Table 1.
The effect that targeted removal of interactions on network noise-processing.
| Interaction removed | Noise-processing (R/R full) | Coherence (c/c full) | Feedback (f/f full) |
|---|---|---|---|
| Tfcp2l1 → Sall4 | 0.02 | 0.68 | 0.64 |
| Sox2 → Oct4 | 0.12 | 0.60 | 0.37 |
| Sall4 → Klf2 | 0.18 | 0.98 | 0.45 |
| Sall4 → Sox2 | 0.18 | 0.86 | 0.18 |
| Stat3 → Gbx2 | 0.32 | 0.98 | 0 |
| Gbx2 → Klf4 | 0.32 | 0.98 | 0 |
| Nanog → Sox2 | 0.83 | 0.94 | 0.18 |
| Klf2 → Nanog | 1 | 1.10 | 0.27 |
| Esrrb → Tfcp2l1 | 1.12 | 0.95 | 0.36 |
| Nanog → Esrrb | 1.19 | 1.25 | 0.45 |
| Klf4 → Tfcp2l1 | 1.99 | 1.06 | 0 |
| Stat3 → Klf4 | 4.70 | 0.98 | 0 |
| Klf4 → Klf2 | 8.68 | 0.92 | 0 |
| Klf2 → Oct4 | 12.10 | 0.86 | 0.18 |
| Stat3 → Tfcp2l1 | 12.18 | 1.03 | 0 |
| Tfcp2l1 → Esrrb | 24.62 | 1.23 | 0.18 |
| Esrrb ⊣ Oct4 | 25.68 | 1.53 | 0.27 |
The first column identifies the edge removed from the network; the second column shows the effect of targeted removal of the given edge on the ratio R by comparison with that of the unperturbed network; the third column shows the effect of targeted removal of the given edge has on network coherence; the fourth column shows the effect of targeted removal of the given edge has on network feedback. Edges that emanate from Oct4 do not contribute to the noise processing capacity of the network and their removal does not affect R so they are excluded from this table. Since all paths from Lif to Oct4 pass through the edge Lif → Stat3 its removal disconnects the network; this edge is also accordingly excluded from the table. Interactions are ordered by column 1.