Extended Data Fig. 4: Overview of single-particle segmentation and fusion of iPALM localization data.
From: Direct observation of the conformational states of PIEZO1
a, Workflow of iPALM data processing, particle segmentation, and super-particle fusion. b, Representative images (from n = 5 cells) showing bead removal and candidate triple labeled PIEZO1 segmentation on pre-processed localizations collected with the 640 nm laser. Top left, pre-processed localizations found at/near the plasma membrane rendered at 4 nm per pixel. Top middle, an image of the total raw photon data collected. The gold fiducial beads are apparent as bright spots since they are constantly emitting photons in each image frame. Top right, fiducial gold beads identified (magenta circles) by the bead removal algorithm from the total raw data image. Bottom left, total raw data overlaid with localizations showing removal of bead-associated localizations. Bottom middle, a 4 nm/pixel rendering of bead fiducial-purged localizations. These localizations are associated with AF647 fluorescence. Bottom right, binary AF647 localizations with candidate PIEZO1 molecules meeting nearest neighbor requirements highlighted with cyan boxes. Scale bars = 20 µm. c, The first 150 segmented triple labeled PIEZO1 particles identified from the segmentation algorithm meeting both nearest neighbor and inter-localization separation distance requirements. Each particle is shown in a 40x40 nm cyan bounding box. d, Overview of each of the three major steps in the particle fusion algorithm from Heydarian, et al.28. e, The total number of localizations per identified triple labeled PIEZO1 particle from all iPALM datasets (n = 5 imaged cells, n = 726 identified particles, n = 8500 total localizations). f, A 2D density grid estimation of the iPALM super-particle shown in Fig. 1c using a 1x1 nm binned grid and colored by local density.
