Skip to main content

Table 7 Parameter estimates with 95% confidence intervals for the Alon yeast regulatory network

From: Testing biological network motif significance with exponential random graph models

Effect Model 1 Model 2 Model 3
Arc \(\mathbf {\underset{(-7.665, -7.313)}{-7.489}}\) \(\mathbf {\underset{(-7.662, -7.318)}{-7.490}}\) \(\mathbf {\underset{(-7.667, -7.313)}{-7.490}}\)
Reciprocity \(\underset{(-15.535, 3.307)}{-6.114}\)
AltInStars \(\underset{(-1.504, 0.577)}{-0.463}\) \(\underset{(-1.451, 0.581)}{-0.435}\) \(\underset{(-1.472, 0.568)}{-0.452}\)
AltOutStars (\(\lambda = 4.5\)) \(\mathbf {\underset{(0.756, 1.261)}{1.008}}\) \(\mathbf {\underset{(0.757, 1.256)}{1.006}}\) \(\mathbf {\underset{(0.752, 1.264)}{1.008}}\)
AltTwoPathsT (\(\lambda = 3\)) \(\underset{(-0.739, 0.076)}{-0.332}\) \(\underset{(-0.670, 0.074)}{-0.298}\) \(\underset{(-0.684, 0.054)}{-0.315}\)
AltKTrianglesT (\(\lambda = 3\)) \(\mathbf {\underset{(0.484, 4.111)}{2.297}}\) \(\mathbf {\underset{(0.043, 3.640)}{1.842}}\) \(\mathbf {\underset{(0.479, 3.763)}{2.121}}\)
  1. Parameter estimates that are statistically significant are shown in bold. In Model 1 only, estimation is conditional on no reciprocated arcs, even though there is a single reciprocated arc in the data. Model 3 is included for illustration, even though it shows poor convergence with respect to the Reciprocity parameter (t-ratio magnitude is greater than 0.3)