Skip to main content
Figure 4 | BMC Bioinformatics

Figure 4

From: Learning transcriptional regulatory networks from high throughput gene expression data using continuous three-way mutual information

Figure 4

Two-way and three-way gene expression patterns and mutual information for representative top two-parent models inferred by MI3 and control methods for T = PSMD7 given R1 = MYC. The first three columns show the three-way and two-way gene expression patterns, and the fourth column the mutual information triangles. The bottom row shows the two-way expression pattern for PSMD7-MYC and the legend for mutual information triangle. This figure gives a concrete example for the difference between MI3 and control scores, echoing the results in Figure 5. For high throughput gene expression data, the BN and MI2 metrics both pick up models with high mutual information between parents and between either parent and the target. MI3 selected relationships with slightly lower I(T;R1, R2) but I(T;R1;R2) much higher than the BN and MI2 metrics.

Back to article page