Evaluation of the performance of classification algorithms for XFEL single-particle imaging data

Yingchen Shi, Ke Yin, Xuecheng Tai, Hasan DeMirci, Ahmad Hosseinizadeh, Brenda Hogue, Haoyuan Li, Abbas Ourmazd, Peter Schwander, Ivan A. Vartanyants, Chun Hong Yoon, Andrew Aquila, Haiguang Liu

Research output: Contribution to journalArticle

Abstract

Using X-ray free-electron lasers (XFELs), it is possible to determine three-dimensional structures of nanoscale particles using single-particle imaging methods. Classification algorithms are needed to sort out the single-particle diffraction patterns from the large amount of XFEL experimental data. However, different methods often yield inconsistent results. This study compared the performance of three classification algorithms: convolutional neural network, graph cut and diffusion map manifold embedding methods. The identified single-particle diffraction data of the PR772 virus particles were assembled in the three-dimensional Fourier space for real-space model reconstruction. The comparison showed that these three classification methods lead to different datasets and subsequently result in different electron density maps of the reconstructed models. Interestingly, the common dataset selected by these three methods improved the quality of the merged diffraction volume, as well as the resolutions of the reconstructed maps.

Original languageEnglish (US)
Pages (from-to)331-340
Number of pages10
JournalIUCrJ
Volume6
DOIs
StatePublished - Jan 1 2019

Fingerprint

X ray lasers
Free electron lasers
free electron lasers
Lasers
X-Rays
Electrons
Imaging techniques
evaluation
Diffraction
x rays
Viruses
Diffraction patterns
Carrier concentration
Space Simulation
Neural networks
viruses
diffraction
Virion
embedding
diffraction patterns

Keywords

  • classification algorithms
  • electron-density map reconstruction
  • single-particle imaging
  • X-ray free-electron lasers (XFELs)

ASJC Scopus subject areas

  • Chemistry(all)
  • Biochemistry
  • Materials Science(all)
  • Condensed Matter Physics

Cite this

Evaluation of the performance of classification algorithms for XFEL single-particle imaging data. / Shi, Yingchen; Yin, Ke; Tai, Xuecheng; DeMirci, Hasan; Hosseinizadeh, Ahmad; Hogue, Brenda; Li, Haoyuan; Ourmazd, Abbas; Schwander, Peter; Vartanyants, Ivan A.; Yoon, Chun Hong; Aquila, Andrew; Liu, Haiguang.

In: IUCrJ, Vol. 6, 01.01.2019, p. 331-340.

Research output: Contribution to journalArticle

Shi, Y, Yin, K, Tai, X, DeMirci, H, Hosseinizadeh, A, Hogue, B, Li, H, Ourmazd, A, Schwander, P, Vartanyants, IA, Yoon, CH, Aquila, A & Liu, H 2019, 'Evaluation of the performance of classification algorithms for XFEL single-particle imaging data' IUCrJ, vol. 6, pp. 331-340. https://doi.org/10.1107/S2052252519001854
Shi, Yingchen ; Yin, Ke ; Tai, Xuecheng ; DeMirci, Hasan ; Hosseinizadeh, Ahmad ; Hogue, Brenda ; Li, Haoyuan ; Ourmazd, Abbas ; Schwander, Peter ; Vartanyants, Ivan A. ; Yoon, Chun Hong ; Aquila, Andrew ; Liu, Haiguang. / Evaluation of the performance of classification algorithms for XFEL single-particle imaging data. In: IUCrJ. 2019 ; Vol. 6. pp. 331-340.
@article{1e1d420a6b1443618fd082ad7e0c0615,
title = "Evaluation of the performance of classification algorithms for XFEL single-particle imaging data",
abstract = "Using X-ray free-electron lasers (XFELs), it is possible to determine three-dimensional structures of nanoscale particles using single-particle imaging methods. Classification algorithms are needed to sort out the single-particle diffraction patterns from the large amount of XFEL experimental data. However, different methods often yield inconsistent results. This study compared the performance of three classification algorithms: convolutional neural network, graph cut and diffusion map manifold embedding methods. The identified single-particle diffraction data of the PR772 virus particles were assembled in the three-dimensional Fourier space for real-space model reconstruction. The comparison showed that these three classification methods lead to different datasets and subsequently result in different electron density maps of the reconstructed models. Interestingly, the common dataset selected by these three methods improved the quality of the merged diffraction volume, as well as the resolutions of the reconstructed maps.",
keywords = "classification algorithms, electron-density map reconstruction, single-particle imaging, X-ray free-electron lasers (XFELs)",
author = "Yingchen Shi and Ke Yin and Xuecheng Tai and Hasan DeMirci and Ahmad Hosseinizadeh and Brenda Hogue and Haoyuan Li and Abbas Ourmazd and Peter Schwander and Vartanyants, {Ivan A.} and Yoon, {Chun Hong} and Andrew Aquila and Haiguang Liu",
year = "2019",
month = "1",
day = "1",
doi = "10.1107/S2052252519001854",
language = "English (US)",
volume = "6",
pages = "331--340",
journal = "IUCrJ",
issn = "2052-2525",
publisher = "International Union of Crystallography",

}

TY - JOUR

T1 - Evaluation of the performance of classification algorithms for XFEL single-particle imaging data

AU - Shi, Yingchen

AU - Yin, Ke

AU - Tai, Xuecheng

AU - DeMirci, Hasan

AU - Hosseinizadeh, Ahmad

AU - Hogue, Brenda

AU - Li, Haoyuan

AU - Ourmazd, Abbas

AU - Schwander, Peter

AU - Vartanyants, Ivan A.

AU - Yoon, Chun Hong

AU - Aquila, Andrew

AU - Liu, Haiguang

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Using X-ray free-electron lasers (XFELs), it is possible to determine three-dimensional structures of nanoscale particles using single-particle imaging methods. Classification algorithms are needed to sort out the single-particle diffraction patterns from the large amount of XFEL experimental data. However, different methods often yield inconsistent results. This study compared the performance of three classification algorithms: convolutional neural network, graph cut and diffusion map manifold embedding methods. The identified single-particle diffraction data of the PR772 virus particles were assembled in the three-dimensional Fourier space for real-space model reconstruction. The comparison showed that these three classification methods lead to different datasets and subsequently result in different electron density maps of the reconstructed models. Interestingly, the common dataset selected by these three methods improved the quality of the merged diffraction volume, as well as the resolutions of the reconstructed maps.

AB - Using X-ray free-electron lasers (XFELs), it is possible to determine three-dimensional structures of nanoscale particles using single-particle imaging methods. Classification algorithms are needed to sort out the single-particle diffraction patterns from the large amount of XFEL experimental data. However, different methods often yield inconsistent results. This study compared the performance of three classification algorithms: convolutional neural network, graph cut and diffusion map manifold embedding methods. The identified single-particle diffraction data of the PR772 virus particles were assembled in the three-dimensional Fourier space for real-space model reconstruction. The comparison showed that these three classification methods lead to different datasets and subsequently result in different electron density maps of the reconstructed models. Interestingly, the common dataset selected by these three methods improved the quality of the merged diffraction volume, as well as the resolutions of the reconstructed maps.

KW - classification algorithms

KW - electron-density map reconstruction

KW - single-particle imaging

KW - X-ray free-electron lasers (XFELs)

UR - http://www.scopus.com/inward/record.url?scp=85062655163&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062655163&partnerID=8YFLogxK

U2 - 10.1107/S2052252519001854

DO - 10.1107/S2052252519001854

M3 - Article

VL - 6

SP - 331

EP - 340

JO - IUCrJ

JF - IUCrJ

SN - 2052-2525

ER -