FIGURE 10. Ensemble interpretation in terms of protein positions, contacts and configuration.

a, Localization volumes of all 456 proteins in the NPC (excluding the FG-repeat regions) in four different views. The diameter of the transport channel and the NPC are also indicated. The proteins are colour-coded according to their assignment to the six NPC modules³⁷. b, Contact frequencies for all pairs of proteins. The contact frequency of a pair of protein types is the fraction of structures in the ensemble that contains at least one protein contact between any protein instances of the two types. c, Contact frequencies between proteins in composite 40. Proteins are nodes connected by edges with the observed contact frequency as the edge weight (indicated by its thickness). Edges that are part of the maximal spanning tree are shown by thick blue lines; the maximal spanning tree is the spanning tree that maximizes the sum of the edge weights. All edges with a statistically significant reduction in contact frequency from their initial values implied by the composite data alone (P-value < 10^-3; Supplementary Information) are indicated by dotted lines with contact frequencies shown in red. d, Protein adjacencies for the whole NPC, with proteins as nodes and edges connecting proteins that are determined to be adjacent to each other. The edge weight is the observed contact frequency. e, Configuration of the proteins in composite 40. The location of a protein corresponds to the average position of the beads representing non-FG repeats of the protein. f, Configuration of Nic96 and the NPC scaffold proteins. g, Localization volume of Nic96 and the NPC scaffold proteins³⁷.