Strainline: full-length de novo viral haplotype reconstruction from noisy long reads

Table 1 Characteristics of benchmarking data sets

Virus mixture	Virus type	#Strain	Genome size (bp)	Coverage	Divergence (%)	Strain abundance (%)
Simulated
5-strain HIV	HIV-1	5	9478–9719	20,000 ×	2.7–5.6	10, 15, 20, 25, 30
6-strain Poliovirus	Poliovirus-2	6	7428–7460	20,000 ×	0.2–5.5	2, 4, 8, 16, 20, 50
6-strain Poliovirus (la1)	Poliovirus-2	6	7428–7460	20,000 ×	0.2–5.5	0.1, 1, 2, 8, 20, 68.9
6-strain Poliovirus (la2)	Poliovirus-2	6	7428–7460	20,000 ×	0.2–5.5	0.01, 0.1, 1, 2, 8, 88.89
10-strain HCV	HCV-1a	10	9273–9311	20,000 ×	2.8–7.4	5, 6, 7, 8, 9, 11, 12, 13, 14, 15
15-strain ZIKV	ZIKV	15	10,251–10,269	20,000 ×	1.1–15.1	2, 4, 5, 5, 5, 6, 6, 6, 7, 7, 8, 8, 9, 10, 12
5-strain SARS-CoV-2	SARS-CoV-2	5	26,574–29,903	20,000 ×	0.3–1.1	10, 15, 20, 25, 30
5-strain SARS-CoV-2 (la)	SARS-CoV-2	5	26,574–29,903	20,000 ×	0.3–1.1	0.1, 1, 5, 10, 83.9
Experimental
5-strain PVY (Mock)	PVY	5	9694–9701	5800 ×	3.6–21.6	9.3, 12.7, 21.1, 24.4, 32.5
SARS-CoV-2 (Real)	SARS-CoV-2	-	-	12,000 ×	-	-

For each benchmarking data set, we specify the name of virus mixture, virus type, number of strains in the mixture, range of genome size, total sequencing coverage, pairwise divergence, and strain abundance spectrum. The pairwise divergence is equal to 1−ANI, where ANI (Average Nucleotide Identity) is calculated by FastANI [31]. In experimental data sets, 5-strain PVY is a mock community, that is the sequencing data is real, but the mixture is synthetic, whereas SARS-CoV-2 (Real) is a real sample so there is no ground truth for the strains. The data sets 6-strain Poliovirus (la1) and 6-strain Poliovirus (la2) are similar with 6-strain Poliovirus, except the lowest abundance (la) of strains extends to 0.1% and 0.01%, respectively. The data set 5-strain SARS-CoV-2 (la) is similar with 5-strain SARS-CoV-2 except the lowest abundance (la) of strains extends to 0.1%

ISSN: 1474-760X