Proteolysis-targeting chimeras (PROTACs) and molecular glues promote targeted protein degradation by recruiting an E3 ligase to proteins of interest (POIs). An accurate 3D structure of the ternary complex formed by E3 li...Proteolysis-targeting chimeras (PROTACs) and molecular glues promote targeted protein degradation by recruiting an E3 ligase to proteins of interest (POIs). An accurate 3D structure of the ternary complex formed by E3 ligase, ligand, and POI is central to the rational design of degraders. Elucidating this structure with crystallography or cryo-EM can be challenging due to conformational flexibility, dynamic protein-protein interactions, and high-dimensional binding landscapes. To facilitate structure-based design in the absence of an experimental structure, computational approaches have been proposed: (i) multistep methods involving traditional docking pipelines, and (ii) single-step methods with deep learning models to directly predict the complex structure. Multistep methods are limited by sampling complexity, accurate input structures, scoring accuracy, and computational cost, while single-step methods are faster but are constrained by training-data scarcity. Here, we examine recent advances and emerging tools in modeling ternary complexes, critically discuss their predictive power and limitations, and highlight remaining challenges.
Curr Opin Struct Biol
· 2026 Feb · PMID 41604887
·
Full text
Cryptic pockets are promising targets for drug discovery that greatly expand the druggable proteome. In particular, they can provide opportunities to target proteins previously thought to be "undruggable" due to a lack o...Cryptic pockets are promising targets for drug discovery that greatly expand the druggable proteome. In particular, they can provide opportunities to target proteins previously thought to be "undruggable" due to a lack of pockets in structures of the ground state. However, their transient and hidden nature renders them difficult to detect through conventional experimental screening methods. Recent advances in computational methodologies and resources have greatly enhanced our ability to identify and characterize such elusive pockets. This review highlights key developments in computational approaches, including physics-based molecular dynamics simulations, artificial intelligence-driven models, and hybrid strategies that integrate both to enhance cryptic pocket discovery and functional interpretation.
Protein design has the potential to revolutionize biotechnology and medicine. While most efforts have focused on proteins with well-defined structures, increased recognition of the functional significance of intrinsicall...Protein design has the potential to revolutionize biotechnology and medicine. While most efforts have focused on proteins with well-defined structures, increased recognition of the functional significance of intrinsically disordered regions, together with improvements in their modeling, has paved the way to their computational design. This review summarizes recent advances in designing intrinsically disordered regions with tailored conformational ensembles, molecular recognition, and phase behavior. We discuss challenges in combining models of predictive accuracy with scalable design workflows and outline emerging strategies that integrate knowledge-based, physics-based, and machine-learning approaches.
Understanding biomolecular function depends on bridging experimental observables with models that capture structural, stationary, and dynamical properties. Molecular dynamics (MD) simulations, in principle provide a brid...Understanding biomolecular function depends on bridging experimental observables with models that capture structural, stationary, and dynamical properties. Molecular dynamics (MD) simulations, in principle provide a bridge, but the sampling problem remains a fundamental roadblock toward this goal. In this mini-review, I outline recent progress in the area of Generative MD (GenMD)-an approach where generative AI (GenAI) is used to mimic the statistical distributions resulting from MD simulations, which are inaccessible using current numerical algorithms. Here, I highlight a few exemplars of GenMD and then outline open problems and current limitations.
Cryo-electron microscopy (cryo-EM) has emerged as one of the most powerful techniques for determining the structures of biological macromolecules. The ultimate goal of cryo-EM is to determine the atomic structure of targ...Cryo-electron microscopy (cryo-EM) has emerged as one of the most powerful techniques for determining the structures of biological macromolecules. The ultimate goal of cryo-EM is to determine the atomic structure of target molecules, where map postprocessing and atomic-model building are two crucial final steps of the cryo-EM pipeline. With the fast development of artificial intelligence, deep learning has been implemented in various stages of cryo-EM. Here, we present a comprehensive overview of recent advances in map postprocessing and model building for cryo-EM maps with focuses on deep learning-based methods. We also discuss the advantages and limitations of current approaches as well as challenges that are left for future research.
Curr Opin Struct Biol
· 2026 Feb · PMID 41401629
·
Full text
Ras guaosine triphosphate hydrolase (GTPase) are central to key cell signaling pathways and, when mutated, drive many cancers. Thought to be undruggable, dramatic progress has been made in the last decade in the design a...Ras guaosine triphosphate hydrolase (GTPase) are central to key cell signaling pathways and, when mutated, drive many cancers. Thought to be undruggable, dramatic progress has been made in the last decade in the design and screening of drugs, in large part thanks to an emerging detailed understanding of Ras conformational changes, excited/sparsely populated states, and allosteric interactions with ligands and protein-binding partners. This perspective reviews this recent progress and how it has been enabled by deep mutational scanning, solution nuclear magnetic resonance (NMR) spectroscopic studies, as well as computational modeling and simulations. We critically discuss these developments over the last 5 years, also for the GTPase-activating proteins (GAP) NF1 and plexin, effector proteins, plexin and Raf, and make suggestions on the gaps in our understanding that still exist.
KRAS, a member of the RAS family of small GTPases, is frequently mutated in cancers and localizes to the inner leaflet of the plasma membrane, where it has been suggested to form dimers and higher-order oligomers (nanocl...KRAS, a member of the RAS family of small GTPases, is frequently mutated in cancers and localizes to the inner leaflet of the plasma membrane, where it has been suggested to form dimers and higher-order oligomers (nanoclusters). These nanoclusters are dynamic, reversible, and may be critical for signal amplification and specificity. In this perspective, we review the current understanding of KRAS oligomerization on membranes and its relevance for downstream signaling. Moreover, we discuss potential KRAS-KRAS interfaces, the effectors contributing to nanoclustering, such as the influence of the membrane lipid composition on KRAS nanoclustering, and outline the effect of small molecules on the RAS signaling pathway and nanoclustering.
Intrinsically disordered proteins (IDPs) play essential roles in regulation, signaling, and phase separation, yet their structural complexity cannot be captured by a single conformation. Instead, they populate dynamic en...Intrinsically disordered proteins (IDPs) play essential roles in regulation, signaling, and phase separation, yet their structural complexity cannot be captured by a single conformation. Instead, they populate dynamic ensembles that encode a context-dependent function. Recent advances in experimental techniques coupled with physics-based simulations, coarse-grained models, and machine learning, have transformed our ability to generate and interpret IDP ensembles. Integrative frameworks now combine complementary data with computational approaches to refine ensembles at both local and global levels. Nevertheless, challenges remain in benchmarking, error estimation, and modeling assemblies involving protein-protein and protein-nucleic acid interactions. We highlight recent progress and outline the emerging directions that will shape the next generation of ensemble determination methods.
Biomolecular condensates formed through protein phase separation are critical for cellular organization and regulation. Recent years have seen rapid growth in computational methods predicting proteins' phase separation p...Biomolecular condensates formed through protein phase separation are critical for cellular organization and regulation. Recent years have seen rapid growth in computational methods predicting proteins' phase separation propensity and condensate localization, fueled by expanding datasets and advances in machine learning. Here, we review recent progress and limitations of state-of-the-art tools. Despite improvements, current models often fail to capture the complexity of phase separation, which depends on molecular interactions and contextual factors such as temperature, ionic strength, and macromolecular crowding. Encouragingly, new approaches are beginning to incorporate these biological variables, moving toward more physiologically relevant predictions. To accelerate progress, we advocate for stricter metadata standards and a coordinated, community-wide benchmarking of predictive tools to ensure robust and reproducible models for inference of protein phase behavior.
Much of our mechanistic understanding of the functions of biological macromolecules is based on static structural experiments, which can be modelled either as single structures or conformational ensembles. While these pr...Much of our mechanistic understanding of the functions of biological macromolecules is based on static structural experiments, which can be modelled either as single structures or conformational ensembles. While these provide us with invaluable insights, they do not directly reveal that molecules are inherently dynamic. Advances in time-dependent and time-resolved experimental methods have made it possible to capture the dynamics of biomolecules at increasingly higher spatial and temporal resolutions. To complement these, computational models can represent the structural and dynamical behaviour of biomolecules at atomistic resolution and femtosecond timescale, and are therefore useful to interpret these experiments. Here, we review the progress in integrating simulations with dynamical experiments, focusing on the combination of simulations with time-resolved and time-dependent experimental data.
Curr Opin Struct Biol
· 2026 Feb · PMID 41370988
·
Full text
Cryogenic electron tomography (cryoET) has emerged as a powerful tool for studying the structural heterogeneity of proteins and their complexes, offering insights into macromolecular dynamics directly within cells. Drive...Cryogenic electron tomography (cryoET) has emerged as a powerful tool for studying the structural heterogeneity of proteins and their complexes, offering insights into macromolecular dynamics directly within cells. Driven by recent computational advances, including powerful machine learning frameworks, researchers can now resolve both discrete structural states and continuous conformational changes from 3D subtomograms and stacks of 2D particle-images acquired across tilt-series. In this review, we survey recent innovations in particle classification and heterogeneous 3D reconstruction methods, focusing specifically on the relative merits of workflows that operate on reconstructed 3D subtomogram volumes compared to those using extracted 2D particle-images. We additionally highlight how these methods have provided specific biological insights into the organization, dynamics, and structural variability of cellular components. Finally, we advocate for the development of benchmarking datasets collected in vitro and in situ to enable a more objective comparison of existent and emerging methods for particle classification and heterogeneous 3D reconstruction.
Accurately predicting protein-ligand binding affinities is a central task in rational drug design, as it directly influences hit discovery, lead optimization, and compound prioritization. Traditional approaches often suf...Accurately predicting protein-ligand binding affinities is a central task in rational drug design, as it directly influences hit discovery, lead optimization, and compound prioritization. Traditional approaches often suffer from limited accuracy, high computational cost, or dependence on heuristic scoring functions. Recent advances in machine learning (ML) have introduced new paradigms for the binding affinity prediction. In this review, we survey the latest developments in ML-based predictions of protein-ligand binding affinities across various directions, including structure-based approaches that leverage three-dimensional conformational data, ligand-based models that utilize mathematical approaches that employ topological invariants, and hybrid or alternative frameworks addressing diverse prediction scenarios. We highlight representative algorithms ranging from traditional supervised learning to deep learning architectures. Additionally, we discuss the current challenges faced in this domain. Finally, we outline emerging trends and potential future directions, which are poised to further enhance the accuracy and applicability of binding affinity prediction in drug discovery pipelines.
Target engagement (TE) assays are essential for confirming on-target activity, guiding medicinal chemistry, and linking molecular interactions to phenotypic outcomes. Despite their success in human drug discovery, their...Target engagement (TE) assays are essential for confirming on-target activity, guiding medicinal chemistry, and linking molecular interactions to phenotypic outcomes. Despite their success in human drug discovery, their application to bacterial and protozoan pathogens remains limited due to biological complexity, technical barriers, and lack of high-quality chemical tools and protein reagents. This review surveys current TE strategies and highlights emerging tools such as live-cell bioluminescence resonance energy transfer, cellular thermal shif assay, and chemoproteomics. Expanding TE in pathogen research will deepen mechanistic insights, reduce development risk, and improve the chances of delivering safer, more effective anti-infective therapies.
The unrealized goal of cryo-electron tomography (cryo-ET) is to visualize every protein within its cellular context. Such capability would enable molecular resolution mapping of three-dimensional protein topography and s...The unrealized goal of cryo-electron tomography (cryo-ET) is to visualize every protein within its cellular context. Such capability would enable molecular resolution mapping of three-dimensional protein topography and structure determination within a native context. Current technology limits the proteins identifiable within an individual tomogram to high-molecular-weight complexes. Localization of smaller target proteins requires the use of labeling systems that act as fiducial markers of target protein localization. Several labeling systems have been developed and recently employed, all of which involve trade-offs. The choice of which system to use depends on the biological question of interest. This review outlines considerations for the design and choice of labeling systems for cryo-ET, highlights recent applications, and outlines areas for future development.
Curr Opin Struct Biol
· 2025 Dec · PMID 41289964
·
Full text
Cryo-electron microscopy has become the preferred method for determining structures of macromolecular complexes both in isolation, using single particle analysis, and in their cellular contexts, using cryo-electron tomog...Cryo-electron microscopy has become the preferred method for determining structures of macromolecular complexes both in isolation, using single particle analysis, and in their cellular contexts, using cryo-electron tomography (Cryo-ET) combined with subvolume averaging (SVA). Collection of tilt series for Cryo-ET introduces challenges such as low signal-to-noise ratios, sample radiation sensitivity, and mechanical imprecision of the microscope stage - particularly at high magnifications. Strategies to improve throughput and resolution include continuous tilt and beam-image-shift parallel acquisition, real-time predictive adjustments, and machine learning-driven targeting. Additionally, montage tomography has increased the observable cellular area, while innovations like rectangular condenser apertures promise improved dose efficiency. Web-based and machine learning-enhanced solutions for automated and remote microscope operation are improving the user experience. Collectively, these advancements represent a critical step towards robust, high-resolution, and user-friendly Cryo-ET, facilitating the visualization of macromolecular assemblies within their authentic biological environments.
Molecular dynamics simulations are crucial for investigating biomolecular mechanisms. The success of these simulations hinges on the accuracy, efficiency, and generalizability of the underlying force field. While classic...Molecular dynamics simulations are crucial for investigating biomolecular mechanisms. The success of these simulations hinges on the accuracy, efficiency, and generalizability of the underlying force field. While classical molecular force fields are efficient yet approximate and quantum mechanics is accurate but computationally prohibitive for large systems, machine learning force fields (MLFFs) have emerged to bridge this gap. We review various MLFFs-from classically parametrized to end-to-end models-evaluating their performance in accuracy and efficiency. However, a significant challenge for MLFFs is generalizability as models trained on specific data often fail to extrapolate to unseen molecules or conformations. To address this, universal MLFFs, such as fragment-based methods like AIBMD designed by Wang et al. and GEMS designed by Unke et al., are being developed. Beyond recent progress, we also discuss the inherent limitations and trade-offs of MLFFs. Looking forward, the integration of MLFFs with virtual cell models and coarse-grained representations is poised to enable whole-cell multiscale simulations.
Curr Opin Struct Biol
· 2025 Dec · PMID 41232168
·
Full text
Receptor tyrosine kinases (RTKs) control myriads of cellular functions. RTKs are paradigmatic examples of receptors where activity is directly dependent on quaternary structure. In most cases, the monomeric RTK is inacti...Receptor tyrosine kinases (RTKs) control myriads of cellular functions. RTKs are paradigmatic examples of receptors where activity is directly dependent on quaternary structure. In most cases, the monomeric RTK is inactive, and function arises only after a ligand binding event leads the RTK to bind to another copy of itself, activating trans-autophosphorylation of tyrosine residues. Such RTK homodimerization can be accompanied by the formation of homomers of higher stoichiometry. However, RTK monomers can also bind to a second type of RTK, forming heterodimers. RTK heteromerization is believed to result in different signaling than homomerization. Despite its importance, we have a poor understanding of the factors that define if an RTK will form homomers or heteromers. This short review covers recent discoveries on the heteromerization of RTK, in what is called the RTK interactome. We discuss its translational potential, and how ligands and membrane lipids affect heteromer formation.
Curr Opin Struct Biol
· 2025 Dec · PMID 41232167
·
Full text
Structural variability plays a crucial role in enabling biological function, as the ability of proteins to adopt multiple conformations allows them to perform diverse cellular tasks. Cryo-electron tomography combined wit...Structural variability plays a crucial role in enabling biological function, as the ability of proteins to adopt multiple conformations allows them to perform diverse cellular tasks. Cryo-electron tomography combined with subtomogram averaging and classification has emerged as a powerful technique for elucidating the conformational dynamics of proteins in their near-native environment. Increased data availability has provided a driving force for improvements in image classification algorithms which have enabled conformational heterogeneity studies of proteins in situ at higher resolutions than previously possible. In particular, the use of 2D particle projections extracted from raw tilt-series paired with constrained classification strategies of projection sets has emerged as a promising strategy for classifying particles in 3D. Despite these efforts, further method development will be needed to extend the applicability of current strategies for 3D classification to more challenging biological targets, including low-molecular weight complexes and membrane proteins.