A random forest based biomarker discovery and power analysis framework for diagnostics research

BMC Medical Genomics

Table 1 A list of the published datasets used in this study

RF mode	Dataset type	Sample number (N)	Feature number (p)	Outcome variable	Pubmed ID	References
Regression	Metabolomics	73	196	Relative liver weight	28,185,575	[21]
Regression	Lipidomics	40	219	infant milk amount	28,190,990	[35]
Classification	Metabolomics	73	196	Relative liver weight class (below or above the mean value)	28,185,575	[21]
	Transcriptomics	68	414	Colorectal cancer (CRC) stages	27,176,004	[36]
	Transcriptomics	20	1386	Primary Sclerosing Cholangitis (PSC) vs. Ulcerative colitis (UC)	32,016,358	[37]
	Transcriptomics	40	25,697	OB/OB vs. wild type genotype mouse	32,646,215	[38]

For each of the RF models, two datasets was considered and the model outcome was compared with the published results

ISSN: 1755-8794