Network Medicine: Disease Module Drug Repositioning

Disease Module (A)

Module size 14

Candidate Drug Target Module (B)

Hop distance from disease seed 0

Network Topology

Attachment rate m (edges/new node) 2

Network-Based Separation

—

d_AB (closest distance)–

S_AB (separation score)–

Network avg degree ⟨k⟩–

Disease module A

Drug target module B

Other proteins

How it works

The graph is a Barabási–Albert network: nodes are added one at a time, each wiring m edges to existing nodes with probability proportional to their current degree (preferential attachment). This reproduces the scale-free, hub-dominated topology real protein–protein interaction networks show — a few "hub" proteins carry many more connections than average.

The disease module A is grown as a breadth-first neighbourhood of a random seed node (disease genes cluster locally in real interactomes). The drug target module B is sampled from the ring of nodes at exactly the chosen hop-distance from A's seed, using unweighted breadth-first search (BFS) distances.

For every pair of nodes the simulator computes the shortest-path (BFS) graph distance. The network-based separation between two modules follows Menche et al., Science 2015:

d_AB = closest-distance between A and B
     = 1/(|A|+|B|) · [ Σ_a min_b d(a,b) + Σ_b min_a d(a,b) ]

d_AA, d_BB = same closest-distance measure within each module

S_AB = d_AB − (d_AA + d_BB) / 2

S_AB < 0 — the modules overlap on the network: they sit closer to each other than to themselves on average. This is the topological signature used to flag drug-repositioning candidates (e.g. metformin's network proximity to several cancer disease modules).
S_AB > 0 — the modules are topologically separated; a drug hitting B is unlikely to perturb disease pathway A.

Module size slider changes |A|; the hop distance slider slides B from directly touching A (hop 0) to several hops away; the attachment rate slider rebuilds the whole network with a different hub density; Show shortest paths highlights the actual A↔B edges the closest-distance sum is computed from.

– nodes – edges

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.