Accurate prediction of metagenome-assembled genome completeness by MAGISTA, a random forest model built on alignment-free intra-bin statistics

Table 1 Datasets used in this study

Dataset	Name^a	Complexity^b	Input material^c	Sequencing output	Read source^d	Assembly tool^e	Binning method	Binning parameters^f
Training	HC227_Cc	227	gDNA evenly	2 × 150 bp PE total: 60 Gb	ERS5705986	SPAdes	CONCOCT	comp
	HC227_Ccc						CONCOCT	comp + cov
	HC227_Xcc						MaxBin	comp + cov
	HC227_Mc						MetaBAT2	comp
	HC227_Mcc						MetaBAT2	comp + cov
Test	BMock12_Mc	12	gDNA unevenly	2 × 150 bp PE total: 64 Gb	SRR8073716	SPAdes	MetaBAT2	comp
	BMock12_Mcc	12	gDNA unevenly	2 × 150 bp PE total: 64 Gb	SRR8073716	SPAdes	MetaBAT2	comp + cov
	Rinke_Mc	54	gDNA evenly	2 × 150 bp PE Total: 13 Gb	Rinke et al. [31]^b	SPAdes	MetaBAT2	comp
	Rinke_Mcc	54	gDNA evenly	2 × 150 bp PE Total: 13 Gb	Rinke et al. [31]^b	SPAdes	MetaBAT2	comp + cov
	MBARC-26_Mc	26	gDNA unevenly	2 × 150 bp PE total: 51.9 Gb	SRR3656745	SPAdes	MetaBAT2	comp
	MBARC-26_Mcc	26	gDNA unevenly	2 × 150 bp PE total: 51.9 Gb	SRR3656745	SPAdes	MetaBAT2	comp + cov
	ZymoCS_Mc	10	gDNA evenly	2 × 150 bp PE total: 3 Gb	ERR2984773	SPAdes	MetaBAT2	comp
	ZymoCS_Mcc	10	gDNA evenly	2 × 150 bp PE total: 3 Gb	ERR2984773	SPAdes	MetaBAT2	comp + cov
	Quince_Mc	210	Simulated reads unevenly	2 × 150 bp PE total: 180 Gb	Quince et al. [33]^b	MEGAHIT	MetaBAT2	comp
	Quince_Mcc	210	Simulated reads unevenly	2 × 150 bp PE total: 180 Gb	Quince et al. [33]^b	MEGAHIT	MetaBAT2	comp + cov

^aLetter code after underscore refers to binning method (upper case) and parameters (lower case)
^bNumber of strains in the mock
^cgDNA: genomic DNA, (un)evenly specifies the distribution of the individual inputs
^dSRR (Sequence Read Archive accession number), ERR (European Nucleotide Archive accession number)
^eSPAdes version 3.14, For MEGAHIT, assemblies were provided with the publication
^fComp, composition; cov, coverage

ISSN: 2524-6372