Table 2. Statistics for de novo assembly of CCS reads.

The “mixed” haplotype assemblies use all reads. The “maternal” and “paternal” assemblies use parent-specific reads from trio binning plus unassigned reads. HG002 concordance is measured against the Genome in a Bottle benchmark. BUSCO gene completeness uses the Mammalia ODB9 gene set. Ensembl genes is the percentage of genes from Ensembl R94 that are full-length, single-copy in the assembly relative to the full-length, single-copy count for GRCh38. Contigs shorter than 13 kb were excluded from genome size and contiguity measurements; contigs shorter than 100 kb were excluded from the concordance measurement. “*” indicates polishing with Arrow.

Haplotype	Assembler	Total size (Gb)	Contigs	N50 (Mb)	NG50 (Mb)	Max (Mb)	E-size (Mb)	HG002 concordance (Phred)	BUSCO genes	Ensembl genes
Mixed	Canu	3.42	18,006	22.78	25.02	108.46	30.16	31.1	92.3%	93.2%
Mixed	FALCON	2.91	2,541	28.95	24.51	110.21	38.04	25.8	87.6%	97.6%
Mixed	wtdbg2	2.79	1,554	15.43	12.62	84.67	22.61	44.6	94.2%	96.1%
Maternal	Canu*	3.04	5,854	18.02	17.04	48.81	19.78	47.2	94.1%	98.1%
Maternal	FALCON*	2.80	924	19.99	15.54	74.33	24.07	43.5	95.1%	97.8%
Maternal	wtdbg2	2.75	2,637	12.10	9.29	66.34	16.55	43.5	93.8%	95.6%
Paternal	Canu*	2.96	6,868	16.14	14.90	64.83	20.19	47.7	93.4%	98.2%
Paternal	FALCON*	2.70	1,489	16.40	14.06	95.34	25.61	43.5	93.6%	97.7%
Paternal	wtdbg2	2.67	1,444	13.96	10.86	50.51	15.36	42.1	92.6%	95.3%