Table 1. Binning performance on various datasets (simulated reads, mock libraries and real samples).

		No. of bins	No. of binned contigs	Precision (%)	Recall (%)	F1 (%)
Simulated dataset	10 Genomes (2,185 contigsa)
	CONCOCT	19	2,185	98.78	97.67	98.2
	MaxBin	10	2,125	93.16	97.17	95.1
	MetaBAT	9	1,653	90.26	95.13	92.6
	MyCC (default)	10	2,185	97.79	97.79	97.8
	MyCC (one stage)	11	2,185	99.17	98.45	98.8
	100 Genomes (8,978 contigsa)
	CONCOCT	79	8,977	59.67	97.40	74.0
	MaxBin	84	7,308	89.64	84.52	87.0
	MetaBAT	105	5,430	92.72	89.59	91.1
	MyCC (default)	93	8,978	87.45	90.54	89.0
	MyCC (5p6 mer, cov)	88	8,978	89.68	94.09	91.8
Mock datasets	25 Genomes, 2 libraries (1,893 contigsa)
	CONCOCT	29	1,892	72.67	97.15	83.1
	MaxBin2	26	1,892	90.00	90.38	90.2
	MetaBAT	31	1,742	93.78	93.57	93.7
	MyCC (default)	23	1,893	88.97	97.35	93.0
	MyCC (4 mer, cov)	24	1,893	95.87	97.28	96.6
	64 Genomes (23,602 contigsa)
	CONCOCT	84	23,585	70.63	93.90	80.6
	MaxBin	56	20,639	84.96	81.83	83.4
	MetaBAT	70	8,722	86.78	77.40	81.8
	MyCC (default)	61	23,602	83.19	88.76	85.9
	MyCC (5p6 mer, cov)	57	23,602	84.36	92.85	88.4
Real dataset	Sharon’s dataset, 18 runs (2,294 contigsa)
	CONCOCT	32	2,291	79.92	97.58	87.9
	GroopM	13	1,687	88.39	86.29	87.3
	MaxBin2	10	2,294	82.94	93.75	88.0
	MetaBAT	10	1,573	85.46	93.66	89.4
	MyCC (4 mer, cov)	14	2,294	86.72	98.68	92.3

^aOnly contigs with a length longer or equal to 1,000 bp.