EMDep: a web-based system for the deposition and validation of high-resolution electron microscopy macromolecular structural information

Deep learning in electron microscopy

Deep learning is transforming most areas of science and technology, including electron microscopy. This review paper offers a practical perspective aimed at developers with limited familiarity. For context, we review popular applications of deep learning in electron microscopy. Following, we discuss hardware and software needed to get started with deep learning and interface with electron microscopes. We then review neural network components, popular architectures, and their optimization. Finally, we discuss future directions of deep learning in electron microscopy.

Evolution of standardization and dissemination of cryo-EM structures and data jointly by the community, PDB, and EMDB

Cryogenic electron microscopy (cryo-EM) methods began to be used in the mid-1970s to study thin and periodic arrays of proteins. Following a half-century of development in cryo-specimen preparation, instrumentation, data collection, data processing, and modeling software, cryo-EM has become a routine method for solving structures from large biological assemblies to small biomolecules at near to true atomic resolution. This review explores the critical roles played by the Protein Data Bank (PDB) and Electron Microscopy Data Bank (EMDB) in partnership with the community to develop the necessary infrastructure to archive cryo-EM maps and associated models. Public access to cryo-EM structure data has in turn facilitated better understanding of structure–function relationships and advancement of image processing and modeling tool development. The partnership between the global cryo-EM community and PDB and EMDB leadership has synergistically shaped the standards for metadata, one-stop deposition of maps and models, and validation metrics to assess the quality of cryo-EM structures. The advent of cryo-electron tomography (cryo-ET) for in situ molecular cell structures at a broad resolution range and their correlations with other imaging data introduce new data archival challenges in terms of data size and complexity in the years to come.

The Protein Data Bank Archive

Protein Data Bank is the single worldwide archive of experimentally determined macromolecular structure data. Established in 1971 as the first open access data resource in biology, the PDB archive is managed by the worldwide Protein Data Bank (wwPDB) consortium which has four partners-the RCSB Protein Data Bank (RCSB PDB; rcsb.org), the Protein Data Bank Japan (PDBj; pdbj.org), the Protein Data Bank in Europe (PDBe; pdbe.org), and BioMagResBank (BMRB; www.bmrb.wisc.edu ). The PDB archive currently includes ~175,000 entries. The wwPDB has established a number of task forces and working groups that bring together experts form the community who provide recommendations on improving data standards and data validation for improving data quality and integrity. The wwPDB members continue to develop the joint deposition, biocuration, and validation system (OneDep) to improve data quality and accommodate new data from emerging techniques such as 3DEM. Each PDB entry contains coordinate model and associated metadata for all experimentally determined atomic structures, experimental data for the traditional structure determination techniques (X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy), validation reports, and additional information on quaternary structures. The wwPDB partners are committed to following the FAIR (Findability, Accessibility, Interoperability, and Reproducibility) principles and have implemented a DOI resolution mechanism that provides access to all the relevant files for a given PDB entry. On average, >250 new entries are added to the archive every week and made available by each wwPDB partner via FTP area. The wwPDB partner sites also develop data access and analysis tools and make these available via their websites. wwPDB continues to work with experts in the community to establish a federation of archives for archiving structures determined using integrative/hybrid method where multiple experimental techniques are used.

The data universe of structural biology

The Protein Data Bank (PDB) has grown from a small data resource for crystallographers to a worldwide resource serving structural biology. The history of the growth of the PDB and the role that the community has played in developing standards and policies are described. This article also illustrates how other biophysics communities are collaborating with the worldwide PDB to create a network of interoperating data resources. This network will expand the capabilities of structural biology and enable the determination and archiving of increasingly complex structures.

Evolving data standards for cryo-EM structures

Electron cryo-microscopy (cryo-EM) is increasingly being used to determine 3D structures of a broad spectrum of biological specimens from molecules to cells. Anticipating this progress in the early 2000s, an international collaboration of scientists with expertise in both cryo-EM and structure data archiving was established (EMDataResource, previously known as EMDataBank). The major goals of the collaboration have been twofold: to develop the necessary infrastructure for archiving cryo-EM-derived density maps and models, and to promote development of cryo-EM structure validation standards. We describe how cryo-EM data archiving and validation have been developed and jointly coordinated for the Electron Microscopy Data Bank and Protein Data Bank archives over the past two decades, as well as the impact of evolving technology on data standards. Just as for X-ray crystallography and nuclear magnetic resonance, engaging the scientific community via workshops and challenging activities has played a central role in developing recommendations and requirements for the cryo-EM structure data archives.

ETDB-Caltech: A blockchain-based distributed public database for electron tomography

Three-dimensional electron microscopy techniques like electron tomography provide valuable insights into cellular structures, and present significant challenges for data storage and dissemination. Here we explored a novel method to publicly release more than 11,000 such datasets, more than 30 TB in total, collected by our group. Our method, based on a peer-to-peer file sharing network built around a blockchain ledger, offers a distributed solution to data storage. In addition, we offer a user-friendly browser-based interface, https://etdb.caltech.edu, for anyone interested to explore and download our data. We discuss the relative advantages and disadvantages of this system and provide tools for other groups to mine our data and/or use the same approach to share their own imaging datasets.

Real-space refinement in PHENIX for cryo-EM and crystallography

This article describes the implementation of real-space refinement in the phenix.real_space_refine program from the PHENIX suite. The use of a simplified refinement target function enables very fast calculation, which in turn makes it possible to identify optimal data-restraint weights as part of routine refinements with little runtime cost. Refinement of atomic models against low-resolution data benefits from the inclusion of as much additional information as is available. In addition to standard restraints on covalent geometry, phenix.real_space_refine makes use of extra information such as secondary-structure and rotamer-specific restraints, as well as restraints or constraints on internal molecular symmetry. The re-refinement of 385 cryo-EM-derived models available in the Protein Data Bank at resolutions of 6 Å or better shows significant improvement of the models and of the fit of these models to the target maps.

Anticipating innovations in structural biology

In this review, we describe how the interplay among science, technology and community interests contributed to the evolution of four structural biology data resources. We present the method by which data deposited by scientists are prepared for worldwide distribution, and argue that data archiving in a trusted repository must be an integral part of any scientific investigation.

EMDataBank unified data resource for 3DEM

Three-dimensional Electron Microscopy (3DEM) has become a key experimental method in structural biology for a broad spectrum of biological specimens from molecules to cells. The EMDataBank project provides a unified portal for deposition, retrieval and analysis of 3DEM density maps, atomic models and associated metadata (emdatabank.org). We provide here an overview of the rapidly growing 3DEM structural data archives, which include maps in EM Data Bank and map-derived models in the Protein Data Bank. In addition, we describe progress and approaches toward development of validation protocols and methods, working with the scientific community, in order to create a validation pipeline for 3DEM data.

The Caltech Tomography Database and Automatic Processing Pipeline

Here we describe the Caltech Tomography Database and automatic image processing pipeline, designed to process, store, display, and distribute electron tomographic data including tilt-series, sample information, data collection parameters, 3D reconstructions, correlated light microscope images, snapshots, segmentations, movies, and other associated files. Tilt-series are typically uploaded automatically during collection to a user's "Inbox" and processed automatically, but can also be entered and processed in batches via scripts or file-by-file through an internet interface. As with the video website YouTube, each tilt-series is represented on the browsing page with a link to the full record, a thumbnail image and a video icon that delivers a movie of the tomogram in a pop-out window. Annotation tools allow users to add notes and snapshots. The database is fully searchable, and sets of tilt-series can be selected and re-processed, edited, or downloaded to a personal workstation. The results of further processing and snapshots of key results can be recorded in the database, automatically linked to the appropriate tilt-series. While the database is password-protected for local browsing and searching, datasets can be made public and individual files can be shared with collaborators over the Internet. Together these tools facilitate high-throughput tomography work by both individuals and groups.

ATTRACT-EM: a new method for the computational assembly of large molecular machines using cryo-EM maps

Many of the most important functions in the cell are carried out by proteins organized in large molecular machines. Cryo-electron microscopy (cryo-EM) is increasingly being used to obtain low resolution density maps of these large assemblies. A new method, ATTRACT-EM, for the computational assembly of molecular assemblies from their components has been developed. Based on concepts from the protein-protein docking field, it utilizes cryo-EM density maps to assemble molecular subunits at near atomic detail, starting from millions of initial subunit configurations. The search efficiency was further enhanced by recombining partial solutions, the inclusion of symmetry information, and refinement using a molecular force field. The approach was tested on the GroES-GroEL system, using an experimental cryo-EM map at 23.5 Å resolution, and on several smaller complexes. Inclusion of experimental information on the symmetry of the systems and the application of a new gradient vector matching algorithm allowed the efficient identification of docked assemblies in close agreement with experiment. Application to the GroES-GroEL complex resulted in a top ranked model with a deviation of 4.6 Å (and a 2.8 Å model within the top 10) from the GroES-GroEL crystal structure, a significant improvement over existing methods.

Data management challenges in three-dimensional EM

This report describes the outcomes of the Data Management Challenges in 3D Electron Microscopy workshop. Key topics discussed include data models, validation and raw-data archiving. The meeting participants agreed that the EMDataBank should take the lead in addressing these issues, and concrete action points were agreed upon that will have a substantial impact on the accessibility of three-dimensional EM data in biology and medicine.

Fitting multimeric protein complexes into electron microscopy maps using 3D Zernike descriptors

A novel computational method for fitting high-resolution structures of multiple proteins into a cryoelectron microscopy map is presented. The method named EMLZerD generates a pool of candidate multiple protein docking conformations of component proteins, which are later compared with a provided electron microscopy (EM) density map to select the ones that fit well into the EM map. The comparison of docking conformations and the EM map is performed using the 3D Zernike descriptor (3DZD), a mathematical series expansion of three-dimensional functions. The 3DZD provides a unified representation of the surface shape of multimeric protein complex models and EM maps, which allows a convenient, fast quantitative comparison of the three-dimensional structural data. Out of 19 multimeric complexes tested, near native complex structures with a root-mean-square deviation of less than 2.5 Å were obtained for 14 cases while medium range resolution structures with correct topology were computed for the additional 5 cases.

A unique spatial arrangement of the snRNPs within the native spliceosome emerges from in silico studies

The spliceosome is a mega-Dalton ribonucleoprotein (RNP) assembly that processes primary RNA transcripts, producing functional mRNA. The electron microscopy structures of the native spliceosome and of several spliceosomal subcomplexes are available; however, the spatial arrangement of the latter within the native spliceosome is not known. We designed a computational procedure to efficiently fit thousands of conformers into the spliceosome envelope. Despite the low resolution limitations, we obtained only one model that complies with the available biochemical data. Our model localizes the five small nuclear RNPs (snRNPs) mostly within the large subunit of the native spliceosome, requiring only minor conformation changes. The remaining free volume presumably accommodates additional spliceosomal components. The constituents of the active core of the spliceosome are juxtaposed, forming a continuous surface deep within the large spliceosomal cavity, which provides a sheltered environment for the splicing reaction.

Macromolecular assembly structures by comparative modeling and electron microscopy

Advances in electron microscopy allow for structure determination of large biological machines at increasingly higher resolutions. A key step in this process is fitting component structures into the electron microscopy-derived density map of their assembly. Comparative modeling can contribute by providing atomic models of the components, via fold assignment, sequence-structure alignment, model building, and model assessment. All four stages of comparative modeling can also benefit from consideration of the density map. In this chapter, we describe numerous types of modeling problems restrained by a density map and available protocols for finding solutions. In particular, we provide detailed instructions for density map-guided modeling using the Integrative Modeling Platform (IMP), MODELLER, and UCSF Chimera.

Consensus among flexible fitting approaches improves the interpretation of cryo-EM data

Cryo-elecron microscopy (cryo-EM) can provide important structural information of large macromolecular assemblies in different conformational states. Recent years have seen an increase in structures deposited in the Protein Data Bank (PDB) by fitting a high-resolution structure into its low-resolution cryo-EM map. A commonly used protocol for accommodating the conformational changes between the X-ray structure and the cryo-EM map is rigid body fitting of individual domains. With the emergence of different flexible fitting approaches, there is a need to compare and revise these different protocols for the fitting. We have applied three diverse automated flexible fitting approaches on a protein dataset for which rigid domain fitting (RDF) models have been deposited in the PDB. In general, a consensus is observed in the conformations, which indicates a convergence from these theoretically different approaches to the most probable solution corresponding to the cryo-EM map. However, the result shows that the convergence might not be observed for proteins with complex conformational changes or with missing densities in cryo-EM map. In contrast, RDF structures deposited in the PDB can represent conformations that not only differ from the consensus obtained by flexible fitting but also from X-ray crystallography. Thus, this study emphasizes that a "consensus" achieved by the use of several automated flexible fitting approaches can provide a higher level of confidence in the modeled configurations. Following this protocol not only increases the confidence level of fitting, but also highlights protein regions with uncertain fitting. Hence, this protocol can lead to better interpretation of cryo-EM data.

ePMV embeds molecular modeling into professional animation software environments

Increasingly complex research has made it more difficult to prepare data for publication, education, and outreach. Many scientists must also wade through black-box code to interface computational algorithms from diverse sources to supplement their bench work. To reduce these barriers we have developed an open-source plug-in, embedded Python Molecular Viewer (ePMV), that runs molecular modeling software directly inside of professional 3D animation applications (hosts) to provide simultaneous access to the capabilities of these newly connected systems. Uniting host and scientific algorithms into a single interface allows users from varied backgrounds to assemble professional quality visuals and to perform computational experiments with relative ease. By enabling easy exchange of algorithms, ePMV can facilitate interdisciplinary research, smooth communication between broadly diverse specialties, and provide a common platform to frame and visualize the increasingly detailed intersection(s) of cellular and molecular biology.

The Protein Data Bank in Europe (PDBe): bringing structure to biology

The Protein Data Bank in Europe (PDBe) is the European partner in the Worldwide PDB and as such handles depositions of X-ray, NMR and EM data and structure models. PDBe also provides advanced bioinformatics services based on data from the PDB and related resources. Some of the challenges facing the PDB and its guardians are discussed, as well as some of the areas on which PDBe activities will focus in the future (advanced services, ligands, integration, validation and experimental data). Finally, some recent developments at PDBe are described.

Exploring the spatial and temporal organization of a cell's proteome

To increase our current understanding of cellular processes, such as cell signaling and division, knowledge is needed about the spatial and temporal organization of the proteome at different organizational levels. These levels cover a wide range of length and time scales: from the atomic structures of macromolecules for inferring their molecular function, to the quantitative description of their abundance, and spatial distribution in the cell. Emerging new experimental technologies are greatly increasing the availability of such spatial information on the molecular organization in living cells. This review addresses three fields that have significantly contributed to our understanding of the proteome's spatial and temporal organization: first, methods for the structure determination of individual macromolecular assemblies, specifically the fitting of atomic structures into density maps generated from electron microscopy techniques; second, research that visualizes the spatial distributions of these complexes within the cellular context using cryo electron tomography techniques combined with computational image processing; and third, methods for the spatial modeling of the dynamic organization of the proteome, specifically those methods for simulating reaction and diffusion of proteins and complexes in crowded intracellular fluids. The long-term goal is to integrate the varied data about a proteome's organization into a spatially explicit, predictive model of cellular processes.

Determining macromolecular assembly structures by molecular docking and fitting into an electron density map

Structural models of macromolecular assemblies are instrumental for gaining a mechanistic understanding of cellular processes. Determining these structures is a major challenge for experimental techniques, such as X-ray crystallography, NMR spectroscopy and electron microscopy (EM). Thus, computational modeling techniques, including molecular docking, are required. The development of most molecular docking methods has so far been focused on modeling of binary complexes. We have recently introduced the MultiFit method for modeling the structure of a multisubunit complex by simultaneously optimizing the fit of the model into an EM density map of the entire complex and the shape complementarity between interacting subunits. Here, we report algorithmic advances of the MultiFit method that result in an efficient and accurate assembly of the input subunits into their density map. The successful predictions and the increasing number of complexes being characterized by EM suggests that the CAPRI challenge could be extended to include docking-based modeling of macromolecular assemblies guided by EM.

Unified data resource for cryo-EM

Three-dimensional (3D) cryoelectron microscopy reconstruction methods are uniquely able to reveal structures of many important macromolecules and macromolecular complexes. EMDataBank.org, a joint effort of the Protein Databank in Europe (PDBe), the Research Collaboratory for Structural Bioinformatics (RCSB), and the National Center for Macromolecular Imaging (NCMI), is a "one-stop shop" resource for global deposition and retrieval of cryo-EM map, model, and associated metadata. The resource unifies public access to the two major EM Structural Data archives: EM Data Bank (EMDB) and Protein Data Bank (PDB), and facilitates use of EM structural data of macromolecules and macromolecular complexes by the wider scientific community.

Conformational dynamics data bank: a database for conformational dynamics of proteins and supramolecular protein assemblies

The conformational dynamics data bank (CDDB, http://www.cdyn.org) is a database that aims to provide comprehensive results on the conformational dynamics of high molecular weight proteins and protein assemblies. Analysis is performed using a recently introduced coarse-grained computational approach that is applied to the majority of structures present in the electron microscopy data bank (EMDB). Results include equilibrium thermal fluctuations and elastic strain energy distributions that identify rigid versus flexible protein domains generally, as well as those associated with specific functional transitions, and correlations in molecular motions that identify molecular regions that are highly coupled dynamically, with implications for allosteric mechanisms. A practical web-based search interface enables users to easily collect conformational dynamics data in various formats. The data bank is maintained and updated automatically to include conformational dynamics results for new structural entries as they become available in the EMDB. The CDDB complements static structural information to facilitate the investigation and interpretation of the biological function of proteins and protein assemblies essential to cell function.

PDBe: Protein Data Bank in Europe

The Protein Data Bank in Europe (PDBe; pdbe.org) is actively involved in managing the international archive of biomacromolecular structure data as one of the partners in the Worldwide Protein Data Bank (wwPDB; wwpdb.org). PDBe also develops new tools to make structural data more widely and more easily available to the biomedical community. PDBe has developed a browser to access and analyze the structural archive using classification systems that are familiar to chemists and biologists. The PDBe web pages that describe individual PDB entries have been enhanced through the introduction of plain-English summary pages and iconic representations of the contents of an entry (PDBprints). In addition, the information available for structures determined by means of NMR spectroscopy has been expanded. Finally, the entire web site has been redesigned to make it substantially easier to use for expert and novice users alike. PDBe works closely with other teams at the European Bioinformatics Institute (EBI) and in the international scientific community to develop new resources with value-added information. The SIFTS initiative is an example of such a collaboration—it provides extensive mapping data between proteins whose structures are available from the PDB and a host of other biomedical databases. SIFTS is widely used by major bioinformatics resources.

EMDataBank.org: unified data resource for CryoEM

Cryo-electron microscopy reconstruction methods are uniquely able to reveal structures of many important macromolecules and macromolecular complexes. EMDataBank.org, a joint effort of the Protein Data Bank in Europe (PDBe), the Research Collaboratory for Structural Bioinformatics (RCSB) and the National Center for Macromolecular Imaging (NCMI), is a global ‘one-stop shop’ resource for deposition and retrieval of cryoEM maps, models and associated metadata. The resource unifies public access to the two major archives containing EM-based structural data: EM Data Bank (EMDB) and Protein Data Bank (PDB), and facilitates use of EM structural data of macromolecules and macromolecular complexes by the wider scientific community.

Integrative structure modeling of macromolecular assemblies from proteomics data

Proteomics techniques have been used to generate comprehensive lists of protein interactions in a number of species. However, relatively little is known about how these interactions result in functional multiprotein complexes. This gap can be bridged by combining data from proteomics experiments with data from established structure determination techniques. Correspondingly, integrative computational methods are being developed to provide descriptions of protein complexes at varying levels of accuracy and resolution, ranging from complex compositions to detailed atomic structures.

Periodic table of virus capsids: implications for natural selection and design

BACKGROUND

For survival, most natural viruses depend upon the existence of spherical capsids: protective shells of various sizes composed of protein subunits. So far, general evolutionary pressures shaping capsid design have remained elusive, even though an understanding of such properties may help in rationally impeding the virus life cycle and designing efficient nano-assemblies.

PRINCIPAL FINDINGS

This report uncovers an unprecedented and species-independent evolutionary pressure on virus capsids, based on the the notion that the simplest capsid designs (or those capsids with the lowest "hexamer complexity", C(h)) are the fittest, which was shown to be true for all available virus capsids. The theories result in a physically meaningful periodic table of virus capsids that uncovers strong and overarching evolutionary pressures, while also offering geometric explanations to other capsid properties (rigidity, pleomorphy, auxiliary requirements, etc.) that were previously considered to be unrelatable properties of the individual virus.

SIGNIFICANCE

Apart from describing a universal rule for virus capsid evolution, our work (especially the periodic table) provides a language with which highly diverse virus capsids, unified only by geometry, may be described and related to each other. Finally, the available virus structure databases and other published data reiterate the predicted geometry-derived rules, reinforcing the role of geometry in the natural selection and design of virus capsids.

PDBe: Protein Data Bank in Europe

The Protein Data Bank in Europe (PDBe) (http://www.ebi.ac.uk/pdbe/) is actively working with its Worldwide Protein Data Bank partners to enhance the quality and consistency of the international archive of bio-macromolecular structure data, the Protein Data Bank (PDB). PDBe also works closely with its collaborators at the European Bioinformatics Institute and the scientific community around the world to enhance its databases and services by adding curated and actively maintained derived data to the existing structural data in the PDB. We have developed a new database infrastructure based on the remediated PDB archive data and a specially designed database for storing information on interactions between proteins and bound molecules. The group has developed new services that allow users to carry out simple textual queries or more complex 3D structure-based queries. The newly designed 'PDBeView Atlas pages' provide an overview of an individual PDB entry in a user-friendly layout and serve as a starting point to further explore the information available in the PDBe database. PDBe's active involvement with the X-ray crystallography, Nuclear Magnetic Resonance spectroscopy and cryo-Electron Microscopy communities have resulted in improved tools for structure deposition and analysis.

Inferential optimization for simultaneous fitting of multiple components into a CryoEM map of their assembly

Models of macromolecular assemblies are essential for a mechanistic description of cellular processes. Such models are increasingly obtained by fitting atomic-resolution structures of components into a density map of the whole assembly. Yet, current density-fitting techniques are frequently insufficient for an unambiguous determination of the positions and orientations of all components. Here, we describe MultiFit, a method used for simultaneously fitting atomic structures of components into their assembly density map at resolutions as low as 25 A. The component positions and orientations are optimized with respect to a scoring function that includes the quality-of-fit of components in the map, the protrusion of components from the map envelope, and the shape complementarity between pairs of components. The scoring function is optimized by our exact inference optimizer DOMINO (Discrete Optimization of Multiple INteracting Objects) that efficiently finds the global minimum in a discrete sampling space. MultiFit was benchmarked on seven assemblies of known structure, consisting of up to seven proteins each. The input atomic structures of the components were obtained from the Protein Data Bank, as well as by comparative modeling based on a 16-99% sequence identity to a template structure. A near-native configuration was usually found as the top-scoring model. Therefore, MultiFit can provide initial configurations for further refinement of many multicomponent assembly structures described by electron microscopy.

Data deposition and annotation at the worldwide protein data bank

The Protein Data Bank (PDB) is the repository for three-dimensional structures of biological macromolecules, determined by experimental methods. The data in the archive is free and easily available via the Internet from any of the worldwide centers managing this global archive. These data are used by scientists, researchers, bioinformatics specialists, educators, students, and general audiences to understand biological phenomenon at a molecular level. Analysis of this structural data also inspires and facilitates new discoveries in science. This chapter describes the tools and methods currently used for deposition, processing, and release of data in the PDB. References to future enhancements are also included.

Data deposition and annotation at the worldwide protein data bank

The Protein Data Bank (PDB) is the repository for the three-dimensional structures of biological macromolecules, determined by experimental methods. The data in the archive are free and easily available via the Internet from any of the worldwide centers managing this global archive. These data are used by scientists, researchers, bioinformatics specialists, educators, students, and lay audiences to understand biological phenomena at a molecular level. Analysis of these structural data also inspires and facilitates new discoveries in science. This chapter describes the tools and methods currently used for deposition, processing, and release of data in the PDB. References to future enhancements are also included.

Multiple subunit fitting into a low-resolution density map of a macromolecular complex using a gaussian mixture model

Recently, electron microscopy measurement of single particles has enabled us to reconstruct a low-resolution 3D density map of large biomolecular complexes. If structures of the complex subunits can be solved by x-ray crystallography at atomic resolution, fitting these models into the 3D density map can generate an atomic resolution model of the entire large complex. The fitting of multiple subunits, however, generally requires large computational costs; therefore, development of an efficient algorithm is required. We developed a fast fitting program, “gmfit”, which employs a Gaussian mixture model (GMM) to represent approximated shapes of the 3D density map and the atomic models. A GMM is a distribution function composed by adding together several 3D Gaussian density functions. Because our model analytically provides an integral of a product of two distribution functions, it enables us to quickly calculate the fitness of the density map and the atomic models. Using the integral, two types of potential energy function are introduced: the attraction potential energy between a 3D density map and each subunit, and the repulsion potential energy between subunits. The restraint energy for symmetry is also employed to build symmetrical origomeric complexes. To find the optimal configuration of subunits, we randomly generated initial configurations of subunit models, and performed a steepest-descent method using forces and torques of the three potential energies. Comparison between an original density map and its GMM showed that the required number of Gaussian distribution functions for a given accuracy depended on both resolution and molecular size. We then performed test fitting calculations for simulated low-resolution density maps of atomic models of homodimer, trimer, and hexamer, using different search parameters. The results indicated that our method was able to rebuild atomic models of a complex even for maps of 30 Å resolution if sufficient numbers (eight or more) of Gaussian distribution functions were employed for each subunit, and the symmetric restraints were assigned for complexes with more than three subunits. As a more realistic test, we tried to build an atomic model of the GroEL/ES complex by fitting 21-subunit atomic models into the 3D density map obtained by cryoelectron microscopy using the C7 symmetric restraints. A model with low root mean-square deviations (14.7 Å) was obtained as the lowest-energy model, showing that our fitting method was reasonably accurate. Inclusion of other restraints from biological and biochemical experiments could further enhance the accuracy.

Software tools for molecular microscopy

In our role as the editors of a special edition of the Journal of Structural Biology published in 1996 and devoted to the development of software tools, we offer our view of past developments and future prospects in this area. The astonishing progress in computer hardware over the past decade has fueled a significant increase in computational power available for the solution of macromolecular structures. At the same time the relatively slow growth and development of the accompanying software reflects the difficulties of developing large, complex and very specialized analytical methods.

The Protein Data Bank: a historical perspective

The Protein Data Bank began as a grassroots effort in 1971. It has grown from a small archive containing a dozen structures to a major international resource for structural biology containing more than 40000 entries. The interplay of science, technology and attitudes about data sharing have all played a role in the growth of this resource.

E-MSD: improving data deposition and structure quality

The Macromolecular Structure Database (MSD) (http://www.ebi.ac.uk/msd/) [H. Boutselakis, D. Dimitropoulos, J. Fillon, A. Golovin, K. Henrick, A. Hussain, J. Ionides, M. John, P. A. Keller, E. Krissinel et al. (2003) E-MSD: the European Bioinformatics Institute Macromolecular Structure Database. Nucleic Acids Res., 31, 458-462.] group is one of the three partners in the worldwide Protein DataBank (wwPDB), the consortium entrusted with the collation, maintenance and distribution of the global repository of macromolecular structure data [H. Berman, K. Henrick and H. Nakamura (2003) Announcing the worldwide Protein Data Bank. Nature Struct. Biol., 10, 980.]. Since its inception, the MSD group has worked with partners around the world to improve the quality of PDB data, through a clean up programme that addresses inconsistencies and inaccuracies in the legacy archive. The improvements in data quality in the legacy archive have been achieved largely through the creation of a unified data archive, in the form of a relational database that stores all of the data in the wwPDB. The three partners are working towards improving the tools and methods for the deposition of new data by the community at large. The implementation of the MSD database, together with the parallel development of improved tools and methodologies for data harvesting, validation and archival, has lead to significant improvements in the quality of data that enters the archive. Through this and related projects in the NMR and EM realms the MSD continues to improve the quality of publicly available structural data.

Automation of macromolecular crystallography beamlines

The production of three-dimensional crystallographic structural information of macromolecules can now be thought of as a pipeline which is being streamlined at every stage from protein cloning, expression and purification, through crystallisation to data collection and structure solution. Synchrotron X-ray beamlines are a key section of this pipeline as it is at these that the X-ray diffraction data that ultimately leads to the elucidation of macromolecular structures are collected. The burgeoning number of macromolecular crystallography (MX) beamlines available worldwide may be enhanced significantly with the automation of both their operation and of the experiments carried out on them. This paper reviews the current situation and provides a glimpse of how a MX beamline may look in the not too distant future.