Figure 1.

PPIs and protein domain compositions are parsimoniously organized in nature. Under each null model, 200 data sets are simulated. The distribution of T is shown as 'violin plot' and p-value is computed based on Gaussian distribution.

Chen et al. BMC Systems Biology 2012 6(Suppl 1):S7   doi:10.1186/1752-0509-6-S1-S7