In a protein-protein interaction network, a hub protein has an unusually high degree (many interaction partners). If this protein is removed, what is the most likely consequence for the network?
ANo significant effect, because other proteins compensate immediately
BThe network fragments into disconnected components, because hub removal disproportionately disrupts connectivity in scale-free networks
COnly the hub protein's immediate neighbors lose function; the rest of the network is unaffected
DThe degree distribution shifts from scale-free to random, but connectivity is maintained
Scale-free networks are robust to random node removal but highly vulnerable to targeted removal of hubs. Because hubs connect many otherwise-distant nodes, losing a hub can disconnect large portions of the network. This has been validated experimentally: essential genes in yeast tend to encode hub proteins in the interaction network. Options (a) and (c) underestimate the cascading effect of hub loss; option (d) describes a statistical property change rather than the functional consequence.
Question 2 True / False
Scale-free biological networks are equally vulnerable to random node failures and targeted hub attacks.
TTrue
FFalse
Answer: False
Scale-free networks have a characteristic asymmetry in vulnerability. Random removal of nodes rarely hits a hub (because most nodes have few connections), so the network tolerates random failures well. However, targeted removal of the few high-degree hubs rapidly fragments the network. This 'robust yet fragile' property is a defining feature of scale-free topology and has implications for understanding genetic diseases (mutations in hub genes tend to be more severe) and drug target selection.
Question 3 Short Answer
Why is betweenness centrality sometimes a better predictor of a protein's biological importance than degree alone?
Think about your answer, then reveal below.
Model answer: Betweenness centrality measures how often a node lies on the shortest path between other nodes, capturing its role as a bridge or bottleneck in information flow. A protein with moderate degree but high betweenness connects otherwise-separated network modules and controls communication between them. Removing such a bottleneck protein can disrupt cross-talk between pathways even if it has fewer direct interactions than a hub. Degree captures local connectivity; betweenness captures global positional importance.
Classic examples include scaffold proteins and signaling adaptors that connect receptor signaling to downstream effector modules. These bridging proteins are often essential even when they have fewer direct partners than the highly connected hubs within each module.
Question 4 Multiple Choice
A researcher builds a protein-protein interaction network and finds it has a power-law degree distribution. She concludes that the network was shaped by preferential attachment during evolution. Is this conclusion justified?
AYes — power-law degree distributions can only arise from preferential attachment
BNo — multiple generative mechanisms (gene duplication, preferential attachment, sampling bias) can produce power-law-like distributions, and the topology alone cannot distinguish between them
CYes — if the distribution fits a power law, the Barabasi-Albert model must have generated it
DNo — because protein interaction networks never truly follow power laws
A power-law degree distribution is consistent with several generative models. Gene duplication followed by divergence naturally produces hub-enriched networks because duplicated genes initially share all interaction partners. Experimental sampling biases (well-studied proteins are tested against more partners) can also inflate apparent hubs. The Barabasi-Albert preferential attachment model is one mechanism, but not the only one. Distinguishing mechanisms requires evolutionary analysis, not just topology.