Table 3.

Performance comparison of five module inference tools

Algorithm name	TP	FP	FN	Jaccard similarity score
HIMIA	7	5	25	0.19
ReMoDiscoery	11	38	21	0.16
GRAM	13	56	19	0.15
MOFA	8	32	24	0.13
SAMBA	10	54	22	0.12

Performance comparison was based on the Jaccard similarity score [12,34], which scores the overlap between an algorithm's output and the list of known stress-responsive TFs. Specifically, the score is defined as TP/(TP+FP+FN) [12,34], where TP stands for true positives, FP for false positives, and FN for false negatives. Note that the higher the Jaccard similarity score, the better the ability of an algorithm to retrieve the known stress-responsive TFs.