skip to main content
research-article

A GPU acceleration of 3-D Fourier reconstruction in cryo-EM

Published: 01 September 2019 Publication History

Abstract

Cryo-electron microscopy is a popular method for macromolecules structure determination. Reconstruction of a 3-D volume from raw data obtained from a microscope is highly computationally demanding. Thus, acceleration of the reconstruction has a great practical value. In this article, we introduce a novel graphics processing unit (GPU)-friendly algorithm for direct Fourier reconstruction, one of the main computational bottlenecks in the 3-D volume reconstruction pipeline for some experimental cases (particularly those with a large number of images and a high internal symmetry). Contrary to the state of the art, our algorithm uses a gather memory pattern, improving cache locality and removing race conditions in parallel writing into the 3-D volume. We also introduce a finely tuned CUDA implementation of our algorithm, using auto-tuning to search for a combination of optimization parameters maximizing performance on a given GPU architecture. Our CUDA implementation is integrated in widely used software Xmipp, version 3.19, reaching 11.4× speedup compared to the original parallel CPU implementation using GPU with comparable power consumption. Moreover, we have reached 31.7× speedup using four GPUs and 2.14×–5.96× speedup compared to optimized GPU implementation based on a scatter memory pattern.

References

[1]
Abrishami V, Bilbao-Castro JR, and Vargas J, et al. (2015) A fast iterative convolution weighting approach for gridding-based direct Fourier three-dimensional reconstruction with correction for the contrast transfer function. Ultramicroscopy 157: 79–87. DOI: 10.1016/j.ultramic.2015.05.018
[2]
Blair J and Edwards C (1974) Stable rational minimax approximations to the modified Bessel functions I0(X) and I1(X): Technical Report AECL–4928. Chalk River, Ontario: Atomic Energy of Canada Ltd, Chalk River Nuclear Labs.
[3]
Crowther RA, DeRosier DJ, and Klug FRS (1970) The reconstruction of a three-dimensional structure from projections and its application to electron microscopy. In: Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences 317(1530): 319–340. DOI: 10.1098/rspa.1970.0119
[4]
Filipovič J, Petrovič F, and Benkner S (2017) Autotuning of OpenCL kernels with global optimizations. In: Proceedings of the 1st workshop on AutotuniNg and aDaptivity AppRoaches for Energy Efficient HPC systems (ANDARE ‘17) (ed Barbosa J), Portland, Oregon, USA, 9 September 2017, New York, NY, USA: ACM.
[5]
Henderson R (2015) Overview and future of single particle electron cryomicroscopy. Archives of Biochemistry and Biophysics 581: 19–24.
[6]
Jonic S, Sorzano COS, and Thévenaz P, et al. (2005) Spline-based image-to-volume registration for three-dimensional electron microscopy. Ultramicroscopy 103(104): 303–317.
[7]
Kimanius D, Forsberg BO, and Scheres S, et al. (2016) Accelerated cryo-EM structure determination with parallelization using GPUs in RELION-2. eLife 5: e18722.
[8]
Li X, Grigorieff N, and Cheng Y (2010) GPU-enabled FREALIGN: accelerating single particle 3D reconstruction and refinement in Fourier space on graphics processors. Journal of Structural Biology 172(3): 407–412. DOI: 10.1016/j.jsb.2010.06.010
[9]
Matej S and Lewitt RM (1995) Efficient 3D grids for image reconstruction using spherically-symmetric volume elements. IEEE Transactions on Nuclear Science 42(4): 1361–1370. DOI: 10.1109/23.467854
[10]
Penczek PA (2010) Chapter one – fundamentals of three-dimensional reconstruction from projections. In: Jensen JG (ed) Cryo-EM, Part B: 3-D Reconstruction, Methods in Enzymology, vol 482. Cambridge, USA: Academic Press, pp. 1–33. DOI: 10.1016/S0076-6879(10)82001-4
[11]
Punjani A, Rubinstein J, and Fleet DJ, et al. (2017) cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nature Methods 14: 290–296.
[12]
Radermacher M (1992) Weighted Back-Projection Methods. Boston: Springer US, pp. 91–115. ISBN 978-1-4757-2163-8. DOI: 10.1007/978-1-4757-2163-8_5
[13]
Sorzano COS, Vargas J, and Otón J, et al. (2017) A survey of the use of iterative reconstruction algorithms in electron microscopy. BioMed Research International 2017: 6482567.
[14]
Su H, Wen W, and Du X, et al. (2016) GeRelion: GPU-enhanced parallel implementation of single particle cryo-EM image processing. bioRxiv 075887. DOI: 10.1101/075887
[15]
Wang Z, Hryc CF, and Bammes B, et al. (2014) An atomic model of brome mosaic virus using direct electron detection and real-space optimization. Nature Communications 5: 4808.
[16]
Zhang X, Zhang X, and Zhou Z (2010) Low cost, high performance GPU computing solution for atomic resolution cryoEM single-particle reconstruction. Journal of Structural Biology 172(3): 400–406. DOI: 10.1016/j.jsb.2010.05.006

Author biographies

Author biographies
David Střelák holds MSc and BSc from Faculty of Informatics, Masaryk University. He is currently a PhD candidate at Universidad Autónoma de Madrid and Faculty of Informatics, Masaryk University. His research interests include high performance computing (heterogeneous computing, algorithm optimization techniques and autotuning) and image processing algorithms.
Carlos Óscar Sánchez Sorzano holds BSc and MSc in Electrical Engineering with two specialities (Electronics and Networking, University of Málaga), BSc in Computer Science (University of Málaga), BSc and MSc in Mathematics, (speciality in Statistics, UNED), PhD in Biomedical Engineering (Universidad Politécnica de Madrid), and PhD in Pharmacy (Universidad CEU San Pablo). In 2006, he received the Ángel Herrera research prize. He is a senior member of the IEEE since 2008 and that same year he was accredited as “profesor titular de universidad” by ANECA. In 2009, he was appointed as “Profesor Agregado” at Universidad CEU San Pablo, awarded a Ramón y Cajal research contract, and appointed as technical director of the INSTRUCT Image Processing Center for Microscopy. In 2011 and 2012, he was the President of the National Association of Ramón y Cajal researchers. He has been coordinating the service of image processing and statistical analysis of the CNB since 2011. In 2013, he was accredited as Full Professor. Since 2017, he is part of the permanent staff of CSIC.
José María Carazo holds MSc in Physics (University of Granada) and PhD in Molecular Biology (Autonomous University of Madrid). He joined the IEEE in 1982, being now a Senior Member. He performed his postdoctoral work at the Department of Health, Wadsworth Center, Albany, NY, USA, under the direction of Dr Joachim Frank (Nobel Laureate in Chemistry in 2017) from 1986 to 1989. In 1989, he set up the BioComputing Unit of the National Center for Biotechnology (CNB) in Madrid, that he heads since then. He is also the Director of the Instruct Image Processing Center and of the CSIC node of Elixir-Spain. He is full professor of Spanish CSIC.
Jiří Filipovič holds BSc and MSc in Applied Informatics (Masaryk University) with specialization on numerical computing and PhD in Informatics (Masaryk University). In 2012, he received the first prize in Joseph Fourier Award in Computer Science. After defending PhD, he worked as a postdoc at Masaryk University and University of Vienna. Since 2017, he has been the head of High Performance Computing research group in CERIT-SC Centre at the Institute of Computer Science, Masaryk University. His research interests include scientific and high-performance computing, in particular methods for auto-tuning, source-to-source code transformation, and heterogeneous computing and computational biology.

Cited By

View all

Recommendations

Comments

Information & Contributors

Information

Published In

cover image International Journal of High Performance Computing Applications
International Journal of High Performance Computing Applications  Volume 33, Issue 5
Sep 2019
303 pages

Publisher

Sage Publications, Inc.

United States

Publication History

Published: 01 September 2019

Author Tags

  1. Cryo-EM
  2. GPU
  3. CUDA
  4. 3-D Fourier reconstruction
  5. auto-tuning

Qualifiers

  • Research-article

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • Downloads (Last 12 months)0
  • Downloads (Last 6 weeks)0
Reflects downloads up to 18 Jan 2025

Other Metrics

Citations

Cited By

View all

View Options

View options

Media

Figures

Other

Tables

Share

Share

Share this Publication link

Share on social media