Table 2.

The number of high-quality bins reconstructed by different binners for CAMI-high, medium, low complexity datasets and chicken gut datasets at purity greater than 0.9 and contamination less than 0.1

The number of reconstructed high-quality bins	CAMI-high datasets	Common strains of CAMI-high	Unique strains of CAMI-high	CAMI-medium datasets	Common strains of CAMI-medium	Unique strains of CAMI-medium	CAMI-low datasets	Common strains of CAMI-low	Unique strains of CAMI-low	Chicken gut metagenomic datasets
Gold standard	596	240	356	132	54	78	40	18	22	/
Groopm 2	*435	112	*323	89	32	57	*25	10	15	*83
MetaBat 2	*366	67	*299	77	27	50	*23	9	14	*87
MaxBin 2	236	20	216	75	21	54	*19	5	14	60
Solidbin	*403	85	*318	83	33	50	10	0	10	**
Vamb	364	69	295	53	12	41	13	2	11	59
MetaWatt	341	58	283	51	9	42	15	0	15	33
Binsanity-refine	35	2	33	16	7	9	18	4	14	27
Autometa	78	18	60	32	10	22	13	4	9	*73
BMC3C	40	0	40	22	0	22	7	0	7	64
COCACOLA	77	0	77	0	0	0	3	0	3	20
CONCOCT	71	2	69	62	9	53	15	0	15	66
Metabat (CAMI)	126	3	123	47	0	47	12	0	12	87
MyCC	56	3	53	45	4	41	14	0	14	20
Binsanity	0	0	0	1	0	1	2	0	2	***
DAStool	439	116	323	94	36	58	29	14	15	91
Binning-refiner	306	73	233	78	28	50	17	4	13	43
MetaWrap	427	104	323	91	32	59	22	7	15	110

*Binning results were used for the input of genome refining binner

**When Solidbin dealt with the chicken metagenomic co-assembly datasets containing more than 110 thousand contigs, it was too computing-intensive to get binning result (all the 112 threads and more than 500GB memory were used, finally Solidbin failed to return binning results)

***Binsanity provide a script Binsanity-lc comprising of binsanity and binsanity-refine to deal with the large metagenomic assemblies (> 100,000 contigs)