The increase in modularity with Venv can be attributed to reactions that are close to nutrients. The sampled genotypes with Venv = 89 typically have additional FCSs compared to sampled genotypes with Venv = 1. The reactions in these additional FCSs are not typical of the whole reaction network, and instead cluster at small distances from the nutrients. (See Methods for the determination of these distances using the Scope algorithm.) The distribution of distances for these reactions in additional FCSs is clearly concentrated at much smaller values than the distribution for all possible reactions; a Kolmogorov-Smirnov (K-S) test yields a p-value of 10-5, allowing us to reject the hypothesis that the two distributions are the same. Furthermore, a two sample Welch t-test gives a p-value of 8 × 10-8, allowing us to also reject the hypothesis that the mean of the two distributions are the same. a) Distribution of scope distances from the nutrients for all possible reactions. b) Distribution of scope distances from the nutrients for those reactions belonging to additional FCSs that differentiate the sampled genotypes at Venv = 89 from those at Venv = 1.
Samal et al. BMC Systems Biology 2011 5:135 doi:10.1186/1752-0509-5-135