Chapter 7
Multi-body Refinement of Cryo-EM Images in RELION
Takanori Nakane and Sjors H. W. Scheres
Abstract
Single-particle analysis of electron cryo-microscopy (cryo-EM) images allows structure determination of
macromolecular complexes. But when these molecules adopt many different conformations, traditional
image processing approaches often lead to blurred reconstructions. By considering complexes to be
comprised of multiple, independently moving rigid bodies, multi-body refinement in RELION enables
structure determination of highly flexible complexes, while at the same time providing a characterization of
the motions in the complex. Here, we describe how to perform multi-body refinement in RELION using a
publicly available example. We outline how to prepare the necessary files, how to run the actual multi-body
calculation, and how to interpret its output. This method can be applied to any cryo-EM data set of flexible
complexes that can be divided into two or more bodies, each with a minimum molecular weight of
100–150 kDa.
Key words Cryo-EM, RELION , Single-particle analysis, Image classification, Molecular motions,
Structural biology, Regularized likelihood optimization
1 Introduction
Relative motions between different domains underlie the functional mechanisms of macromolecular complexes that are involved
in many processes in the living cell. Structural characterization of
these complexes in their multiple states is key to understanding how
they work. Electron cryo-microscopy (cryo-EM) allows visualization of individual complexes in a thin layer of vitreous water. 3D
classification of the resulting particles into subsets of structurally
homogeneous complexes can then be used to obtain atomic structures of these complexes in their different states [1]. However,
subdivision of the cryo-EM data into a discrete number of subsets
is not well suited for the description of continuous types of molecular motion, for which an infinite number of subsets would, in
principle, be needed.
We recently proposed multi-body refinement as an alternative
to discrete classification to describe continuous molecular motions
in cryo-EM projection images [2]. The underlying model of
Tamir Gonen and Brent L. Nannenga (eds.), CryoEM: Methods and Protocols, Methods in Molecular Biology, vol. 2215,
https://doi.org/10.1007/978-1-0716-0966-8_7, © Springer Science+Business Media, LLC, part of Springer Nature 2021
145
Multi-body Refinement of Cryo-EM Images in RELION
Takanori Nakane and Sjors H. W. Scheres
Abstract
Single-particle analysis of electron cryo-microscopy (cryo-EM) images allows structure determination of
macromolecular complexes. But when these molecules adopt many different conformations, traditional
image processing approaches often lead to blurred reconstructions. By considering complexes to be
comprised of multiple, independently moving rigid bodies, multi-body refinement in RELION enables
structure determination of highly flexible complexes, while at the same time providing a characterization of
the motions in the complex. Here, we describe how to perform multi-body refinement in RELION using a
publicly available example. We outline how to prepare the necessary files, how to run the actual multi-body
calculation, and how to interpret its output. This method can be applied to any cryo-EM data set of flexible
complexes that can be divided into two or more bodies, each with a minimum molecular weight of
100–150 kDa.
Key words Cryo-EM, RELION , Single-particle analysis, Image classification, Molecular motions,
Structural biology, Regularized likelihood optimization
1 Introduction
Relative motions between different domains underlie the functional mechanisms of macromolecular complexes that are involved
in many processes in the living cell. Structural characterization of
these complexes in their multiple states is key to understanding how
they work. Electron cryo-microscopy (cryo-EM) allows visualization of individual complexes in a thin layer of vitreous water. 3D
classification of the resulting particles into subsets of structurally
homogeneous complexes can then be used to obtain atomic structures of these complexes in their different states [1]. However,
subdivision of the cryo-EM data into a discrete number of subsets
is not well suited for the description of continuous types of molecular motion, for which an infinite number of subsets would, in
principle, be needed.
We recently proposed multi-body refinement as an alternative
to discrete classification to describe continuous molecular motions
in cryo-EM projection images [2]. The underlying model of
Tamir Gonen and Brent L. Nannenga (eds.), CryoEM: Methods and Protocols, Methods in Molecular Biology, vol. 2215,
https://doi.org/10.1007/978-1-0716-0966-8_7, © Springer Science+Business Media, LLC, part of Springer Nature 2021
145