Image processing techniques for high-resolution structure determination from badly ordered 2D crystals

Abstract

Abstract2D electron crystallography can be used to study small membrane proteins in their native environment. Obtaining highly ordered 2D crystals is difficult and time-consuming.However, 2D crystals diffracting to only 10-12 Å can be prepared relatively conveniently in most cases. We have developed image-processing algorithms allowing to generate a high resolution 3D structure from cryo-electron crystallography images of badly ordered crystals.These include movie-mode unbending, refinement over sub-tiles of the images in order to locally refine the sample tilt geometry; implementation of different CTF correction schemes;and an iterative method to apply known constraints in the real and reciprocal space to approximate amplitudes and phases in the so-called missing cone regions. These algorithms applied to a dataset of the potassium channel MloK1 show significant resolution improvements to approximately 5Å.. CC-BY-NC 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint . http://dx.doi.org/10.1101/113415 doi: bioRxiv preprint first posted online Mar. 3, 2017; Biyani et al., Image processing techniques for high-resolution 2D electron crystallography 3

A Journey Towards Understanding Biology Holistically at the Nanoscale

“…An advantage of 2D-electron crystallography is that the crystals formed can be poorly ordered and still give valuable structural information by using image processing (179,180). While the real space image of the protein crystal is typically low contrast and has a high amount of noise, the power spectrum of the image shows the structural information in discrete spots from the periodicity of the crystal lattice (181). Fourier filtering can be used to remove excess noise from the image, at which point crystal defects and lattice distortions are observable (181).…”

“…While the real space image of the protein crystal is typically low contrast and has a high amount of noise, the power spectrum of the image shows the structural information in discrete spots from the periodicity of the crystal lattice (181). Fourier filtering can be used to remove excess noise from the image, at which point crystal defects and lattice distortions are observable (181). These distortions can be corrected through a lattice unbending process which uses cross correlation of a theoretical, or measured from a small subsection of the image, unit cell to calculate the shift vectors of the lattice distortions (181).…”

“…Fourier filtering can be used to remove excess noise from the image, at which point crystal defects and lattice distortions are observable (181). These distortions can be corrected through a lattice unbending process which uses cross correlation of a theoretical, or measured from a small subsection of the image, unit cell to calculate the shift vectors of the lattice distortions (181). These distortions can then be corrected for and, through iterations, as the lattice is refined the unit cell can be determined more accurately.…”