Journal of Applied Crystallography

X-ray fiber diffraction patterns exhibit four-fold symmetry that can be exploited, through folding and averaging, to improve signal-to-noise ratio. Accurate folding requires a precise sub-pixel estimate of the symmetry center and precise orientation of the meridional pattern axis to the fiber axis: small center or angular errors blur diffraction features, reduce layer-line sharpness, and introduce errors in spacing measurements. A pixel-level estimate is often too imprecise for this purpose, and detector gaps further complicate the alignment objective. We formulate the masked quadrant-folding problem, define a four-quadrant symmetry loss that consistently excludes invalid pixels, and evaluate several refinement strategies: hierarchical coarse-to-fine grid search; ECC-based rigid registration with global center/orientation correction fitting; ECC registration followed by local gradient refinement; and a hybrid that appends a local grid search on a cropped pattern. Direct gradient optimization from the rough QF alignment was found to be unreliable. Grid search provides a robust, interpretable baseline that directly minimizes the folding objective but is substantially slower than registration; ECC gives a fast near-correct alignment, and the hybrid closes the accuracy gap to brute-force search at a fraction of its runtime. On real datasets with calibration data, applying a calibration center with optimized rotation is effectively optimal. The hybrid center-refinement method has been integrated into the MuscleX package.

Three-dimensional electron diffraction (3DED) is an emerging and useful technique for solving molecular structures of small and biological macro-molecules from nanometre-sized crystals. We present our automated data processing workflow for 3DED datasets collected at Diamond Light Sources electron Bio-Imaging Centre (eBIC). For this purpose, we developed a package called AutoED. The processing pipeline includes data collection, analysis of the beam position, metadata gathering, file conversion, and finally data processing using xia2 (which supports both DIALS and XDS). The processing results are captured in a summary report produced by AutoED. Our main goal is to reduce the workload of electron diffraction scientists, but also to enforce good standards already used in macromolecular crystallography (MX). All the collected 3DED datasets are automatically converted into NeXus data format which is considered a Gold Standard for MX. This standardized data format allows for all the relevant metadata about the experiment to be kept together with diffraction images. We also discuss the methods used in AutoED to determine the electron beam position on diffraction images.

Time-resolved crystallography can be inaccessible for membrane proteins and other difficult-to-crystallize targets for which suitable crystals are scarce. Here we present an acquisition strategy that records consecutive diffraction images during a small crystal rotation synchronized to a single activation pulse. Matching image segments are merged across crystals to reconstruct successive time windows, enabling a single crystal ensemble to capture an extended structural trajectory. Using the Mesh&Collect workflow at a synchrotron, we obtained ten consecutive 7.5-ms time windows at 2 [A] resolution from 1,255 microcrystals of a bacteriorhodopsin triple mutant at room temperature. An additional 790 crystals extended the observation window to 500 ms. Rotation-based processing, retrospective time binning and dose control maximize information extracted from each crystal, while compatible measurements can be pooled across acquisitions. This approach makes millisecond time-resolved crystallography practical for crystal-limited targets at synchrotrons, with further development potentially extending the temporal resolution down to 1 ms.

Directional anisotropy in electron density provides key information about chemical bonding that is not readily accessible from conventional electron-density maps. Here, a model-independent framework is presented for decomposing experimental structure factors into angular components using spherical harmonics. Reciprocal-space projection onto spherical harmonics followed by standard Fourier synthesis yields angularly filtered density maps. The{ell} = 0 component captures the isotropic part of the density, while the{ell} = 1 components resemble px, py and pz-like dipolar functions that highlight directional electronic structure. Applications to high-resolution datasets, including urea, the Gly-Ala dipeptide and a 0.97 [A]{beta}-lactamase structure, reveal chemically interpretable dipolar features associated with carbonyl and amide bonds, N-H interactions and aromatic{pi} systems. Quantitative analysis using bond-centred sampling demonstrates stable dipolar signatures that remain detectable under moderate resolution truncation. These results establish spherical-harmonic angular decomposition as a practical framework for extracting directional electronic information from crystallographic electron-density maps. SynopsisAngular decomposition of experimental structure factors reveals dipolar anisotropy and directional electron-density features that are directly meaningful for chemical interpretation.

Accurate assessment of ligand coordinate-density consistency across different resolutions remains challenging in macromolecular crystallography. We introduce the atomic Box Correlation Coefficient (aBCC), an atom-level metric for evaluating the consistency between ligand atomic coordinates and electron density in a resolution-standardized framework. To predict aBCC values from electron-density maps, we developed QAEmap, a machine-learning model based on three-dimensional convolutional neural networks (3D-CNNs). The model was trained using Fourier-truncated electron-density maps and corresponding ligand coordinates generated from high-resolution structures in the Protein Data Bank. It was evaluated using both Fourier-truncated electron-density maps and experimentally determined PDB structures. was evaluated using both Fourier-truncated electron-density maps and experimentally determined PDB structures.The prediction accuracy gradually decreased with decreasing resolution, but remained reliable up to [~]3.5 [A]. These results demonstrate that aBCC enables resolution-standardized atom-wise evaluation of coordinate-density consistency across different resolutions and provide a foundation for further development and refinement of machine learning-based coordinate validation. SynopsisWe introduce the atomic box correlation coefficient (aBCC), a machine learning-based metric for the resolution-standardized atom-level evaluation of ligand coordinate-density consistency in crystallographic structures. aBCC provides a common framework for assessing and communicating the local coordinate reliability between structural biologists and researchers in structure-based drug discovery.

Damage to biological specimens by the electron beam is the fundamental resolution-limiting factor in cryoelectron microscopy (cryo-EM) single particle analysis. There is, however, currently no method to accurately infer fluence-dependent changes to the specimen structure during electron irradiation. We develop a Bayesian framework to fit a sequence of atomic models to a series of cryo-EM reconstructions produced at increasing fluence. In particular, our algorithm is able to infer the ensemble average position and atomic displacement parameter of every atom in the macromolecule as a function of fluence. Application of the algorithm to cryo-EM datasets shows that the molecule expands during imaging and identifies environment-dependent variations in beam-induced damage. We use our results to propose a stochastic process model of this phenomenon. We envisage that our method will lead to a better mechanistic understanding of radiation damage to biological specimens and may contribute to efforts to mitigate its effects.

This paper presents and critically reviews the results of original and some literature based experimental studies conducted by the authors last years on the structural organization of DNA in dormant (starvation stress), anabiotic dormant (4 HR treatment) E. coli cells, as well as the K12 {Delta}dps strain, which lacks the Dps protein (Dps null E. coli). The experimental data includes small-angle synchrotron radiation diffraction (SAXS) and transmission electron microscopy (TEM) data. Synchrotron radiation diffraction experiments on K12{Delta}dps cells allowed us to conclude that peaks at 44.3, 22.1, and 14.8 angstrom resolutions are associated exclusively with ordered DNA organization. Peaks at 44.3, 22.1, and 14.8 angstrom resolutions are also observed for samples of dormant (starvation stress) cells and anabiotically dormant cells. Therefore, this ordered DNA organization also applies to samples of dormant and anabiotically dormant cells. A model is proposed that considers the ordered DNA organization in the cell as a cholesteric liquid crystal. The powder diffraction pattern calculated based on this model is compared with experimental small angle X ray scattering (SAXS) data obtained on Dps-null cell samples. The model completely reproduces the key features of the experimental diffraction pattern from Dps-null cell samples. Accordingly, the cholesteric liquid crystal model corresponds to DNA packaging in dormant and anabiotically dormant cells. Cholesteric liquid crystal ordering should be further considered in all models of cellular DNA packaging. To address the question of which structural organization of DNA predominates in the cell: the cholesteric liquid crystal or nanocrystalline or whether they coexist and fully manifest themselves under different external conditions, it is necessary to utilize the latest methodological advances in structural analysis.

Time-resolved serial crystallography (TR-SX) has emerged as a powerful method for capturing ultrafast structural dynamics in proteins. TR-SX continues to produce remarkable studies, revealing previously unobserved transient states and providing deeper insights into processes such as drug targeting, DNA repair, and photosynthesis. However, extracting weak structural signals from noisy time-resolved datasets remains a major challenge. Robust computational methods are therefore required to isolate the signals associated with the underlying transient states. Importantly, this should be performed in reciprocal space to preserve compatibility with established downstream structure refinement workflows. Here, we introduce a framework for kinetic decomposition directly in reciprocal space that enables separation of kinetically distinct structural states. The method decomposes crystallographic data according to a predefined kinetic model, improving the recovery of weak transient signals and enhancing mechanistic interpretation from limited time-resolved datasets. We validate the framework using simulated data based on a previously published time-resolved crystallography study and demonstrate its application to a new TR-SX dataset comprising 17 time points. We show that the method separates the reciprocal space signatures of four intermediates by incorporating kinetic information from a predefined reaction model. This establishes a workflow for extracting kinetic states directly from time-resolved X-ray diffraction data that can be seamlessly integrated into existing crystallographic structure-determination pipelines.

Here, we present MC-Bayes, a Python-based script for processing cryo-electron microscopy EER movies on one or more GPUs using MotionCor3 in a user-friendly manner. Further, it generates the .star files necessary for RELION to perform Bayesian polishing (a.k.a.: reference-based motion correction) with EER movies. Until now, Bayesian polishing of EER data was only possible if the CPU-based "RELIONCor" implementation of MotionCor2 was used, which is sub-optimal on GPU-heavy cryo-EM processing systems. This wrapper was created for those facilities and/or users who may have (many) powerful GPUs, but for whatever reason have few CPU cores or less system RAM. Leveraging MotionCor3, MC-Bayes allows motion correction of EER data 2 or more times faster (depending on system) than the RELION CPU implementation, except in circumstances where dozens or hundreds of CPU cores with high quantities of system RAM can be utilised.

Proteins are molecular machines whose dynamics govern their function. Using time-resolved crystallography, the basis of protein mechanics can be uncovered by physical perturbation on protein crystals. Perturbations are often vector fields that, in crystals, break symmetry. Lacking existing tools for analyzing symmetry breaking, we built open-source software to do so. We applied this method to three time-resolved crystallography experiments and analyzed the effect of vectorial perturbations on macromolecular crystal structure. We demonstrate that our method can handle symmetry reduction in all 65 Sohncke space groups.

Carbohydrates are essential molecules involved in a wide range of biological processes. While their structural representation in the Protein Data Bank (PDB) has been hindered due to their highly diverse and flexible nature, advancements on X-ray crystallography and CryoEM are accelerating the obtaining of their experimental structures, which in turn supports improved analyses of their structural properties, such as the role of their non-chair conformations on biological processes. In this context, the current work provides a comprehensive and up to date overview of carbohydrate structures within the PDB through classification, quantification, and characterization of monosaccharides and glycan chains, incorporating Cremer-Pople puckering coordinates to evaluate pyranose conformations. Our findings show that, even with a steep increase in CryoEM structures including more complex oligosaccharides, carbohydrate data remains concentrate dinsmallglycanchains, with a about 4% of non-chair conformation sintopresolution structures. Free energy landscapes for the pyranose ring in the presence and absence of nine different binding proteins, including both enzymes and non-enzymes, reinforce the preferred stability of chairs and the rarity of conformational induction upon binding.

Reported relative biological effectiveness (RBE) values for low-energy X-rays disagree, assays scoring initial DNA double-strand breaks (DSBs) returning about 1.1 and chromosome-level assays 2 to 4. Whether radiation quality varies within the diagnostic range, and how its comparison with a megavoltage reference depends on target scale, has not been quantified on a tube-potential series. A tungsten-anode tube with 1 mm Be and 2.5 mm Al filtration, with copper added in some cases, was modelled in PHITS for 40 to 200 kV. The spectra were transported into a water phantom in which absorbed dose, lineal-energy densities and cluster size distributions were scored for target diameters of 3 nm to 1 micrometre against a cobalt-60 reference; DSB yields were computed in the electron track-structure mode with the PHITS DNA damage tally. Between 40 and 120 kV the depth-dose ratio changed by a factor of 5.7 and the tube output by a factor of 42, whereas the dose-mean lineal energy varied by 2.5 % at 1 micrometre and 1.2 % at 3 nm against a reproducibility of 0.3 %. Relative to cobalt-60 it was 2.05 times larger at 1 micrometre but only 1.08 times larger at 3 nm, while DSB yields per unit dose were 5 to 7 % higher and constant across the range within the 2 % bound set by the statistics. Tube potential therefore changes the amount and distribution of dose but not its physical quality, and a stated RBE is incomplete without the target scale implied by the endpoint.

Proteins naturally adopt multiple conformations in mediating cellular processes, and ensemble models are used to fit X-ray crystallography data that captures this heterogeneity. In practice, ensemble refinement produces only minor improvements in agreement with experimental data (R-free) over single-conformation models. It has recently been shown that ensemble models are universally trapped, or "tangled"; refinement algorithms strain each individual conformation in the model to fit the electron density in its immediate vicinity, missing more harmonious ways to arrange the collection of protein conformations to fit the electron density. Here, we demonstrate that this type of trap may be escaped by formulating the construction of low-energy conformations from individual conformer coordinates as an integer linear programming problem. The method successfully recovered the two original protein conformations from a previously published synthetic dataset that traps current refinement methods. Inclusion of the method in an automated refinement procedure with real data is shown to improve R-free and reduce geometric strain in a four-conformation model by comparison with controls. Applying this method in combination with human input and fitting low-occupancy waters to density features in the bulk solvent, we produce models for deposited datasets of three separate 14-19 kDa proteins with greatly improved geometry and R-factors. This includes a 0.77 [A] six-conformation model of the SARS-CoV-2 macrodomain Mac1 (PDB ID: 44PS) with an R-work of 4.7% and an R-free of 6.4%.

Cryo-electron tomography of 100-200 nm lamellae prepared by focused ion beam milling enables structural analysis of macromolecules within cells, tissues, and small organisms, but lamella preparation remains time-consuming and limits throughput. Although automated workflows enable unsupervised fabrication, they commonly produce variable results because fixed, pre-calibrated parameters are applied without accounting for lamella-to-lamella variability. To overcome these limitations, we developed an Adaptive Milling approach that adapts milling parameters based on imaging feedback. We implemented this approach for the critical polishing stage as Adaptive Polishing, which uses machine-learning-based segmentation of scanning electron microscopy images to identify lamella features predictive of imminent collapse. These are used to periodically assess whether to continue or stop thinning during polishing. We performed a systematic comparison of the quality of lamellae prepared with Adaptive Polishing and conventional automation, and demonstrate that Adaptive Polishing produces thinner lamellae with greater consistency in thickness, and performs well for in situ cryo-electron tomography. Adaptive Polishing was tested on three different microscopes with minimal model retraining required, demonstrating transferability of the approach. We implemented Adaptive Polishing as an open-source plugin for fibsemOS together with machine-learning model training software and pre-trained models.

Atomistic simulations often struggle to capture the full conformational landscape of biomolecular complexes on experimentally relevant timescales, leaving gaps in our understanding of functionally relevant thermodynamics and kinetics. Cryo-electron microscopy (cryo-EM) provides high-resolution structural snapshots spanning multiple conformational states, but translating these static images into dynamic mechanisms remains challenging. Existing approaches for integrating cryo-EM data with molecular simulations either perturb the underlying dynamics through artificial biasing potentials or remain confined to conformations already present in the prior ensemble. We introduce cryoWEight, an automated and extensible Python framework that couples iterative weighted ensemble (WE) simulations with ensemble reweighting against cryo-EM particle images to steer simulation trajectories towards conformational states consistent with cryo-EM populations. We have found that reweighting in the image space introduces a directional bias towards the target distribution even when the prior ensemble has no overlap with the target distribution in the collective variable space. The reweighting requires only the state likelihoods that set the ensemble weights, and any observable that defines these likelihoods can serve as the target distribution, therefore generalizing beyond the application of cryo-EM. cryoWEight is validated on three proteins: chignolin, the N-terminal domain of ribosomal protein L9, and adenylate kinase. The framework converges towards the target conformational states across different sampling strategies and noise levels, providing a scalable route to integrate cryo-EM data with enhanced sampling approaches.

Structural model building and refinement need many decisions between different programs. These decisions often depend on an experienced researcher who knows how the programs work together and how to read experimental data. We used Claude Code with Claude Opus 5.5 to supervise this work for 24 cryo-EM and X-ray cases. The work finished in about 36 hours on one workstation. The agent prepared the inputs, ran established programs, read the validation results, compared candidate models and chose the next step. For 15 deposited cryo-EM models with weak starting geometry, the median MolProbity score improved from 2.62 to 1.76. For three crystal structures, the agent models came close to the published Rfree values. Chemical information such as ligands, ions and modifications must be provided by the researcher, in the same way as the sequence. With this information, the agent can place and check these components. These results show that agent supervision is acceptable when every decision can be inspected. The researcher makes the final scientific judgment. We suggest practical requirements for this kind of use. Future AI agents may help researchers obtain high quality and reliable structures of biological macromolecules.

The primary question in this work is whether a turbidity trace contains more than one resolved kinetic event before any rate or aggregate-mass information is supplied. A blind, three-gate hierarchical Gompertz procedure applied to human interleukin-1 receptor antagonist (IL-1ra) at 53 {degrees}C in phosphate retained one phase at 1 and 2 mg/mL, but selected two phases at 5 and 10 mg/mL. Thus the first resolved multiphase behavior appears at the low protein concentration of 5 mg/mL. Source optical-rate markers were withheld until after selection. Those observed optical rates were then compared with blind-fit phase rates, and the same comparison was expressed on a conditional mass-equivalent scale using the published 5.3-fold Type-I/Type-II turbidity response ratio. A separate 50 {degrees}C citrate series (4, 6, 8, 10, and 14 mg/mL) also selected two phases in every retained trace; its printed Vmax values were used only as post-selection comparators. Finally, two challenging 40 {degrees}C high-concentration traces (180 and 200 mg/mL) were digitized from the instrument-output plot and each retained two phases. All three conditions are analyzed independently. Blind recognition is therefore the central result; the mass-equivalent conversion is a secondary, post-selection interpretation of optical phase rates. The absolute {micro}M/min scale remains conditional on an endpoint aggregate fraction measured elsewhere.

Solution X-ray scattering provides unique structural information on biomolecules under biological conditions, resolving conformational heterogeneity and time-resolved structural changes. The scattering profiles contain limited information, and interpretation largely relies on fitting candidate structures guided by priors. Direct reconstruction of electron density maps is desirable, but so far has been prevented by the difficulty of incorporating such prior knowledge. Here we propose XSSDense, a framework that couples a variational autoencoder trained on electron densities from predicted or simulated protein ensembles with a genetic algorithm to refine densities against scattering data. We validate XSSDense on synthetic data for crambin, recover the conformational heterogeneity of the unfolded state of Avena sativa light-oxygen-voltage sensing domain 2, resolve a de-novo density for the pre-unfolding state of the same protein, and provide a new structural description of the signalling-state ensemble of photoactive yellow protein. XSSDense enables structurally grounded electron density reconstructions that intrinsically capture conformational heterogeneity.

High-speed atomic force microscopy (HS-AFM) experiments provide unique insights into nanoscale biological processes by directly visualizing the surface topography dynamics of biomolecules, such as proteins, over extended periods of time and generating large imaging data sets. However, the lack of computational methods for automated data analysis has thus far limited the full utilization of these data for the quantitative understanding of single-molecule dynamic. Here, we present the AFM Analysis Arena (A3) within BioAFMviewer as a software platform for automated, high-throughput quantitative analysis of AFM topographic imaging data within a user-friendly interactive graphical interface, implementing a convenient workflow from raw HS-AFM (AFM) imaging to statistical analysis. In several applications to HS-AFM imaging data, we demonstrate methods of automatized molecule detection and dynamic tracking to highlight the available capacities, aiming to transform HS-AFM measurements from observations towards an analytical tool in nanoscale biology.

The visualisation of large biological samples in cryo-electron tomography requires both few-nanometre resolution and a wide, ideally >10 micrometrre, field of view. For a single image, increasing magnification trades increased resolution for a smaller field of view. Montaging, where high-magnification images are stitched together to increase field of view, addresses the trade-off but is dose-inefficient with round beams that are standard in transmission electron microscopy (TEM), since this illumination shape does not tessellate. Our workflow greatly simplifies montaging schemes for dose-sensitive specimens, allowing large montages of any shape - from individual cells or other features of interest up to whole lamellae - to be collected at high resolution without compromising electron dose efficiency. We demonstrate the workflow by acquiring montage tomograms of whole yeast (Saccharomyces cerevisiae) cells in cryo-FIB lamellae and of human neurons.