Table 1.

Important TF motifs discovered by PEP-Motif (but not by TargetFinder) to be of top 5% feature importance in at least two cell lines (the upper part) and those of top 10% feature importance in at least three cell lines (the lower part)

Cell lines	Potential novel predictive TF of top 5% importance (in PEP-Motif but not in TargetFinder)
2 (E)	(PRGR, ANDR, GCR), BHE23, (TBX4,TBR1,TBX21),
	CEBPE, EGR3, ENOA,(EVX1, FOXG1,FOXL1), FOXO1,
	HAND1, HBP1, HOXA1, MCR, (NFAT5,NFAC3), STAT6,
	(NR1I2, NR1I3), RHXF1, SMAD3, SOX10, SOX4,
	BRAC, TEAD3, ZNF713
3 (E)	EHF, ZSC16
2 (P)	AP2D, (KLF4, KLF1,KLF3), NKX32, PLAG1
3 (P)	(HOXD8, POU5F1B, POU2F2, POU3F3), (MNT,SPIC)

Cell lines	Potential novel predictive TF of top 10% importance (in PEP-Motif but not in TargetFinder)
3 (E)	BHE23, ERR2, ENOA, EOMES, FEV, GCM1,
	(HSF1, HSF2, HOXB2, HSF4), HOXA1, (KLF3, KLF1, KLF4),
	MCR, MITF, MYOD1, (RORG, RORA, NR1D1),
	(NR1I2, NR1I3), SMAD3, TEAD3, TF7L2, ZBT18, ZNF713
4 (E)	(ANDR, PRGR, GCR), EGR3, ETV5, HAND1, ZSC16
5 (E)	EHF
3 (P)	(HOXD8, POU5F1B, POU2F2, POU3F3), PLAG1
4 (P)	(KLF4, KLF1, KLF3), (MNT, SPIC)

Note: Each TF is represented by one or multiple motifs. If the corresponding motif is associated with a motif cluster, all members of the cluster are displayed and the motif reaching the specified importance level is in italic. ‘E’ represents enhancer regions and ‘P’ represents promoter regions. The row name represents the exact number of cell lines where the motif reaches the specified importance level, e.g. ‘2 (E)’ denotes that the feature in the enhancer region has top 5% importance in exactly two cell lines.