Unsupervised Cryo-EM Images Denoising and Clustering Based on Deep Convolutional Autoencoder and K-Means+

Cryo-electron microscopy (cryo-EM) is a widely used structural determination technique. Because of the extremely low signal-to-noise ratio (SNR) of images captured by cryo-EM, clustering single-particle cryo-EM images with high accuracy is challenging. To address this, we proposed an iterative denoising and clustering method based on a deep convolutional variational autoencoder and K-means++. The proposed method contains two modules: a denoising ResNet variational autoencoder (DRVAE), and Balance size K-means++ (BSK-means++). First, the DRVAE is trained in a fully unsupervised manner to initialize the neural network and obtain preliminary denoised images. Second, BSK-means++ is built for clustering denoised images, and images closer to class centers are divided into reliable samples. Third, the training of DRVAE is continued, while the class-average images are used as pseudo supervision of reliable samples to reserve more detailed information of denoised images. Finally, the second and third steps mentioned above can be performed jointly and iteratively until convergence occurs. The experimental results showed that the proposed method can generate reliable class average images and achieve better clustering accuracy and normalized mutual information than current methods. This study confirmed that DRVAE with BSK-means++ could achieve a good denoise performance on single-particle cryo-EM images, which can help researchers obtain information such as symmetry and heterogeneity of the target particles. In addition, the proposed method avoids the extreme imbalance of class size, which improves the reliability of the clustering result.

Machine Learning for Structure Determination in Single-Particle Cryo-Electron Microscopy: A Systematic Review

Recently, single-particle cryo-electron microscopy (cryo-EM) has become an indispensable method for determining macromolecular structures at high resolution to deeply explore the relevant molecular mechanism. Its recent breakthrough is mainly because of the rapid advances in hardware and image processing algorithms, especially machine learning. As an essential support of single-particle cryo-EM, machine learning has powered many aspects of structure determination and greatly promoted its development. In this article, we provide a systematic review of the applications of machine learning in this field. Our review begins with a brief introduction of single-particle cryo-EM, followed by the specific tasks and challenges of its image processing. Then, focusing on the workflow of structure determination, we describe relevant machine learning algorithms and applications at different steps, including particle picking, 2-D clustering, 3-D reconstruction, and other steps. As different tasks exhibit distinct characteristics, we introduce the evaluation metrics for each task and summarize their dynamics of technology development. Finally, we discuss the open issues and potential trends in this promising field.

Bayesian Random Tomography of Particle Systems

Random tomography is a common problem in imaging science and refers to the task of reconstructing a three-dimensional volume from two-dimensional projection images acquired in unknown random directions. We present a Bayesian approach to random tomography. At the center of our approach is a meshless representation of the unknown volume as a mixture of spherical Gaussians. Each Gaussian can be interpreted as a particle such that the unknown volume is represented by a particle cloud. The particle representation allows us to speed up the computation of projection images and to represent a large variety of structures accurately and efficiently. We develop Markov chain Monte Carlo algorithms to infer the particle positions as well as the unknown orientations. Posterior sampling is challenging due to the high dimensionality and multimodality of the posterior distribution. We tackle these challenges by using Hamiltonian Monte Carlo and a global rotational sampling strategy. We test the approach on various simulated and real datasets.

Joint Angular Refinement and Reconstruction for Single-Particle Cryo-EM

Single-particle cryo-electron microscopy (cryo-EM) reconstructs the three-dimensional (3D) structure of biomolecules from a large set of 2D projection images with random and unknown orientations. A crucial step in the single-particle cryo-EM pipeline is 3D refinement, which resolves a highresolution 3D structure from an initial approximate volume by refining the estimation of the orientation of each projection. In this work, we propose a new approach that refines the projection angles on the continuum. We formulate the optimization problem over the density map and the orientations jointly. The density map is updated using the efficient alternating-direction method of multipliers, while the orientations are updated through a semicoordinate- wise gradient descent for which we provide an explicit derivation of the gradient. Our method eliminates the requirement for a fine discretization of the orientation space and does away with the classical but computationally expensive templatematching step. Numerical results demonstrate the feasibility and performance of our approach compared to several baselines.

Advances in image processing for single-particle analysis by electron cryomicroscopy and challenges ahead

Electron cryomicroscopy (cryoEM) is essential for the study and functional understanding of non-crystalline macromolecules such as proteins. These molecules cannot be imaged using X-ray crystallography or other popular methods. CryoEM has been successfully used to visualize macromolecular complexes such as ribosomes, viruses, and ion channels. Determination of structural models of these at various conformational states leads to insight on how these molecules function. Recent advances in imaging technology have given cryoEM a scientific rebirth. As a result of these technological advances image processing and analysis have yielded molecular structures at atomic resolution. Nevertheless there continue to be challenges in image processing, and in this article we will touch on the most essential in order to derive an accurate three-dimensional model from noisy projection images. Traditional approaches, such as k-means clustering for class averaging, will be provided as background. We will then highlight new approaches for each image processing subproblem, including a 3D reconstruction method for asymmetric molecules using just two projection images and deep learning algorithms for automated particle picking.

Building Proteins in a Day: Efficient 3D Molecular Structure Estimation with Electron Cryomicroscopy

Discovering the 3D atomic-resolution structure of molecules such as proteins and viruses is one of the foremost research problems in biology and medicine. Electron Cryomicroscopy (cryo-EM) is a promising vision-based technique for structure estimation which attempts to reconstruct 3D atomic structures from a large set of 2D transmission electron microscope images. This paper presents a new Bayesian framework for cryo-EM structure estimation that builds on modern stochastic optimization techniques to allow one to scale to very large datasets. We also introduce a novel Monte-Carlo technique that reduces the cost of evaluating the objective function during optimization by over five orders of magnitude. The net result is an approach capable of estimating 3D molecular structure from large-scale datasets in about a day on a single CPU workstation.

cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination

Single-particle electron cryomicroscopy (cryo-EM) is a powerful method for determining the structures of biological macromolecules. With automated microscopes, cryo-EM data can often be obtained in a few days. However, processing cryo-EM image data to reveal heterogeneity in the protein structure and to refine 3D maps to high resolution frequently becomes a severe bottleneck, requiring expert intervention, prior structural knowledge, and weeks of calculations on expensive computer clusters. Here we show that stochastic gradient descent (SGD) and branch-and-bound maximum likelihood optimization algorithms permit the major steps in cryo-EM structure determination to be performed in hours or minutes on an inexpensive desktop computer. Furthermore, SGD with Bayesian marginalization allows ab initio 3D classification, enabling automated analysis and discovery of unexpected structures without bias from a reference map. These algorithms are combined in a user-friendly computer program named cryoSPARC (http://www.cryosparc.com).

Unsupervised Cryo-EM Data Clustering through Adaptively Constrained K-Means Algorithm

In single-particle cryo-electron microscopy (cryo-EM), K-means clustering algorithm is widely used in unsupervised 2D classification of projection images of biological macromolecules. 3D ab initio reconstruction requires accurate unsupervised classification in order to separate molecular projections of distinct orientations. Due to background noise in single-particle images and uncertainty of molecular orientations, traditional K-means clustering algorithm may classify images into wrong classes and produce classes with a large variation in membership. Overcoming these limitations requires further development on clustering algorithms for cryo-EM data analysis. We propose a novel unsupervised data clustering method building upon the traditional K-means algorithm. By introducing an adaptive constraint term in the objective function, our algorithm not only avoids a large variation in class sizes but also produces more accurate data clustering. Applications of this approach to both simulated and experimental cryo-EM data demonstrate that our algorithm is a significantly improved alterative to the traditional K-means algorithm in single-particle cryo-EM analysis.

A Stochastic Hill Climbing Approach for Simultaneous 2D Alignment and Clustering of Cryogenic Electron Microscopy Images

A critical step in the analysis of novel cryogenic electron microscopy (cryo-EM) single-particle datasets is the identification of homogeneous subsets of images. Methods for solving this problem are important for data quality assessment, ab initio 3D reconstruction, and analysis of population diversity due to the heterogeneous nature of macromolecules. Here we formulate a stochastic algorithm for identification of homogeneous subsets of images. The purpose of the method is to generate improved 2D class averages that can be used to produce a reliable 3D starting model in a rapid and unbiased fashion. We show that our method overcomes inherent limitations of widely used clustering approaches and proceed to test the approach on six publicly available experimental cryo-EM datasets. We conclude that, in each instance, ab initio 3D reconstructions of quality suitable for initialization of high-resolution refinement are produced from the cluster centers.

Bayesian inference of initial models in cryo-electron microscopy using pseudo-atoms

Single-particle cryo-electron microscopy is widely used to study the structure of macromolecular assemblies. Tens of thousands of noisy two-dimensional images of the macromolecular assembly viewed from different directions are used to infer its three-dimensional structure. The first step is to estimate a low-resolution initial model and initial image orientations. This is a challenging global optimization problem with many unknowns, including an unknown orientation for each two-dimensional image. Obtaining a good initial model is crucial for the success of the subsequent refinement step. We introduce a probabilistic algorithm for estimating an initial model. The algorithm is fast, has very few algorithmic parameters, and yields information about the precision of estimated model parameters in addition to the parameters themselves. Our algorithm uses a pseudo-atomic model to represent the low-resolution three-dimensional structure, with isotropic Gaussian components as moveable pseudo-atoms. This leads to a significant reduction in the number of parameters needed to represent the three-dimensional structure, and a simplified way of computing two-dimensional projections. It also contributes to the speed of the algorithm. We combine the estimation of the unknown three-dimensional structure and image orientations in a Bayesian framework. This ensures that there are very few parameters to set, and specifies how to combine different types of prior information about the structure with the given data in a systematic way. To estimate the model parameters we use Markov chain Monte Carlo sampling. The advantage is that instead of just obtaining point estimates of model parameters, we obtain an ensemble of models revealing the precision of the estimated parameters. We demonstrate the algorithm on both simulated and real data.

Cryogenic electron microscopy and single-particle analysis

About 20 years ago, the first three-dimensional (3D) reconstructions at subnanometer (<10-Å) resolution of an icosahedral virus assembly were obtained by cryogenic electron microscopy (cryo-EM) and single-particle analysis. Since then, thousands of structures have been determined to resolutions ranging from 30 Å to near atomic (<4 Å). Almost overnight, the recent development of direct electron detectors and the attendant improvement in analysis software have advanced the technology considerably. Near-atomic-resolution reconstructions can now be obtained, not only for megadalton macromolecular complexes or highly symmetrical assemblies but also for proteins of only a few hundred kilodaltons. We discuss the developments that led to this breakthrough in high-resolution structure determination by cryo-EM and point to challenges that lie ahead.

Bayesian analysis of individual electron microscopy images: towards structures of dynamic and heterogeneous biomolecular assemblies

We develop a method to extract structural information from electron microscopy (EM) images of dynamic and heterogeneous molecular assemblies. To overcome the challenge of disorder in the imaged structures, we analyze each image individually, avoiding information loss through clustering or averaging. The Bayesian inference of EM (BioEM) method uses a likelihood-based probabilistic measure to quantify the consistency between each EM image and given structural models. The likelihood function accounts for uncertainties in the molecular position and orientation, variations in the relative intensities and noise in the experimental images. The BioEM formalism is physically intuitive and mathematically simple. We show that for experimental GroEL images, BioEM correctly identifies structures according to the functional state. The top-ranked structure is the corresponding X-ray crystal structure, followed by an EM structure generated previously from a superset of the EM images used here. To analyze EM images of highly flexible molecules, we propose an ensemble refinement procedure, and validate it with synthetic EM maps of the ESCRT-I-II supercomplex. Both the size of the ensemble and its structural members are identified correctly. BioEM offers an alternative to 3D-reconstruction methods, extracting accurate population distributions for highly flexible structures and their assemblies. We discuss limitations of the method, and possible applications beyond ensemble refinement, including the cross-validation and unbiased post-assessment of model structures, and the structural characterization of systems where traditional approaches fail. Overall, our results suggest that the BioEM framework can be used to analyze EM images of both ordered and disordered molecular systems.

Assembly of macromolecular complexes by satisfaction of spatial restraints from electron microscopy images

To obtain a structural model of a macromolecular assembly by single-particle EM, a large number of particle images need to be collected, aligned, clustered, averaged, and finally assembled via reconstruction into a 3D density map. This process is limited by the number and quality of the particle images, the accuracy of the initial model, and the compositional and conformational heterogeneity. Here, we describe a structure determination method that avoids the reconstruction procedure. The atomic structures of the individual complex components are assembled by optimizing a match against 2D EM class-average images, an excluded volume criterion, geometric complementarity, and optional restraints from proteomics and chemical cross-linking experiments. The optimization relies on a simulated annealing Monte Carlo search and a divide-and-conquer message-passing algorithm. Using simulated and experimentally determined EM class averages for 12 and 4 protein assemblies, respectively, we show that a few class averages can indeed result in accurate models for complexes of as many as five subunits. Thus, integrative structural biology can now benefit from the relative ease with which the EM class averages are determined.

A Bayesian adaptive basis algorithm for single particle reconstruction

Traditional single particle reconstruction methods use either the Fourier or the delta function basis to represent the particle density map. This paper proposes a more flexible algorithm that adaptively chooses the basis based on the data. Because the basis adapts to the data, the reconstruction resolution and signal-to-noise ratio (SNR) is improved compared to a reconstruction with a fixed basis. Moreover, the algorithm automatically masks the particle, thereby separating it from the background. This eliminates the need for ad hoc filtering or masking in the refinement loop. The algorithm is formulated in a Bayesian maximum-a-posteriori framework and uses an efficient optimization algorithm for the maximization. Evaluations using simulated and actual cryogenic electron microscopy data show resolution and SNR improvements as well as the effective masking of particle from background.

A Bayesian view on cryo-EM structure determination

Three-dimensional (3D) structure determination by single-particle analysis of cryo-electron microscopy (cryo-EM) images requires many parameters to be determined from extremely noisy data. This makes the method prone to overfitting, that is, when structures describe noise rather than signal, in particular near their resolution limit where noise levels are highest. Cryo-EM structures are typically filtered using ad hoc procedures to prevent overfitting, but the tuning of arbitrary parameters may lead to subjectivity in the results. I describe a Bayesian interpretation of cryo-EM structure determination, where smoothness in the reconstructed density is imposed through a Gaussian prior in the Fourier domain. The statistical framework dictates how data and prior knowledge should be combined, so that the optimal 3D linear filter is obtained without the need for arbitrariness and objective resolution estimates may be obtained. Application to experimental data indicates that the statistical approach yields more reliable structures than existing methods and is capable of detecting smaller classes in data sets that contain multiple different structures.