Skip to main content
Figure 5 | BMC Bioinformatics

Figure 5

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

Figure 5

Two-way and three-way mutual information distributions for top models selected by MI3 and control methods. For each MYC target gene, the top 5 R2 or MYC cofactors were selected by applying different scoring metrics to the microarray dataset generated by Basso et al [15]. (a). I(R1;R2) vs. I(R2;T), i.e. two way mutual information between R2 and R1 or T, (b). I(T;R1, R2) vs. I(T;R1, R2)-I(T;R1)-I(T;R2), i.e. the correlative and coordinative components of MI3 score for the top 5 models selected by MI3 or control methods. Each ellipse represents the distribution of top 5 models in the specified mutual information coordinates, with mean as center and standard deviations as width and height. Note that I(R1;T) scores are the same for all methods hence not shown in (a).

Back to article page