Classification of crystallization outcomes using deep convolutional neural networks

Robust image descriptor for machine learning based data reduction in serial crystallography

Serial crystallography experiments at synchrotron and X-ray free-electron laser (XFEL) sources are producing crystallographic data sets of ever-increasing volume. While these experiments have large data sets and high-frame-rate detectors (around 3520 frames per second), only a small percentage of the data are useful for downstream analysis. Thus, an efficient and real-time data classification pipeline is essential to differentiate reliably between useful and non-useful images, typically known as 'hit' and 'miss', respectively, and keep only hit images on disk for further analysis such as peak finding and indexing. While feature-point extraction is a key component of modern approaches to image classification, existing approaches require computationally expensive patch preprocessing to handle perspective distortion. This paper proposes a pipeline to categorize the data, consisting of a real-time feature extraction algorithm called modified and parallelized FAST (MP-FAST), an image descriptor and a machine learning classifier. For parallelizing the primary operations of the proposed pipeline, central processing units, graphics processing units and field-programmable gate arrays are implemented and their performances compared. Finally, MP-FAST-based image classification is evaluated using a multi-layer perceptron on various data sets, including both synthetic and experimental data. This approach demonstrates superior performance compared with other feature extractors and classifiers.

Deep learning applications in protein crystallography

Deep learning techniques can recognize complex patterns in noisy, multidimensional data. In recent years, researchers have started to explore the potential of deep learning in the field of structural biology, including protein crystallography. This field has some significant challenges, in particular producing high-quality and well ordered protein crystals. Additionally, collecting diffraction data with high completeness and quality, and determining and refining protein structures can be problematic. Protein crystallographic data are often high-dimensional, noisy and incomplete. Deep learning algorithms can extract relevant features from these data and learn to recognize patterns, which can improve the success rate of crystallization and the quality of crystal structures. This paper reviews progress in this field.

20 years of crystal hits: progress and promise in ultrahigh-throughput crystallization screening

Diffraction-based structural methods contribute a large fraction of the biomolecular structural models available, providing a critical understanding of macromolecular architecture. These methods require crystallization of the target molecule, which remains a primary bottleneck in crystal-based structure determination. The National High-Throughput Crystallization Center at Hauptman-Woodward Medical Research Institute has focused on overcoming obstacles to crystallization through a combination of robotics-enabled high-throughput screening and advanced imaging to increase the success of finding crystallization conditions. This paper will describe the lessons learned from over 20 years of operation of our high-throughput crystallization services. The current experimental pipelines, instrumentation, imaging capabilities and software for image viewing and crystal scoring are detailed. New developments in the field and opportunities for further improvements in biomolecular crystallization are reflected on.

Not getting in too deep: A practical deep learning approach to routine crystallisation image classification

Using a relatively small training set of ~16 thousand images from macromolecular crystallisation experiments, we compare classification results obtained with four of the most widely-used convolutional deep-learning network architectures that can be implemented without the need for extensive computational resources. We show that the classifiers have different strengths that can be combined to provide an ensemble classifier achieving a classification accuracy comparable to that obtained by a large consortium initiative. We use eight classes to effectively rank the experimental outcomes, thereby providing detailed information that can be used with routine crystallography experiments to automatically identify crystal formation for drug discovery and pave the way for further exploration of the relationship between crystal formation and crystallisation conditions.

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization

Artificial intelligence and specifically machine learning applications are nowadays used in a variety of scientific applications and cutting-edge technologies, where they have a transformative impact. Such an assembly of statistical and linear algebra methods making use of large data sets is becoming more and more integrated into chemistry and crystallization research workflows. This review aims to present, for the first time, a holistic overview of machine learning and cheminformatics applications as a novel, powerful means to accelerate the discovery of new crystal structures, predict key properties of organic crystalline materials, simulate, understand, and control the dynamics of complex crystallization process systems, as well as contribute to high throughput automation of chemical process development involving crystalline materials. We critically review the advances in these new, rapidly emerging research areas, raising awareness in issues such as the bridging of machine learning models with first-principles mechanistic models, data set size, structure, and quality, as well as the selection of appropriate descriptors. At the same time, we propose future research at the interface of applied mathematics, chemistry, and crystallography. Overall, this review aims to increase the adoption of such methods and tools by chemists and scientists across industry and academia.

Potential of Deep Learning Methods for Deep Level Particle Characterization in Crystallization

Crystalline particle properties, which are defined throughout the crystallization process chain, are strongly tied to the quality of the final product bringing along the need of detailed particle characterization. The most important characteristics are the size, shape and purity, which are influenced by agglomeration. Therefore, a pure size determination is often insufficient and a deep level evaluation regarding agglomerates and primary crystals bound in agglomerates is desirable as basis to increase the quality of crystalline products. We present a promising deep learning approach for particle characterization in crystallization. In an end-to-end fashion, the interactions and processing steps are minimized. Based on instance segmentation, all crystals containing single crystals, agglomerates and primary crystals in agglomerates are detected and classified with pixel-level accuracy. The deep learning approach shows superior performance to previous image analysis methods and reaches a new level of detail. In experimental studies, L-alanine is crystallized from aqueous solution. A detailed description of size and number of all particles including primary crystals is provided and characteristic measures for the level of agglomeration are given. This can lead to a better process understanding and has the potential to serve as cornerstone for kinetic studies.

Deploying Big Data to Crack the Genotype to Phenotype Code

Mechanistically connecting genotypes to phenotypes is a longstanding and central mission of biology. Deciphering these connections will unite questions and datasets across all scales from molecules to ecosystems. Although high-throughput sequencing has provided a rich platform on which to launch this effort, tools for deciphering mechanisms further along the genome to phenome pipeline remain limited. Machine learning approaches and other emerging computational tools hold the promise of augmenting human efforts to overcome these obstacles. This vision paper is the result of a Reintegrating Biology Workshop, bringing together the perspectives of integrative and comparative biologists to survey challenges and opportunities in cracking the genotype to phenotype code and thereby generating predictive frameworks across biological scales. Key recommendations include promoting the development of minimum "best practices" for the experimental design and collection of data; fostering sustained and long-term data repositories; promoting programs that recruit, train, and retain a diversity of talent; and providing funding to effectively support these highly cross-disciplinary efforts. We follow this discussion by highlighting a few specific transformative research opportunities that will be advanced by these efforts.

Machine learning for neutron scattering at ORNL*

Machine learning (ML) offers exciting new opportunities to extract more information from scattering data. At neutron scattering user facilities, ML has the potential to help accelerate scientific productivity by empowering facility users with insight into their data which has traditionally been supplied by scattering experts. Such support can help in both speeding up common modeling problems for users, as well as help solve harder problems that are normally time consuming and difficult to address with standard methods. This article explores the recent ML work undertaken at Oak Ridge National Laboratory involving neutron scattering data. We cover materials structure modeling for diffuse scattering, powder diffraction, and small-angle scattering. We also discuss how ML can help to model the response of the instrument more precisely, as well as enable quick extraction of information from neutron data. The application of super-resolution techniques to small-angle scattering and peak extraction for diffraction will be discussed.

Application of ultraviolet, visible, and infrared light imaging in protein-based biopharmaceutical formulation characterization and development studies

Imaging is increasingly more utilized as analytical technology in biopharmaceutical formulation research, with applications ranging from subvisible particle characterization to thermal stability screening and residual moisture analysis. This review offers a comprehensive overview of analytical imaging for scientists active in biopharmaceutical formulation research and development, where it presents the unique information provided by the ultraviolet (UV), visible (Vis), and infrared (IR) sections in the electromagnetic spectrum. The main body of this review consists of an outline of UV, Vis, and IR imaging techniques for several (bio)physical properties that are commonly determined during protein-based biopharmaceutical formulation characterization and development studies. The review concludes with a future perspective of applied imaging within the field of biopharmaceutical formulation research.

Polo: an open-source graphical user interface for crystallization screening

A multi-platform open-source Python-based graphical user interface has been developed to provide access to automated classification and data management tools for biomolecular crystallization screening.

Polo is a Python-based graphical user interface designed to streamline viewing and analysis of images to monitor crystal growth, with a specific target to enable users of the High-Throughput Crystallization Screening Center at Hauptman-Woodward Medical Research Institute (HWI) to efficiently inspect their crystallization experiments. Polo aims to increase efficiency, reducing time spent manually reviewing crystallization images, and to improve the potential of identifying positive crystallization conditions. Polo provides a streamlined one-click graphical interface for the Machine Recognition of Crystallization Outcomes (MARCO) convolutional neural network for automated image classification, as well as powerful tools to view and score crystallization images, to compare crystallization conditions, and to facilitate collaborative review of crystallization screening results. Crystallization images need not have been captured at HWI to utilize Polo’s basic functionality. Polo is free to use and modify for both academic and commercial use under the terms of the copyleft GNU General Public License v3.0.

IceBear: an intuitive and versatile web application for research-data tracking from crystallization experiment to PDB deposition

The web-based IceBear software is a versatile tool to monitor the results of crystallization experiments and is designed to facilitate supervisor and student communications. It also records and tracks all relevant information from crystallization setup to PDB deposition in protein crystallography projects. Fully automated data collection is now possible at several synchrotrons, which means that the number of samples tested at the synchrotron is currently increasing rapidly. Therefore, the protein crystallography research communities at the University of Oulu, Weizmann Institute of Science and Diamond Light Source have joined forces to automate the uploading of sample metadata to the synchrotron. In IceBear, each crystal selected for data collection is given a unique sample name and a crystal page is generated. Subsequently, the metadata required for data collection are uploaded directly to the ISPyB synchrotron database by a shipment module, and for each sample a link to the relevant ISPyB page is stored. IceBear allows notes to be made for each sample during cryocooling treatment and during data collection, as well as in later steps of the structure determination. Protocols are also available to aid the recycling of pins, pucks and dewars when the dewar returns from the synchrotron. The IceBear database is organized around projects, and project members can easily access the crystallization and diffraction metadata for each sample, as well as any additional information that has been provided via the notes. The crystal page for each sample connects the crystallization, diffraction and structural information by providing links to the IceBear drop-viewer page and to the ISPyB data-collection page, as well as to the structure deposited in the Protein Data Bank.

A fully automated crystallization apparatus for small protein quantities

In 2003, a fully automated protein crystallization and monitoring system (PXS) was developed to support the structural genomics projects that were initiated in the early 2000s. In PXS, crystallization plates were automatically set up using the vapor-diffusion method, transferred to incubators and automatically observed according to a pre-set schedule. The captured images of each crystallization drop could be monitored through the internet using a web browser. While the screening throughput of PXS was very high, the demands of users have gradually changed over the ensuing years. To study difficult proteins, it has become important to screen crystallization conditions using small amounts of proteins. Moreover, membrane proteins have become one of the main targets for X-ray crystallography. Therefore, to meet the evolving demands of users, PXS was upgraded to PXS2. In PXS2, the minimum volume of the dispenser is reduced to 0.1 µl to minimize the amount of sample, and the resolution of the captured images is increased to five million pixels in order to observe small crystallization drops in detail. In addition to the 20°C incubators, a 4°C incubator was installed in PXS2 because crystallization results may vary with temperature. To support membrane-protein crystallization, PXS2 includes a procedure for the bicelle method. In addition, the system supports a lipidic cubic phase (LCP) method that uses a film sandwich plate and that was specifically designed for PXS2. These improvements expand the applicability of PXS2, reducing the bottleneck of X-ray protein crystallography.

The low-cost Shifter microscope stage transforms the speed and robustness of protein crystal harvesting

Despite the tremendous success of X-ray cryo-crystallography in recent decades, the transfer of crystals from the drops in which they are grown to diffractometer sample mounts remains a manual process in almost all laboratories. Here, the Shifter, a motorized, interactive microscope stage that transforms the entire crystal-mounting workflow from a rate-limiting manual activity to a controllable, high-throughput semi-automated process, is described. By combining the visual acuity and fine motor skills of humans with targeted hardware and software automation, it was possible to transform the speed and robustness of crystal mounting. Control software, triggered by the operator, manoeuvres crystallization plates beneath a clear protective cover, allowing the complete removal of film seals and thereby eliminating the tedium of repetitive seal cutting. The software, either upon request or working from an imported list, controls motors to position crystal drops under a hole in the cover for human mounting at a microscope. The software automatically captures experimental annotations for uploading to the user's data repository, removing the need for manual documentation. The Shifter facilitates mounting rates of 100-240 crystals per hour in a more controlled process than manual mounting, which greatly extends the lifetime of the drops and thus allows a dramatic increase in the number of crystals retrievable from any given drop without loss of X-ray diffraction quality. In 2015, the first in a series of three Shifter devices was deployed as part of the XChem fragment-screening facility at Diamond Light Source, where they have since facilitated the mounting of over 120 000 crystals. The Shifter was engineered to have a simple design, providing a device that could be readily commercialized and widely adopted owing to its low cost. The versatile hardware design allows use beyond fragment screening and protein crystallography.

Multivariate Analysis Applications in X-ray Diffraction

Multivariate analysis (MA) is becoming a fundamental tool for processing in an efficient way the large amount of data collected in X-ray diffraction experiments. Multi-wedge data collections can increase the data quality in case of tiny protein crystals; in situ or operando setups allow investigating changes on powder samples occurring during repeated fast measurements; pump and probe experiments at X-ray free-electron laser (XFEL) sources supply structural characterization of fast photo-excitation processes. In all these cases, MA can facilitate the extraction of relevant information hidden in data, disclosing the possibility of automatic data processing even in absence of a priori structural knowledge. MA methods recently used in the field of X-ray diffraction are here reviewed and described, giving hints about theoretical background and possible applications. The use of MA in the framework of the modulated enhanced diffraction technique is described in detail.

Multiobjective nonlinear model predictive control of pharmaceutical batch crystallizers

A rigorous multiobjective nonlinear model predictive control procedure is implemented in solving problems involving batch crystallizations. This technique does not involve the use of weighting functions and additional restrictive constraints. Three cases are considered. The first is the unseeded batch crystallization involving paracetamol, the second is the seeded batch crystallization concerning potassium nitrate while the third problem deals with a temperature controlled batch crystallizer that involves citric acid anyhydrate. The optimization language pyomo with GAMS interface is used to solve the problems.

An automated platform for in situ serial crystallography at room temperature

Direct observation of functional motions in protein structures is highly desirable for understanding how these nanomachineries of life operate at the molecular level. Because cryogenic temperatures are non-physiological and may prohibit or even alter protein structural dynamics, it is necessary to develop robust X-ray diffraction methods that enable routine data collection at room temperature. We recently reported a crystal-on-crystal device to facilitate in situ diffraction of protein crystals at room temperature devoid of any sample manipulation. Here an automated serial crystallography platform based on this crystal-on-crystal technology is presented. A hardware and software prototype has been implemented, and protocols have been established that allow users to image, recognize and rank hundreds to thousands of protein crystals grown on a chip in optical scanning mode prior to serial introduction of these crystals to an X-ray beam in a programmable and high-throughput manner. This platform has been tested extensively using fragile protein crystals. We demonstrate that with affordable sample consumption, this in situ serial crystallography technology could give rise to room-temperature protein structures of higher resolution and superior map quality for those protein crystals that encounter difficulties during freezing. This serial data collection platform is compatible with both monochromatic oscillation and Laue methods for X-ray diffraction and presents a widely applicable approach for static and dynamic crystallographic studies at room temperature.

COVID-19 Control by Computer Vision Approaches: A Survey

The COVID-19 pandemic has triggered an urgent call to contribute to the fight against an immense threat to the human population. Computer Vision, as a subfield of artificial intelligence, has enjoyed recent success in solving various complex problems in health care and has the potential to contribute to the fight of controlling COVID-19. In response to this call, computer vision researchers are putting their knowledge base at test to devise effective ways to counter COVID-19 challenge and serve the global community. New contributions are being shared with every passing day. It motivated us to review the recent work, collect information about available research resources, and an indication of future research directions. We want to make it possible for computer vision researchers to find existing and future research directions. This survey article presents a preliminary review of the literature on research community efforts against COVID-19 pandemic.

A Searchable Database of Crystallization Cocktails in the PDB: Analyzing the Chemical Condition Space

Nearly 90% of structural models in the Protein Data Bank (PDB), the central resource worldwide for three-dimensional structural information, are currently derived from macromolecular crystallography (MX). A major bottleneck in determining MX structures is finding conditions in which a biomolecule will crystallize. Here, we present a searchable database of the chemicals associated with successful crystallization experiments from the PDB. We use these data to examine the relationship between protein secondary structure and average molecular weight of polyethylene glycol and to investigate patterns in crystallization conditions. Our analyses reveal striking patterns of both redundancy of chemical compositions in crystallization experiments and extreme sparsity of specific chemical combinations, underscoring the challenges faced in generating predictive models for de novo optimal crystallization experiments.

Protein Crystallization Segmentation and Classification Using Subordinate Color Channel in Fluorescence Microscopy Images

The accuracy of detecting protein crystals for fluorescence microscopy images is very critical for high throughput and automated systems. Although the trace fluorescent labeling method could highlight protein crystals, reflection and emission from the fluorescence dye is not always due to crystal regions. Therefore, the analysis of the peak wavelength in the emission spectra of a fluorophore may not always yield effective results. In this paper, we show that using the subordinate color intensity corresponding to longer wavelengths than the peak wavelength of the emission spectra could improve the accuracy of protein crystal detection. Hence, we have built a segmentation method based on the percentile intensity of the subordinate color for trace fluorescently labeled (TFL'd) protein crystallization trial images. Compared to using the dominant color channel, our segmentation method on subordinate color channel was able to reduce the misclassification rate of likely-leads or crystals as non-crystals by the percentage of from 9.71% to 2.02% depending on the classifier. Similarly, the accuracy of classifiers were increased by the percentage of from 1.77% to 5.53%. Our method reached around 94% accuracy while keeping misclassification of likely-leads and crystals as non-crystals below 1%. Moreover, to evaluate the generalizability of our method, we have conducted new wet lab experiments on two proteins, Concanavalin A (Con A) and Ab inorganic pyrophosphate (AbIPPase), and the misclassification rate was below 1%. Our experiments show that using the subordinate channel may be more helpful for TFL'd protein trial image classification.

Tools to Ease the Choice and Design of Protein Crystallisation Experiments

The process of macromolecular crystallisation almost always begins by setting up crystallisation trials using commercial or other premade screens, followed by cycles of optimisation where the crystallisation cocktails are focused towards a particular small region of chemical space. The screening process is relatively straightforward, but still requires an understanding of the plethora of commercially available screens. Optimisation is complicated by requiring both the design and preparation of the appropriate secondary screens. Software has been developed in the C3 lab to aid the process of choosing initial screens, to analyse the results of the initial trials, and to design and describe how to prepare optimisation screens.

High-Throughput Crystallization Pipeline at the Crystallography Core Facility of the Institut Pasteur

The availability of whole-genome sequence data, made possible by significant advances in DNA sequencing technology, led to the emergence of structural genomics projects in the late 1990s. These projects not only significantly increased the number of 3D structures deposited in the Protein Data Bank in the last two decades, but also influenced present crystallographic strategies by introducing automation and high-throughput approaches in the structure-determination pipeline. Today, dedicated crystallization facilities, many of which are open to the general user community, routinely set up and track thousands of crystallization screening trials per day. Here, we review the current methods for high-throughput crystallization and procedures to obtain crystals suitable for X-ray diffraction studies, and we describe the crystallization pipeline implemented in the medium-scale crystallography platform at the Institut Pasteur (Paris) as an example.

BraggNet: integrating Bragg peaks using neural networks

Neutron crystallography offers enormous potential to complement structures from X-ray crystallography by clarifying the positions of low-Z elements, namely hydrogen. Macromolecular neutron crystallography, however, remains limited, in part owing to the challenge of integrating peak shapes from pulsed-source experiments. To advance existing software, this article demonstrates the use of machine learning to refine peak locations, predict peak shapes and yield more accurate integrated intensities when applied to whole data sets from a protein crystal. The artificial neural network, based on the U-Net architecture commonly used for image segmentation, is trained using about 100 000 simulated training peaks derived from strong peaks. After 100 training epochs (a round of training over the whole data set broken into smaller batches), training converges and achieves a Dice coefficient of around 65%, in contrast to just 15% for negative control data sets. Integrating whole peak sets using the neural network yields improved intensity statistics compared with other integration methods, including k-nearest neighbours. These results demonstrate, for the first time, that neural networks can learn peak shapes and be used to integrate Bragg peaks. It is expected that integration using neural networks can be further developed to increase the quality of neutron, electron and X-ray crystallography data.

Time-Dependent Multi-Light-Source Image Classification Combined With Automated Multidimensional Protein Phase Diagram Construction for Protein Phase Behavior Analysis

Image-based protein phase diagram analysis is key for understanding and exploiting protein phase behavior in the biopharmaceutical field. However, required data analysis has become a notorious time-consuming task since high-throughput screening approaches were implemented. A variety of computational tools have been developed to support analysis, but these tools primarily use end point visible light images. This study investigates the combined effect of end point and time-dependent image features obtained from cross-polarized and ultraviolet light features, supplementary to visible light, on protein phase diagram image classification. In addition, external validation was performed to evaluate the classification algorithm's applicability to support protein phase diagram scoring. The predicted protein phase behavior classes were subsequently used to automatically construct multidimensional protein phase diagrams to prevent image information loss without complicating the used image classification algorithm. Combining end point and time-dependent features from 3 light sources resulted in a balanced accuracy of 86.4 ± 4.3%, which is comparable to or better than more complex classifiers reported in literature. External validation resulted in a correct formulation classification rate of 91.7%. Subsequent automated construction of the multidimensional protein phase diagrams, using predicted classes, allowed visualization of details such as crystallization rate and protein phase behavior type coexistence.

DeepCentering: fully automated crystal centering using deep learning for macromolecular crystallography

High-throughput protein crystallography using a synchrotron light source is an important method used in drug discovery. Beamline components for automated experiments including automatic sample changers have been utilized to accelerate the measurement of a number of macromolecular crystals. However, unlike cryo-loop centering, crystal centering involving automated crystal detection is a difficult process to automate fully. Here, DeepCentering, a new automated crystal centering system, is presented. DeepCentering works using a convolutional neural network, which is a deep learning operation. This system achieves fully automated accurate crystal centering without using X-ray irradiation of crystals, and can be used for fully automated data collection in high-throughput macromolecular crystallography.

From invisibility to readability: Recovering the ink of Herculaneum

The noninvasive digital restoration of ancient texts written in carbon black ink and hidden inside artifacts has proven elusive, even with advanced imaging techniques like x-ray-based micro-computed tomography (micro-CT). This paper identifies a crucial mistaken assumption: that micro-CT data fails to capture any information representing the presence of carbon ink. Instead, we show new experiments indicating a subtle but detectable signature from carbon ink in micro-CT. We demonstrate a new computational approach that captures, enhances, and makes visible the characteristic signature created by carbon ink in micro-CT. This previously “unseen” evidence of carbon inks, which can now successfully be made visible, is a discovery that can lead directly to the noninvasive digital recovery of the lost texts of Herculaneum.

Volumetric Segmentation via Neural Networks Improves Neutron Crystallography Data Analysis

Crystallography is the powerhouse technique for molecular structure determination, with applications in fields ranging from energy storage to drug design. Accurate structure determination, however, relies partly on determining the precise locations and integrated intensities of Bragg peaks in the resulting data. Here, we describe a method for Bragg peak integration that is accomplished using neural networks. The network is based on a U-Net and identifies peaks in three-dimensional reciprocal space through segmentation, allowing prediction of the full 3D peak shape from noisy data that is commonly difficult to process. The procedure for generating appropriate training sets is detailed. Trained networks achieve Dice coefficients of 0.82 and mean IoUs of 0.69. Carrying out integration over entire datasets, it is demonstrated that integrating neural network-predicted peaks results in improved intensity statistics. Furthermore, using a second dataset, the possibility of transfer learning between datasets is shown. Given the ubiquity and growing complexity of crystallography, we anticipate integration by machine learning to play an increasingly important role across the physical sciences. These early results demonstrate the applicability of deep learning techniques for integrating crystallography data and suggest a possible role in the next generation of crystallography experiments.

An inexpensive system for imaging the contents of multi-well plates

This paper describes the construction of a low-cost, open-source system for imaging the contents of multi-well plates.

An inexpensive system for automated imaging of the contents of 12-, 24- and 96-well plates has been built. The xyz stage is constructed from parts from a light-duty computer numerical control wood-carving/engraving machine, and the Arduino-based board was wired so that it can trigger still images or movies though a microscope-mounted digital camera. The translation stage provides reproducible three-dimensional movement of the sample over a volume of 160 mm in x, 100 mm in y and 40 mm in z. A Python script generates the G-code command file that scans the plate and collects a series of z-stacked images of each sample. A second Python script automates the calculation of images with a digitally enhanced depth of field. The imaging system is currently being used to facilitate screening for protein crystals, but it could be used to automate the imaging of many other types of samples in multi-well plates.

Synchrotron Big Data Science

The rapid development of synchrotrons has massively increased the speed at which experiments can be performed, while new techniques have increased the amount of raw data collected during each experiment. While this has created enormous new opportunities, it has also created tremendous challenges for national facilities and users. With the huge increase in data volume, the manual analysis of data is no longer possible. As a result, only a fraction of the data collected during the time- and money-expensive synchrotron beam-time is analyzed and used to deliver new science. Additionally, the lack of an appropriate data analysis environment limits the realization of experiments that generate a large amount of data in a very short period of time. The current lack of automated data analysis pipelines prevents the fine-tuning of beam-time experiments, further reducing their potential usage. These effects, collectively known as the "data deluge," affect synchrotrons in several different ways including fast data collection, available local storage, data management systems, and curation of the data. This review highlights the Big Data strategies adopted nowadays at synchrotrons, documenting this novel and promising hybridization between science and technology, which promise a dramatic increase in the number of scientific discoveries.