Skip to main content

Table 1 The impact of genome-wide signals on prediction accuracy (MOST vs. MOST+, K=9)

From: MOST+: A de novo motif finding approach combining genomic sequence and heterogeneous genome-wide signatures

TF ChIP-seq Peaks Ranking Predicted sites Co-factors
CTCF 39,609 1→1 27,458→43,150 0→2
ESRRB 21,647 1→1 17,144→18,998 3→4
KLF4 10,875 1→1 7,662→10,900 5→8
OCT4 3,761 1→1 2,051→2,802 3→7
cMyc 3,422 2→1 1,120→1,342 3→6
nMyc 7,182 2→1 1,853→2,519 3→4
SMAD1 1,126 9→4 119→135 4→8
E2F1 20,699 - - 3→5
NANOG 10,343 2→1 2,844→3,012 2→3
SOX2 4,526 1→1 3,515→3,490 3→6
STAT3 2,546 1→1 1,486→1,560 4→7
TCFCP2L1 26,910 1→1 1,4568→1,4780 3→7
ZFX 10,338 1→1 4,269→7,684 3→4
  1. Column 2-5: 2) the number of ChIP-seq peaks; 3) the change of ranks for the major TFBSs (from MOST to MOST+); 4) number of predicted binding sites for each TF; 5) the change of numbers of co-factors predicted by MOST and MOST+.