Lecture 8:  Tomography (part 2) 1. Interpretation of EM tomograms 2. Denoising algorithms 3. Segmentation approaches 4. Identification of features of interest 5. Subtomogram averaging techniques 6. Methods of EELS and EF-TEM Examples of EM Tomograms SNARE‐mediated membrane fusion Bharat, TAM et al. (2009) EMBO Rep., 15, 2014 Examples of EM Tomograms Bacterial flagellar motor Murphy, GE et al. (2006) Nature, 442, 1062 Examples of EM Tomograms Red Cell Cytoskeleton Nans, A et al. (2011) Biophys. J., 101, 2341 Examples of EM Tomograms Golgi apparatus March, BJ (2005) Biochim. Biophys. Acta, 1744, 273 Examples of EM Tomograms Polyribosomes in human glia cells Brandt, F et al. (2010) Mol. Cell, 39, 560 Examples of EM Tomograms Gag lattice of the immature HIV virion Briggs, JA et al. (2011) J.Mol.Biol., 410, 491 Schur, FK et al. (2015) Nature, 517, 505 Denoising algorithms Linear filters:  averaging neighborhood voxels  Gaussian filter or other function Median filters: local filters that estimate the voxel value based on the neighbors wavelet filtering  non‐linear anisotropic diffusion bilateral filtering no filter bilateral median NAD Segmentation Techniques • thresholding and masking • manual segmentation • watershed segmentation • segmentation with eigenvectors • segmentation using prior knowledge ‐ tubular structures ‐ membranes 3D Template Matching Further considerations: Missing wedge Local variance at each angle Peak detection Validation Subtomogram Averaging Subtomograms Modelling of subtomograms EELS & EF‐TEM Techniques 1) Expose specimen to mono-energetic electron radiation 2) Inelastic scattering in the specimen poly-energetic electron beam 3) Image-forming electrons are selected by scattering angle Diffraction contrast (interference of scattered and unscattered electrons) sample inelastically scattered e‐ elastically scattered e‐ Energy Filters In‐column energy filter (JEOL) Post‐column energy filter (Gatan) Advantages: • Less aberrations  • Larger scattering angles  • Larger fields of view Advantages: • Can fit any microscope • Recording of filtered and  unfiltered data is possible Zero‐Loss Imaging (EF‐TEM) Slit aperture is centered on the zero-loss peak of the EELS spectrum  typical slit width = 10-20 eV only electrons that suffered no energy loss in the specimen can pass  only elastically scattered electrons arrive at the electron detector Imaging of thick specimen: Microstructure of a Ti-Al-V alloy When inelastically scattered electrons also reach the image plane: - image is affected by chromatic aberration => unfocused image - image is blurred and background is diffused => low contrast => Energy filtering allows imaging of thick specimen (both materials and biological samples) Electron Energy Loss Spectroscopy (EELS) Imaging the energy‐dispersive plane of the energy filter onto the image plane 1) Identification of elements (EELS elements tables) 2) EELS Quantification (thickness measurements) EELS LOG-RATIO TECHNIQUE I0 … electrons in the zero‐loss peak It … electrons in the EEL spectrum  … mean free path for inelastic scattering t … specimen thickness  Electron Spectroscopic Imaging (ESI) chemical analysis by imaging with element-specific energy-loss windows Elemental Mapping Images • characteristic edges in the energy-loss spectrum • onset energy characteristic of atomic species • concentration of an element can be determined from EELS spectrum • subtract background for each pixel: three-window technique EF‐TEM Tomography Silica–alumina porous composite: 3D elemental mapping 6 images collected per each tilt image: zero‐loss image L23 edge of Al (59, 70, 81 eV) L23 edge of Si (99 and 110 eV) Ersen & Hirlimann, Microscopy, 1572‐79 EF‐TEM Tomography