Skip to main content
Figure 4 | BMC Systems Biology

Figure 4

From: Modeling protein network evolution under genome duplication and domain shuffling

Figure 4

Combining whole genome duplication and domain shuffling of multi-domain proteins. A. Protein-domain interaction network. Nodes now correspond to single binding domains in a protein-domain interaction network (solid lines). Multi-binding-domain proteins are introduced through a new type of links corresponding to covalent peptide bonds between protein domains (black dashed lines). This provides a graphical framework to distinguish mutually exclusive, direct interactions ("XOR") between protein domains from cummulative, indirect interactions ("AND") within multi-protein complexes (red dashed lines). B&C. Comparison with protein direct & indirect interaction data for Yeast from BIND [38] database (B&C filled symbols, indirect interactions from [75,76]) and Ref [77] (C open symbols, see Supporting Information). Data are statistically averaged as in Fig. 3B&C to account for gaps in connectivities for large k ≥ 20, due to the finite size of Yeast PPI network. B. Two-parameter fit of both direct connectivity distribution p k and average direct connectivity of first neighbor proteins g k [50] (see Fig. 3C and text). Numerical predictions are averaged over 1,000 network realizations (central green and blue lines). Numerical averages plus or minus two standard deviations are also displayed to show the predicted dispersions (upper and lower green and blue lines). The two adjusted parameters (γ = 0.1 and λ = 0.3) correspond to a network growth rate of 20% and an average of 1.5 protein-binding sites (domains) per protein. The connectivity distribution of the underlying single-domain network (corresponding to γ = 0.1 and λ = 0.0) is also displayed (brown line) to illustrate its relation to the full multi-domain protein network (see text). C. Two-parameter fit of both direct & indirect "matrix" connectivity distribution p k and average direct & indirect "matrix" connectivity of first neighbor proteins g k [50] (see text). Same two adjusted parameters (γ = 0.1 and λ = 0.3) as in B while a selection of indirect interactions is added up to a total of 28,000 direct and indirect interactions (see Supporting Information).

Back to article page