Table 3.

Comparison to alternative technologies. We obtain similar results to alternative techniques that used hundreds of pooled samples (Nielsen et al.) or potentially inaccurate binning approaches (Albertsen et al., Iverson et al.). We also analyze strains at the resolution of individual variants and haplotypes rather than strains or species.

	Our method	Nielsen et al.	Albertsen et al.	Iverson et al.	Sharon et al.
Sample type	Gut microbiome	Gut microbiome	Environmental	Environmental	Environmental
# of samples	1	18–396	2	1	3 independent
Seq. platform	Tru-seq SLR	Illumina WGS	Illumina WGS	SOLiD mate-pairs Sanger sequencing	Tru-seq SLR
Seq. amount	8 Gbp (long reads) ×40 (subassembly)	4.5 Gbp/sample	86 Gbp	59 Gbp	1.5 Gbp (long reads) ×40 (subassembly)
Analysis type	De-novo assembly; Phasing	Correlation across multiple samples	DNA extraction efficiency binning	Tetranucleotide binning	De-novo assembly
Resolution	Individual SNV	Strain	Species with diff. GC content	Family	Strain
Longest scaffold	3.9 Mbp	733 Kbp	3.6 Mbp	2.2 Mbp	<20 Kbp
Scaffold N50	49 Kbp	39 Kbp1	4.1 Kbp overall ~100 Kbp for top species	6.8 Kbp	8.2 Kbp
Bases assembled	656 Mbp	45 Mbp (genes) 35 Gbp (total)	423 Mbp	300 Mbp	500 Mbp/sample
# Variants	200K	n/a	n/a	n/a	n/a
# Haplotypes	5K	n/a	n/a	n/a	n/a