Gctf: Real-time CTF determination and correction.

Zhang K. · 2016

Read Gctf: Real-time CTF determination and correction. by Zhang K. (2016) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4711343.txt.

Introduction

Kai Zhang presented Gctf, a program for measuring a key imaging distortion in cryo-electron microscopy, in the Journal of Structural Biology in 2016. Cryo-EM rebuilds molecular structures from thousands of fuzzy particle images. Each image passes through the microscope's contrast transfer function, which scrambles detail in a known way. Estimating that function exactly is essential for near-atomic results, and doing it live is better still.

Gctf runs on graphics processors and matches simulated functions against each micrograph's spectrum. A two-step search adds speed, on top of gains the paper puts at ten to fiftyfold. It then refines focus values for single particles and single movie frames. Built-in consistency checks flag bad estimates. Outputs drop neatly into the field's standard packages, making Gctf a practical front step for high-resolution work.

Catalog record

CREATORZhang K.
YEAR2016
COLLECTIONPAPERS
FORMATTXT
FILE SIZE60 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC4711343

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Gctf: Real-time CTF determination and correction By Kai Zhang Published in: Journal of Structural Biology DOI: 10.1016/j.jsb.2015.11.003 ABSTRACT Abstract Accurate estimation of the contrast transfer function (CTF) is critical for a near-atomic resolution cryo electron microscopy (cryoEM) reconstruction. Here, a GPU-accelerated computer program, Gctf, for accurate and robust, real-time CTF determination is presented. The main target of Gctf is to maximize the cross-correlation of a simulated CTF with the logarithmic amplitude spectra (LAS) of observed micrographs after background subtraction. Novel approaches in Gctf improve both speed and accuracy. In addition to GPU acceleration (e.g. 10–50×), a fast ‘1-dimensional search plus 2-dimensional refinement (1S2R)’ procedure further speeds up Gctf. Based on the global CTF determination, the local defocus for each particle and for single frames of movies is accurately refined, which improves CTF parameters of all particles for subsequent image processing. Novel diagnosis method using equiphase averaging (EPA) and self-consistency verification procedures have also been implemented in the program for practical use, especially for aims of near-atomic reconstruction. Gctf is an independent program and the outputs can be easily imported into other cryoEM software such as Relion (Scheres, 2012) and Frealign (Grigorieff, 2007). The results from several representative datasets are shown and discussed in this paper. Abbreviations: 1D (2D, 3D), one(two, three) dimensional; cryoEM, cryo-electron microscopy; CCC, cross-correlation coefficient; CCD, Charge Coupled Device; CCF, cross-correlation function; CTF, contrast transfer function; DQE, detective quantum efficiency; EPA, equiphase averaging; FFT, Fast Fourier Transform; GPU, Graphic Processing Unit; HAV, hepatitis A virus; LAS, logarithmic amplitude spectra; NFS, Network File System; SNR, signal to noise ratio; SSD, Solid State Disk Keywords: Contrast transfer function, Cryo-electron microscopy, GPU program, CTF determination INTRODUCTION Recent progress has allowed cryo-electron microscopy (cryoEM) to determine structures of bio-macromolecules to near-atomic resolution (Nogales and Scheres, 2015). This is due to developments in multiple fields, but especially better detectors and image processing methods (Bai et al., 2015). The significantly improved detective quantum efficiency (DQE) of direct detectors, such as Falcon II and K2 summit, makes the quality of the cryoEM reconstructions much better than when using traditional CCD or film (Bai et al., 2015). Recording movies on these detectors allows motion correction of entire micrographs or individual particles, which makes critical improvements for high resolution reconstruction. More and more structures at near-atomic or atomic resolution are being solved recently by cryoEM (Amunts et al., 2015, Bartesaghi et al., 2015, Jiang et al., 2015, Paulsen et al., 2015, Taylor et al., 2015, Urnavicius et al., 2015, Zhao et al., 2015). In contrast to a simple projection of a 3-dimensional object, the cryoEM image of vitrified specimen is modulated by contrast transfer function (CTF). Because of the thin vitreous ice film, the image formation can be well described by weak-phase approximation (Wade, 1992). Based on this approximation, the phase contrast is dominant while the amplitude contrast is very small. Therefore, the major factors that affect the CTF of cryoEM image formation are the defocus and aberration of lens. The effect of these factors makes CTF a frequency-dependent oscillatory function, modulating both the amplitudes and phases of the image. Original information of the images must be corrected using accurate CTF parameters in order to obtain a reliable 3D reconstruction. The oscillation of the CTF becomes more severe at higher frequency or under a higher defocus. For this reason, information restoration from cryoEM image is quite challenging, especially at high frequency, which makes accurate CTF determination an important factor for near-atomic 3D reconstructions. There are currently several programs available for CTF determination (Ludtke et al., 1999, Mallick et al., 2005, Mindell and Grigorieff, 2003, Penczek et al., 2014, Shaikh et al., 2008, Sorzano et al., 2004, Tang et al., 2007, Vargas et al., 2013, Voortman et al., 2011). In a recent work, researchers systematically studied the performance of different programs (Marabini et al., 2015). Each of the programs has its own advantages for certain purposes. The popular program CTFFIND3 (Mindell and Grigorieff, 2003) shows the most self-consistent results using real datasets in this benchmark test, in spite of a slightly lower rank using simulated micrographs. However, with the fast development of cryoEM, a lot of new challenges are being required for daily image processing. One challenging requirement is to further improve CTF accuracy for 3D reconstruction at near-atomic or real atomic resolution. Higher speed without sacrificing the accuracy is also helpful to facilitate data processing with the development of automatic data collection at higher throughput. Besides, automatic self-consistency verification of the CTF determination and quality evaluation of the micrographs will greatly facilitate the ultimate goal of automation in cryoEM. Here a robust GPU-accelerated computer program called Gctf for CTF determination, refinement and correction is presented. GPU acceleration as well as an optimized programming strategy makes Gctf very fast. It can easily process thousands of micrographs within minutes using a single GPU card. The accuracy of the global CTF determination was verified by both manual scrutiny and automatic verification. Astigmatism-based rotational averaging or what is called equiphase averaging (EPA) in Gctf makes the visibility of Thon rings significantly improved for better diagnosis. Gctf was tested using a variety of parameters, showing stable ranges of parameter selection and thus its potential power for CTF automation of many types of micrographs. Micrographs from many datasets collected at the MRC-LMB (Cambridge) and several other collaborating institutes proved its accuracy, speed, convenience and robustness in practical use. In almost all cases, there was no need for parameter optimization. Gctf also performed well with a number of deliberately selected challenging micrographs. Local refinement and movie processing have also been implemented in Gctf. Local defocus refinement for each single particle makes significant improvements for 3D reconstructions carried out with datasets that have large defocus variation. Refinement of defocus of each frame in a movie provides a way of tracking frame movement in the Z-direction during imaging. Beside the determination and refinement of defocus in Gctf, CTF correction, automatic self-consistency verification and micrographs quality evaluation are also available for better automation of cryoEM data processing. THEORY AND METHODS Application, usage and typical input/output of Gctf 2.1. Gctf is basically designed to estimate the unknown CTF parameters of EM micrographs. A summary of capability in the current version (1.0) of Gctf is listed as follows: - (a)Determine overall CTF parameters of(a.1)each micrograph (basic application);(a.2)each particle stack from the same micrograph;(a.3)each movie stack by(a.3.1)coherent averaging;(a.3.2)incoherent averaging; - (a.1)each micrograph (basic application); - (a.2)each particle stack from the same micrograph; - (a.3)each movie stack by(a.3.1)coherent averaging;(a.3.2)incoherent averaging; - (a.3.1)coherent averaging; - (a.3.2)incoherent averaging; - (b)Refine user-provided CTF parameters of (a.1), (a.2), (a.3); - (c)Refine local CTF parameters based on (a) for(c.1)each frames from the same movie;(c.2)each particle(c.2.1)using coordinate files as input;(c.2.2)using particle stacks as input;(c.2.3)from auto-detection by(c.2.3.1)user provided templates(c.2.3.2)Gaussian convolution - (c.1)each frames from the same movie; - (c.2)each particle(c.2.1)using coordinate files as input;(c.2.2)using particle stacks as input;(c.2.3)from auto-detection by(c.2.3.1)user provided templates(c.2.3.2)Gaussian convolution - (c.2.1)using coordinate files as input; - (c.2.2)using particle stacks as input; - (c.2.3)from auto-detection by(c.2.3.1)user provided templates(c.2.3.2)Gaussian convolution - (c.2.3.1)user provided templates - (c.2.3.2)Gaussian convolution - (d)Auto-check CTF determination results after (a), (b) or (c) by(d.1)self-consistency verification;(d.2)evaluation of the micrograph quality;(d.3)overall defocus or astigmatism variation;(d.4)local defocus variation; - (d.1)self-consistency verification; - (d.2)evaluation of

Read the complete text