Fig 1: Molecular Modeling for Improved Image Processing (Super-STORM)(A–D) Side-view schematics of E-H+, E+H-, and E+H+ ß2-integrins with KIM127 and mAb24 Fabs bound (A), EM images (B), ribbon (C), and space filling (D).(E) Space-filling structure, top view; see also Videos S1, S2, and S3. EM images were adapted from Chen et al. (2010, 2012). The Fab of CBR LFA-1/2 in the EM images of E+H- and E+H+ and the ligand C3c in the EM image of E+H+ were removed with Photoshop for clarity (original images are shown in Figure S9). E-H+ ß2-integrin is from the published crystal structure (PDB: 4NEH) (Sen et al., 2013); E+H- is modeled by unfolding the headpiece of E-H- (PDB: 3K6S) (Xie et al., 2010) to E+. For E+H+ modeling, the hybrid domain swing out is superposed (PDB: 3FCU) (Zhu et al., 2008). a chain is in gray, and ß chain is in white. mAb24 and KIM127 Fabs are docked to their binding sites (Lu et al., 2001; Yang et al., 2004).(F and G) Distance distributions of localizations emanating from the same Fab (open) or from Fabs bound to adjacent integrin molecules (filled) of mAb24 (F) or KIM127 (G), based on simulations of 64,000 randomly oriented integrin molecules.(H) Zoomed-in pointillism map gated for areas within 50 nm of the substrate.(I) Pointillism map after removing multiple blinks from (H), based on the cutoffs indicated in (F) and (G). The fields of view in (H) and (I) are the same as in Figure 2C. Scale bars are 10 nm for (A)–(E) and 30 nm for (H) and (I). See also Figures S5 and S9–S11, Tables S2 and S3, and Videos S1, S2, and S3.