Table 1.

Comparative studies evaluating differences across enrichment methods

No.	Review	Methods tested	Datasets	Database (# of gene sets/pathways)	Types of evaluated methods
1	[13]	7	36	KEGG (116)	Topology- and non-topology-based methods
2	[2]	10	75	KEGG (323) and GO (4631)	ORA and FCS methods
3	[3]	7	118	KEGG (232)	Topology-based methods
4	[14]	6	20	KEGG (86)	Topology- and non-topology-based methods
5	[15]	9	3	KEGG (114)	Topology-based methods
6	[16]	13	6	GO gene set collection extracted from MSigDB [17] v6.1 (5917)	Widely used pathway enrichment methods
7	[18]	8	3	MSigDB v5.0 (10,295)	Widely used pathway enrichment methods
8	[9]	10	86	KEGG; 150 pathways for all methods except 130 for PathNet [19] and 186 for CePa [20, 21]	Topology- and non-topology-based methods
9	[22]	11	1	C2 collection from MSigDB v4.0 (4722)	Methods differing based on null hypothesis
10	[23]	16	42	KEGG (259) and Metacore™ (88)	ORA and FCS methods
11	[24]	5	6	KEGG (192)	ORA and FCS methods
12	[25]	7	38	KEGG (189)	ORA and FCS methods

In the third column, we report the number of enrichment methods compared in each study (see Supplementary Tables 2 and 3, available online at https://academic.oup.com/bib, for details on the methods tested). Here, we would like to note that we differentiate between methods and tools/web applications based on Geistlinger et al. [2]. In the fourth column, we report the number of datasets each study performed comparisons on, all of which were experimental datasets except in [3, 13, 14, 18, 22], which included both experimental and simulated datasets. Finally, the fifth column reports the pathway databases used in each study while the number of pathways is shown between parentheses.