Considerations for membrane thickness measurements in cryo-electron tomograms. (A) Area-weighted histograms of per-triangle thickness measurements of IMM thickness measurements performed on tomograms reconstructed at different pixel sizes. (B) Table of peak histogram values for membrane thickness calculations across distinct organellar membranes (OMM, IMM, and ER) performed on tomograms reconstructed at different pixel sizes. (C) Comparison of the resulting voxel segmentation output from MemBrain-Seg (Lamm et al., 2024, Preprint) performed on tomograms reconstructed at different pixel sizes 6.65, 9.98, and 13.3 Å/pixel. Voxel segmentations derived from tomograms reconstructed at 13.3 and 9.98 Å/pixel performed similarly accurately label the underlying membrane density visible in the tomogram. Voxel segmentations generated from tomograms reconstructed at 6.65 Å/pixel exhibit inaccuracies, such as voxels corresponding to membrane-protruding proteins mistakenly labeled as membrane. (D) 2D histogram of thickness measurement versus verticality of membrane shows no correlation between these measurements, suggesting that the missing wedge, which may artificially impact the thickness of membranes in these regions, is not driving significant errors in our data. (E) Triangulated surface mesh before edge filtering (left) and after edge filtering (right). The wire mesh shows the region of surface within 8 nm nearest to the edge of the membrane mesh that is artificially removed. Arrows highlight areas where measurements could not be made at all.