigs with length ≥ N. LG50 is the number of contigs with length ≥ NG50. NGA50 is an NG50 of aligned blocks that are obtained by breaking contigs at misassembly events and removing all unaligned bases. LGA50 is the number of aligned blocks with length ≥ NGA50. Misassemblies and QV are evaluated by QUAST, where QV is defined as \(-10\times {{\text{log}}}_{10}(\frac{\# {\text{mismatches}} {\text{per}} 100 {\text{kbp}} + \# {\text{indels}} {\text{per}} 100 {\text{kbp}}}{100 {\text{kbp}}})\). Gene completeness is represented by the complete BUSCO values. QV and gene completeness were evaluated using the polished assemblies and other metrics were evaluated using the raw assemblies. The genomes of A. thaliana, D. melanogaster, and O. sativa were assembled on the same computer with 60 CPUs and 504 GB RAM of memory. The Z. mays genome, assembled by NextDenovo, Necat, and Canu, was run on a computer cluster with 7 nodes each with 32 CPUs and 256 GB RAM and assembled by Fly and Wtdbg2 run on a fat computer node. Best results for each metric are highlighted in bold

ISSN: 1474-760X