GreenHill: a de novo chromosome-level scaffolding and phasing tool using Hi-C

Table 4 Benchmarking results for zebra finch data

Assembler	Input for contigs	Size (bp)	Max Length (bp)	N50 (bp)	QV	Switch error rate	Phasing accuracy
Platanus-allee + GreenHill	10X, CLR	2,309,934,319	152,477,488	61,881,567	35.55	0.57	0.953
FALCON-Unzip + GreenHill	CLR	2,025,894,925	150,748,938	70,617,212	35.97	0.79	0.886
FALCON-Unzip + 3D-DNA		2,165,277,869	153,652,046	73,366,558	35.29	0.87	0.639
FALCON-Unzip + SALSA2		2,153,682,369	67,065,260	14,220,490	35.35	0.87	0.717
Canu + GreenHill	CLR	1,995,878,168	148,688,728	70,920,789	35.98	2.28	0.849
Canu + 3D-DNA		2,283,995,575	163,319,697	74,641,957	35.57	2.22	0.589
Canu + SALSA2		2,263,303,629	27,304,387	6,544,809	35.61	2.22	0.658
Hifiasm + GreenHill	HiFi	2,139,611,083	152,614,629	62,848,794	49.40	0.02	0.914
Hifiasm + 3D-DNA		2,667,759,650	77,682,004	1,858,283	49.83	0.02	0.929
Hifiasm + SALSA2		2,352,582,037	64,427,564	11,305,740	49.84	0.02	0.784
Hifiasm Hi-C mode		2,172,318,724	33,557,543	7,896,913	50.51	0.01	0.997

Size and N50 were calculated for sequences whose length ≥ 500 bp. Size represents the size of assemblies generated by each assembler. QV and switch error rate were calculated by Merqury. Phasing accuracy represents the proportion of the majority of hap-mers in a scaffold, and a high value of this indicator suggests large-scale haplotype consistency. A bold value indicates the best one for each input assembly

ISSN: 1474-760X