Table 1. Datasets used in this study.

Two pooled datasets composed of multiple studies are abbreviated as CRC-16S [73–75] and CRC-WGS [1,74,76–79], whereas the American Gut Project (AGP) [34] and the Hispanic Community Health Study (HCHS) [72] are each from a single source study and have several potential confounders [33].

Phenotype	Joined dataset	Number of samples	Number of studies	Sequencing method	Published Sources
Body mass index	American Gut Project (AGP)	6,722	1 (multiple sequencing batches)	16S	[34]
Antibiotic history	American Gut Project (AGP)	12,619	1 (multiple sequencing batches)	16S	[34]
Body mass index	Hispanic Community Health Study (HCHS)	1,769	1 (multiple sequencing batches)	16S	[72]
Colorectal Cancer	CRC-16S	574	3	16S	[73–75]
Colorectal Cancer	CRC-WGS	813	7	WGS	[1,74,76–79]