Table 3.

Coverage-based binning can enhance the recovery of individual genomes from metagenomes

Abundance class (Bin #)	% Binned reads in global assembly	% Binned reads from “key player”	Bin coverage (estimated by CB-Cl)	Bin coverage measured (from read mapping)	Bin size (estimated by CB-Cl)	Bin size measured (assembly length)	%Reads used in assembly of individual bins
Bin I	0.2 %	91.3 %	581	656	6.8 Mbp	6.5 Mbp	91.3 %
Bin II	0.5 %	1.3 %	317	306	7.0 Mbp	8.9 Mbp	86.3 %
Bin III	1.6 %	0.0 %	140	152	20.4 Mbp	21.2 Mbp	86.2 %
Bin IV	49.8 %	0.0 %	47	44	38.9 Mbp	37.6 Mbp	71.4 %
Bin V	37.7 %	0.0 %	14	17	117 Mbp	39 Mbp	30.4 %

Raw reads generated from a poly-aromatic hydrocarbon degrading enrichment culture [18] were assembled (with the ALLPATHS program [20]) globally on one hand, and segregated into 5 abundance classes (bins) followed by targeted assembly of reads from individual classes on the other hand. The assembly of reads with the highest coverage (Bin I) led to the reconstruction of a single 6.5 Mbp genome (key player), which is missing in the global metagenome assembly