Skip to main content

Table 1 Results of the bottom-up approach for modeling MITF regulation using Mixed Integer Linear Programming

From: SOX5 is involved in balanced MITF regulation in human melanoma cells

No. of TFs Predicted TFs Performance*
1 SOX5 0.83
2 ESR2, SOX5 0.87
3 ESR2, PAX2, SOX5 0.88
4 ESR2, NFKB1.1, PAX2, SOX5 0.89
5 ESR2, NFKB1.1, PAX2, SOX5, ZEB1 0.90
6 ESR2, NFKB1.1, ONECUT2, POU3F2, SOX5, ZEB1 0.91
7 ESR2, NFKB1.1, ONECUT2, PAX2, POU3F2, SOX5, ZEB1 0.91
8 ESR2, GLI2, NFKB1.1, ONECUT2, PAX3, POU3F2, SOX5, ZEB1 0.91
9 ESR2, GLI2, NFKB1.1, ONECUT2, PAX2, PAX3, POU3F2, SOX5, ZEB1 0.90
10 ESR2, GLI2, IRF1, NFKB1.1, ONECUT2, PAX2, PAX3, POU3F2,
SOX5, ZEB1
0.92
11 BHLHE40, ESR2, GLI2, IRF1, NFKB1.1, ONECUT2, PAX2, PAX3, POU3F2, SOX5, ZEB1 0.92
12 ESR2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1,
POU3F2, SOX5, SOX9, ZEB1
0.92
13 ESR2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1,
POU3F2, SOX5, SOX9, TCF4, ZEB1
0.91
14 BHLHE40, ESR2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX5, SOX9, TCF4, ZEB1 0.91
15 BHLHE40, ESR2, GLI2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX5, SOX9, TCF4, ZEB1 0.90
16 BHLHE40, ESR2, GLI2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX10, SOX5, SOX9, TCF4, ZEB1 0.89
17 BHLHE40, ESR2, GLI2, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX10, SOX2, SOX5, SOX9, TCF4, ZEB1 0.89
18 BHLHE40, ESR2, GLI2, IRF1, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX10, SOX2, SOX5, SOX9, TCF4, ZEB1 0.87
19 BHLHE40, CREB1, ESR2, GLI2, IRF1, LEF1, NFKB1.1, ONECUT2, PAX2, PAX3, PAX6, PDX1, POU3F2, SOX10, SOX2, SOX5, SOX9, TCF4, ZEB1 0.86
  1. *Averaged Pearson correlation of the model from the training data compared to the validation data