Extended Data Fig. 4: Subvolume averaging generates a plausible model of NEC lattice order.

Shown is a top view of a single perinuclear vesicle (also in Figs. 2, 3 and Supplementary Fig. 8). The surface (generated and visualised using Open3D) was coloured with the intersecting voxel densities of either a, volume where an NEC average volume was backplotted using subvolume averaging particle positions or b, c, the original data. Orange lines highlight the same areas in a and b where particles were removed due to their relatively low cross correlation coefficient. A long-range hexagonal order is apparent outside these areas. Assessing the nature of disordered regions is more challenging. To guide the reader’s eye, some densities in putative disordered regions were highlighted with red lines. A single hexagonal region was highlighted in blue for comparison. Note: A direct interpretation of tomogram densities on this scale can be misleading and should be used with caution. This example is intended to highlight that there is likely NEC in the disordered regions. What the structure of this lattice may be is not clear.