CONTRAST TRANSFER FUNCTION (CTF) CORRECTION FOR ELECTRON TOMOGRAPHY

点击翻译按钮获取中文摘要

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. With conventional optics (i.e., in the absence of a phase plate), image contrast in low dose cryo-ET is mainly achieved by defocusing the microscope significantly, typically by 3-8 microns. This large defocus shifts the first zero of the CTF to a much lower spatial frequency, and as a result the acquired tomogram does not have directly usable information beyond about 3-4 nm resolution due to phase inversion after the first zero. CTF correction is one of several complementary approaches to achieve 2-3 nm resolution in cryotomograms. We have developed a GUI program that can robustly determine the center defocus (thus the first zero of CTF) for a subset of views from a tilt series. Our method computes the 1D power spectrum of the images using periodogram averaging and then subtracts from it the quantum noise power spectrum of the microscope and CCD camera. To improve the signal to noise ratio, power spectra from strips at different Z heights are shifted into approximate registration with those from the center of the tilted images. The user interactively sets limits for fitting curves before and after the first zero, and those limits can then be applied when analyzing different subsets of views through the range of tilt angles. Experiments show that the computed defocus agrees well with the defocus reported by the SerialEM program, on both our F20 and F30 microscopes. A second program was written to reverse the phase inversion after the first zero of CTF in the center is determined. The correction method takes into account the defocus gradient along the direction perpendicular to the tilt axis. It tessellates an image into strips parallel to the tilt axis so that defocus variation within a strip is small comparing with the defocus. Then it corrects the strip assuming a constant defocus that is computed from the tilt geometry. At this stage we still can not quantify the resolution gain due to CTF correction. The plan is to first validate our correction algorithm using images from some known crystal protein, then apply it to our tomograms to see how much it can improve the final result of our particle averaging program.