Figure 2

The example distributions of absolute correlation coefficients between regulatees of a regulator detected by different target functions. a) The average absolute correlation coefficient between regulatees in the real network (red) is significantly higher than it in the random network (blue). b) There is no significant difference between the absolute average correlation coefficients in the real network and the random network. However, there is a small bump at the right hand side, which means a small subset of highly correlated regulatees. The ratio scoring function was designed to detect such small subsets of regulatees.