bims-enlima Biomed News
on Engineered living materials
Issue of 2025–10–12
thirty-six papers selected by
Rahul Kumar, Tallinna Tehnikaülikool
  1. Boolean logic-gated protein presentation through autonomously compiled molecular topology.
  2. Bacterially grown living materials with resistant and on-demand functionality.
  3. Spatiotemporal toughness modulation in hydrogels through on-demand cross-linking.
  4. Automated Strain Construction for Biosynthetic Pathway Screening in Yeast.
  5. Generalized design of sequence-ensemble-function relationships for intrinsically disordered proteins.
  6. In Situ Stiffness Tunable DNA Hydrogels Based on Ring-Opening Polymerization.
  7. Generative AI for navigating synthesizable chemical space.
  8. Reprogrammable snapping morphogenesis in ribbon-cluster meta-units using stored elastic energy.
  9. High-strength, multi-mode processable bamboo molecular bioplastic enabled by solvent-shaping regulation.
  10. citrOgen: a synthesis-free polysaccharide and protein antigen-presentation to antibody-induction platform.
  11. Metabolic interplay drives population cycles in a cross-feeding microbial community.
  12. Recovering gene regulatory networks in single-cell multiomics data with PRISM-GRN.
  13. Programmable DNA shell scaffolds for directional membrane budding.
  14. SHAPE-Based Chemical Probes for Studying preQ1-RNA Interactions in Living Bacteria.
  15. Rigidity Governs Entrainment of Bacteria Cells in Biopolymer Scaffolds.
  16. Enzyme specificity prediction using cross attention graph neural networks.
  17. Machine learning in chemistry.
  18. Spatial self-organization of confined bacterial suspensions.
  19. PEDOT:PSS Microparticles for Extrudable and Bioencapsulating Conducting Granular Hydrogel Bioelectronics.
  20. Durable epigenetic silencing with a TALE-based editor.
  21. Chemistry Nobel for scientists who developed massively porous 'super sponge' materials.
  22. Combinatorial Discovery of RAFT Cationic Polymers for mRNA Delivery: Structure-Function Insights from High-Throughput Screening and Machine Learning.
  23. Shape Switching and Tunable Oscillations of Adaptive Droplets.
  24. Programmable material via thiol-ene polymerization initiated by electric-field induced thiyl radical on piezoelectric ZnO.
  25. Transformative Mechanical and Piezoelectric Properties in Organic-Inorganic Hybrid Materials for Biomedical Applications.
  26. Enhancing Escherichia coli production of material proteins using circular mRNAs.
  27. 'Google for DNA' brings order to biology's big data.
  28. Cellulose-mediated ionic liquid crystallization enables tough-stiff switchable ionogels.
  29. Hydrolytic degradation of PEG-based thiol-norbornene hydrogels enables multi-modal controlled release.
  30. Customizable AI systems that anyone can adapt bring big opportunities - and even bigger risks.
  31. Thermochemical heterolytic hydrogenation catalysis proceeds through polarization-driven hydride transfer.
  32. Stiff and Fracture-Resistant Ion Gels Enabled by Synergetic Physical Entanglement and Hydrogen Bonding.
  33. Comprehensive genetic interaction analysis of the Bacillus subtilis envelope using double-CRISPRi.
  34. Water-Modulated Construction of Nanoclay/g-C3N4 Heterostructures for Selective Generation of Singlet Oxygen in Peroxymonosulfate Activation Processes.
  35. Transient Emulsion-Assisted Fabrication of Liquid Metal Microsphere Arrays for Mechanically Adaptive Anisotropic Conductive Packaging.
  36. Graph neural networks for predicting metal-ligand coordination of transition metal complexes.
  1. Nat Chem Biol. 2025 Oct 09.
    Boolean logic-gated protein presentation through autonomously compiled molecular topology.
    Ryan Gharios, Murial L Ross, Annabella Li, Shivani P Kottantharayil, Jack Hoye, Cole A DeForest.
      Stimulus-responsive materials have enabled advanced applications in biosensing, tissue engineering and therapeutic delivery. Although controlled molecular topology has been demonstrated as an effective route toward creating materials that respond to prespecified input combinations, prior efforts suffer from a reliance on complicated and low-yielding multistep organic syntheses that dramatically limit their utility. Harnessing the power of recombinant expression, we integrate emerging chemical biology tools to create topologically specified protein cargos that can be site-specifically tethered to and conditionally released from biomaterials following user-programmable Boolean logic. Critically, construct topology is autonomously compiled during expression through spontaneous intramolecular ligations, enabling direct and scalable synthesis of advanced operators. Using this framework, we specify protein release from biomaterials following all 17 possible YES/OR/AND logic outputs from input combinations of three orthogonal protease actuators, multiplexed delivery of three distinct biomacromolecules from hydrogels, five-input-based conditional cargo liberation and logically defined protein localization on or within living mammalian cells.
    DOI:  https://doi.org/10.1038/s41589-025-02037-5
  2. Sci Adv. 2025 Oct 10. 11(41): eadw8278
    Bacterially grown living materials with resistant and on-demand functionality.
    Jeong-Joo Oh, Franka H van der Linden, Koray Malcı, Ramon A van der Valk, Tom Ellis, Marie-Eve Aubin-Tam.
      Inspired by naturally occurring biomaterials, autonomously grown engineered living materials (ELMs) feature cell-driven growth and programmable biological functions. However, the "livingness" of cells poses a short life span and low tolerance to harsh conditions, limiting the practical use of such materials. Here, we developed materials with programmable and dormant functionalities, grown from a mixture of Komagataeibacter rhaeticus and Bacillus endospores under engineered medium conditions. K. rhaeticus produces the bacterial cellulose (BC) matrix that integrates Bacillus spores within, whereas the confined spores keep dormant and are resistant to harsh conditions in the environment. Bacillus spores can germinate and confer desired functions to the materials. Modulating the binding affinity of spores to the BC matrix with genetic engineering can improve cell loading and therefore enhance the material functionality. These materials can serve as a versatile on-demand platform for applications as biosensors, biocatalytic materials, and in situ transformation of mechanically robust cellulose-based composites.
    DOI:  https://doi.org/10.1126/sciadv.adw8278
  3. Sci Adv. 2025 Oct 10. 11(41): eadz0440
    Spatiotemporal toughness modulation in hydrogels through on-demand cross-linking.
    Jihun Lee, Rogerio M Castilho, Sungmin Nam.
      Tough hydrogels are promising for soft robotics, bioelectronics, and tissue adhesives due to their exceptional resilience and biocompatibility, yet precise spatiotemporal control of their mechanics remains challenging. Here, we present a hydrogel platform that enables spatiotemporal modulation of toughness through a latent ionic cross-linking mechanism. By embedding calcium carbonate (CaCO3) microparticles in alginate/polyacrylamide double-network hydrogels, we create a system where localized calcium release and thus ionic cross-linking can be programmed in both space and time. Spatial control is achieved by direct ink writing of CaCO3, while temporal activation is triggered by glucono-δ-lactone, a biocompatible acidifier that releases calcium on demand. This strategy allows user-defined tuning of stiffness and toughness, enabling fabrication of three-dimensional (3D) hydrogels with tailored mechanical profiles. The resulting materials offer a versatile platform for anisotropic impact shielding, directional strain sensing, and 3D-printed tissue adhesives, representing a paradigm shift for adaptive, reconfigurable, and multifunctional soft materials.
    DOI:  https://doi.org/10.1126/sciadv.adz0440
  4. ACS Synth Biol. 2025 Oct 08.
    Automated Strain Construction for Biosynthetic Pathway Screening in Yeast.
    Maria C T Astolfi, Sam D Yoder, Marina Delfa-Lalaguna, Peter H Winegar, Sara K F Holm, Mengziang Lei, Xixi Zhao, Stephen E Tan, Randy Louie, Nathan J Hillson, Graham A Hudson, Jay D Keasling.
      Automation accelerates the Design-Build-Test-Learn (DBTL) cycle for synthetic biology; however, most strain construction pipelines lack robotic integration. Here, we present the workflow design and source code for a modular, integrated protocol that automates the Build step in Saccharomyces cerevisiae. We programmed the Hamilton Microlab VANTAGE to integrate off-deck hardware via its central robotic arm, enabling automated steps that increased throughput to 2,000 transformations per week. We developed a user interface with the Hamilton VENUS software to support on-demand parameter customization. As a proof of concept, we screened a gene library in an engineered yeast strain producing verazine, a key intermediate in the biosynthesis of steroidal alkaloids. Our pipeline rapidly identified pathway bottlenecks and genes that enhanced verazine production by 2.0- to 5-fold. This technical note provides resources for synthetic biologists designing yeast workflows for biofoundries to screen libraries for pathway discovery/optimization, combinatorial biosynthesis, and protein engineering.
    Keywords:  automation; high-throughput screen; metabolic engineering; microbial biosynthesis; robotics integration; strain engineering
    DOI:  https://doi.org/10.1021/acssynbio.5c00554
  5. Nat Comput Sci. 2025 Oct 06.
    Generalized design of sequence-ensemble-function relationships for intrinsically disordered proteins.
    Ryan K Krueger, Michael P Brenner, Krishna Shrinivas.
      The design of folded proteins has advanced substantially in recent years. However, many proteins and protein regions are intrinsically disordered and lack a stable fold, that is, the sequence of an intrinsically disordered protein (IDP) encodes a vast ensemble of spatial conformations that specify its biological function. This conformational plasticity and heterogeneity makes IDP design challenging. Here we introduce a computational framework for de novo design of IDPs through rational and efficient inversion of molecular simulations that approximate the underlying sequence-ensemble relationship. We highlight the versatility of this approach by designing IDPs with diverse properties and arbitrary sequence constraints. These include IDPs with target ensemble dimensions, loops and linkers, highly sensitive sensors of physicochemical stimuli, and binders to target disordered substrates with distinct conformational biases. Overall, our method provides a general framework for designing sequence-ensemble-function relationships of biological macromolecules.
    DOI:  https://doi.org/10.1038/s43588-025-00881-y
  6. Small Methods. 2025 Oct 05. e01478
    In Situ Stiffness Tunable DNA Hydrogels Based on Ring-Opening Polymerization.
    Ziwei Shi, Jiarui Li, Miaomiao Qiu, Lianqiang Dong, Dongsheng Liu, Lijin Xu, Yuanchen Dong.
      DNA hydrogels are promising artificial extracellular matrices (ECMs) due to their programmability and biocompatibility. However, most current stiffness modulation strategies are static, with limited dynamic regulation due to the restricted responsiveness of the building blocks. Here, a ring-opening polymerization strategy is presented based on supramolecular dimer rings containing functional domains to achieve in situ regulation of DNA hydrogel stiffness. The rings consist of complementary regions, flexible spacers, and sticky ends. Upon the addition of linkers, the rings polymerize into linear polymers that form a hydrogel through physical entanglement. Hybridization with trigger strands induces ring-opening, leading to network remodeling and enhanced stiffness, while strand displacement enables reversible stiffness reduction. This approach allows dynamic and programmable mechanical regulation under physiological conditions, providing a biomimetic platform to mimic dynamic ECM stiffening.
    Keywords:  3D cell culture; DNA nanotechnology; DNA supramolecular hydrogel; ring‐opening polymerization; stiffness regulation
    DOI:  https://doi.org/10.1002/smtd.202501478
  7. Proc Natl Acad Sci U S A. 2025 Oct 14. 122(41): e2415665122
    Generative AI for navigating synthesizable chemical space.
    Wenhao Gao, Shitong Luo, Connor W Coley.
      We introduce SynFormer, a generative modeling framework designed to efficiently explore and navigate synthesizable chemical space. Unlike traditional molecular generation approaches, we generate synthetic pathways for molecules to ensure that designs are synthetically tractable. By incorporating a scalable transformer architecture and a diffusion module for building block selection, SynFormer surpasses existing models in synthesizable molecular design. We demonstrate SynFormer's effectiveness in two key applications: 1) local chemical space exploration, where the model generates synthesizable analogs of a query molecule, and 2) global chemical space exploration, where the model aims to identify optimal molecules according to a black-box property prediction oracle. Additionally, we demonstrate the scalability of our approach via the improvement in performance as more computational resources become available. With our code and trained models openly available, we hope that SynFormer will find use across applications in drug discovery and materials science.
    Keywords:  deep learning; generative AI; molecular design; synthetic accessibility
    DOI:  https://doi.org/10.1073/pnas.2415665122
  8. Nat Mater. 2025 Oct 10.
    Reprogrammable snapping morphogenesis in ribbon-cluster meta-units using stored elastic energy.
    Yaoye Hong, Caizhi Zhou, Haitao Qing, Yinding Chi, Jie Yin.
      Snapping, driven by stored elastic energy, enables versatile and rapid shape changes in nature; yet replicating such autonomous, reprogrammable morphogenesis in free-standing volumetric structures remains elusive. Here we report a lantern-shaped ribbon-cluster meta-unit that harnesses programmable and reprogrammable elastic energy to achieve over 13 distinct volumetric snapping morphologies from a single unit. Governed by three Euler angles, the meta-unit post-fabrication offers a tunable mechanical design space spanning up to quadrastable states. Unlike single-ribbon or mechanism-based designs, our system autonomously selects snapping pathways via nastic coupling between multiple ribbons, enabling the inverse design of complex snapping morphologies. We harness magnetically actuated bud-to-bloom and tristable morphogenesis to enable fast, non-invasive grasping and remote flow regulation in confined environments. These results establish a general framework for architected materials with programmable shape, stability and function, offering potential applications in soft robotics, deployable devices and mechanical logic.
    DOI:  https://doi.org/10.1038/s41563-025-02370-z
  9. Nat Commun. 2025 Oct 07. 16(1): 8729
    High-strength, multi-mode processable bamboo molecular bioplastic enabled by solvent-shaping regulation.
    Hongying Tang, Zhihan Tong, Rui Zhang, Xiaona Li, Suqing Zeng, Dawei Zhao, Haipeng Yu.
      The global reliance on petrochemical plastics has led to severe environmental crises, necessitating sustainable alternatives that combine high performance with circularity. While bioplastics derived from biomass show promise, their widespread adoption is hindered by inferior mechanical properties, limited processability, and reliance on food-competing feedstocks. Here, we present a molecular engineering strategy to fabricate high-strength bamboo molecular plastics (BM-plastics) through a solvent-regulated shaping process. By employing deep eutectic solvents to disassemble bamboo cellulose's hydrogen-bond network and ethanol-mediated molecular stimulation to reconstruct dense hydrogen-bond interactions, we achieve a bioplastic with exceptional mechanical strength (tensile strength: 110 MPa, flexural modulus: 6.41 GPa), thermal stability (>180 °C), and versatile processability via injection, molding, and machining techniques. The BM-plastic outperforms most commercial plastics and bioplastics in mechanical and thermo-mechanical metrics while maintaining full biodegradability in soil within 50 days and closed-loop recyclability with 90% retained strength. Techno-economic analysis confirms its cost competitiveness, bridging the gap between sustainability and industrial scalability. This work establishes a method for transforming abundant bamboo cellulose into high-performance, eco-friendly materials, offering a viable pathway to mitigate plastic pollution and fossil resource dependence.
    DOI:  https://doi.org/10.1038/s41467-025-63904-2
  10. Nat Commun. 2025 Oct 06. 16(1): 8886
    citrOgen: a synthesis-free polysaccharide and protein antigen-presentation to antibody-induction platform.
    Joshua L C Wong, Julia Sanchez-Garrido, Jaie Rattle, Jonathan Bradshaw, Vishwas Mishra, Gad Frankel.
      Existing technologies employed to generate antibodies against bacterial polysaccharides and proteins rely on the availability of purified or synthetic antigens. Here, we present a genetics-based platform that utilises Citrobacter rodentium (CR), an enteric mouse pathogen, to both produce and present complex heterologous polysaccharides and protein antigen complexes during natural infection. As proof of concept, we use lipopolysaccharides (O), capsular polysaccharides (K) and type 3 fimbrial (T3F) antigens expressed by the WHO critical priority pathogens Klebsiella pneumoniae (KP) and Escherichia coli (EC). Following one infection cycle (28 days), CR induces specific IgG antibodies against KPO1, ECO25b, KPK2 and KPT3F. We demonstrate that the antibodies are functional in downstream applications, including protection against pathogenic KP challenge, KP capsular serotyping and KP biofilm inhibition. Whilst KP and EC antigens were used as prototypical examples, this modular platform is now readily adaptable to generate antibodies against diverse polysaccharide and protein antigens, with basic science, public health and therapeutic applications.
    DOI:  https://doi.org/10.1038/s41467-025-63922-0
  11. Nat Commun. 2025 Oct 07. 16(1): 8919
    Metabolic interplay drives population cycles in a cross-feeding microbial community.
    Tyler D Ross, Hanhyeok Im, Brennan G Keogh, Christopher A Klausmeier, Ophelia S Venturelli.
      Population cycles are prevalent in ecosystems and play key roles in determining their functions. While multiple mechanisms have been theoretically shown to generate population cycles, there are limited examples of mutualisms driving self-sustained oscillations. Using an engineered microbial community that cross-feeds essential amino acids, we experimentally demonstrate cycles in strain abundance that are robust across environmental conditions. A nonlinear dynamical model that incorporates the experimentally observed cross-inhibition of amino acid production recapitulates the population cycles. The model shows that the cycles represent internally generated relaxation oscillations, which emerge when fast resource dynamics with positive feedback drive slow changes in strain abundance. The temporal structure of the resource dynamics prevents nonproducing cheaters from persisting within the oscillating community. Our findings highlight the critical role of resource dynamics and feedback in shaping population cycles in microbial communities and have implications for biotechnology.
    DOI:  https://doi.org/10.1038/s41467-025-63986-y
  12. Genome Res. 2025 Oct 09. pii: gr.280757.125. [Epub ahead of print]
    Recovering gene regulatory networks in single-cell multiomics data with PRISM-GRN.
    Wenhao Zhang, Lan Cao, Xiaoxuan Gu, Yongyu Long, Ying Wang.
      Understanding Gene Regulatory Networks (GRNs) is crucial for deciphering cellular heterogeneity and the mechanisms underlying development and disease. However, current GRN inference methods fail to utilize multiomics data and prior knowledge from a biologically-interpretable insight. Therefore, we propose PRISM-GRN, a Bayesian model that seamlessly incorporates known GRNs, along with scRNA-seq and scATAC-seq data, into a probabilistic framework to reconstruct cell type-specific GRNs. PRISM-GRN employs a biologically interpretable architecture firmly rooted in the established gene regulatory mechanism, which asserts that gene expression is influenced by TF expression levels and gene chromatin accessibility through GRNs. Accordingly, PRISM-GRN decomposes observable data into biologically meaningful latent variables through a mechanism-informed generation process and a prior-GRN-primed inference process, enabling precise and robust GRN reconstruction. We evaluate PRISM-GRN on four benchmarking datasets with paired scRNA-seq and scATAC-seq data, demonstrating its superior performance over seven baseline methods in GRN reconstruction, especially its higher precision under the inherently imbalanced scenario where the true regulatory interaction is sparse. Furthermore, benchmarking on directed GRNs highlights PRISM-GRN's ability to capture causality in gene regulation derived from the biologically-interpretable architecture. More importantly, PRISM-GRN performs well with unpaired omics data and limited prior GRN information, showcasing its flexibility and adaptability across various biological contexts. Finally, biological analyses on PBMC datasets demonstrate PRISM-GRN's potential to facilitate the identification of cell type-specific or context-specific GRNs across broader real-world biological research applications. Overall, PRISM-GRN provides a novel paradigm for precise, robust, and interpretable exploration of causal GRNs with prior knowledge and multiomics data.
    DOI:  https://doi.org/10.1101/gr.280757.125
  13. Nat Commun. 2025 Oct 09. 16(1): 8972
    Programmable DNA shell scaffolds for directional membrane budding.
    Michael T Pinner, Hendrik Dietz.
      In the pursuit of replicating biological processes at the nanoscale, controlling cellular membrane dynamics has emerged as a key area of interest. Here, we report a system mimicking virus assembly to control directional membrane budding. We employ three-dimensional DNA origami techniques to construct cholesterol-modified triangles that self-assemble into polyhedral shells on lipid vesicles, resulting in gradual curvature induction, bud formation, and spontaneous neck scission. Strategic positioning of cholesterols on the triangle surface provides control over the directionality of bud growth and yields daughter vesicles with DNA endo- or exoskeletons reminiscent of clathrin-coated vesicles. This process occurs with rapid kinetics and across various lipid compositions. When combined into a two-step process, nested bivesicular objects with DNA shells encapsulated between lipid vesicles could be produced. Our work replicates key aspects of natural endocytic and exocytic pathways, opening new avenues for exploring membrane mechanics and applications in targeted drug delivery and synthetic biology.
    DOI:  https://doi.org/10.1038/s41467-025-64298-x
  14. ACS Chem Biol. 2025 Oct 10.
    SHAPE-Based Chemical Probes for Studying preQ1-RNA Interactions in Living Bacteria.
    José A Reyes Franceschi, Emilio L Cárdenas, Brandon J C Klein, Chase A Weidmann, Amanda L Garner.
      Interrogating RNA-small molecule interactions inside cells is critical for advancing RNA-targeted drug discovery. In particular, chemical probing technologies that both identify small molecule-bound RNAs and define their binding sites in the complex cellular environment will be key to establishing the on-target activity necessary for successful hit-to-lead campaigns. Using the small molecule metabolite preQ1 and its cognate riboswitch RNA as a model, herein we describe a chemical probing strategy for filling this technological gap. Building on well-established RNA acylation chemistry employed by in vivo click-selective 2'-hydroxyl acylation analyzed by primer extension (icSHAPE) probes, we developed an icSHAPE-based preQ1 probe that retains biological activity in a preQ1 riboswitch reporter assay and successfully enriches the preQ1 riboswitch from living bacterial cells. Further, we map the preQ1 binding site on probe-modified riboswitch RNA by mutational profiling (MaP). As the need for rapid profiling of on- and off-target small molecule interactions continues to grow, this chemical probing strategy offers a method to interrogate cellular RNA-small molecule interactions and supports the future development of RNA-targeted therapeutics.
    DOI:  https://doi.org/10.1021/acschembio.5c00548
  15. ACS Biomater Sci Eng. 2025 Oct 06.
    Rigidity Governs Entrainment of Bacteria Cells in Biopolymer Scaffolds.
    Katarina Matic, Nimisha Krishnan, Moumita Das, Michael J Rust, Megan T Valentine, Jennifer L Ross, Rae M Robertson-Anderson.
      Embedding cells in biopolymer networks is a promising route for designing responsive, dynamic materials for applications such as tissue engineering. The ability of cells to tune material properties relies on the physical properties of the biopolymers and their interactions with each other and the embedded cells. Here, we investigate how biopolymer stiffness and cross-linking influences cell entrainment within the network, and how the network is restructured by the embedded cells. Specifically, we design composites of Escherichia coli cells and cytoskeleton filaments, either semiflexible actin filaments or rigid microtubules; and characterize the effect of cell volume fraction and filament cross-linking on the structure and interactions of the components. We find that entropically driven depletion interactions between cells and filaments lead to emergent filament bundling at intermediate cell densities, which is more pronounced for actin compared to microtubules. Moreover, cross-linking of actin causes enhanced clustering of cells and colocalization of cells and filaments compared to the entangled case. Conversely, cross-linking of microtubules modestly suppresses cell-induced restructuring and colocalization compared to the entangled case. Our work highlights the ability of biopolymer stiffness and connectivity to selectively tune composite properties and cell entrainment for use in diverse applications from tissue engineering to biocement.
    Keywords:  actin; cells; composites; confocal fluorescence microscopy; microtubules; networks
    DOI:  https://doi.org/10.1021/acsbiomaterials.5c01302
  16. Nature. 2025 Oct 08.
    Enzyme specificity prediction using cross attention graph neural networks.
    Haiyang Cui, Yufeng Su, Tanner J Dean, Tianhao Yu, Zhengyi Zhang, Jian Peng, Diwakar Shukla, Huimin Zhao.
      Enzymes are the molecular machines of life, and a key property that governs their function is substrate specificity-the ability of an enzyme to recognize and selectively act on particular substrates. This specificity originates from the three-dimensional (3D) structure of the enzyme active site and complicated transition state of the reaction1,2. Many enzymes can promiscuously catalyze reactions or act on substrates beyond those for which they were originally evolved1,3-5. However, millions of known enzymes still lack reliable substrate specificity information, impeding their practical applications and comprehensive understanding of the biocatalytic diversity in nature. Herein, we developed a cross-attention-empowered SE(3)-equivariant graph neural network architecture named EZSpecificity for predicting enzyme substrate specificity, which was trained on a comprehensive tailor-made database of enzyme-substrate interactions at sequence and structural levels. EZSpecificity outperformed the existing machine learning models for enzyme substrate specificity prediction, as demonstrated by both an unknown substrate and enzyme database and seven proof-of-concept protein families. Experimental validation with eight halogenases and 78 substrates revealed that EZSpecificity achieved a 91.7% accuracy in identifying the single potential reactive substrate, significantly higher than that of the state-of-the-art model ESP (58.3%). EZSpecificity represents a general machine learning model for accurate prediction of substrate specificity for enzymes related to fundamental and applied research in biology and medicine.
    DOI:  https://doi.org/10.1038/s41586-025-09697-2
  17. Proc Natl Acad Sci U S A. 2025 Oct 14. 122(41): e2517062122
    Machine learning in chemistry.
    Pablo G Debenedetti, Juan J de Pablo, George C Schatz.
      
    DOI:  https://doi.org/10.1073/pnas.2517062122
  18. Proc Natl Acad Sci U S A. 2025 Oct 14. 122(41): e2503983122
    Spatial self-organization of confined bacterial suspensions.
    Babak Vajdi Hokmabad, Alejandro Martínez-Calvo, Sebastian Gonzalez La Corte, Sujit S Datta.
      Lab studies of bacteria usually focus on cells in spatially extended, nutrient-replete settings, such as in liquid cultures and on agar surfaces. By contrast, many biological and environmental settings-ranging from mucus in the body to ocean sediments and the soil beneath our feet-feature multicellular bacterial populations that are confined to tight spots where essential metabolic substrates (e.g., oxygen) are scarce. What influence does such confinement have on a bacterial population? Here, we address this question by studying suspensions of motile Escherichia coli confined to quasi two-dimensional (2D) droplets. We find that when the droplet size and cell concentration are both large enough, the initially uniform suspension spatially self-organizes into a concentrated, immotile inner "core" that coexists with a more dilute, highly motile surrounding "shell." By simultaneously measuring cell concentration, oxygen concentration, and motility-generated fluid flow, we show that this behavior arises from the interplay between oxygen transport through the droplet from its boundary, uptake by the cells, and corresponding changes in their motility in response to oxygen variations. Furthermore, we use biophysical theory and simulations to quantitatively describe this interplay. Our work thus sheds light on the rich collective behaviors that emerge for bacterial populations in confined environments, with implications for understanding ecological niches and engineering artificial systems.
    Keywords:  active matter; bacteria; pattern formation; reaction–diffusion; self-organization
    DOI:  https://doi.org/10.1073/pnas.2503983122
  19. Small. 2025 Oct 08. e06438
    PEDOT:PSS Microparticles for Extrudable and Bioencapsulating Conducting Granular Hydrogel Bioelectronics.
    Anna P Goestenkors, Justin S Yu, Jae Park, Yuqing Wu, Cielo J Vargas Espinoza, Lianna C Friedman, Somtochukwu S Okafor, Tianran Liu, Suman Chatterjee, Avishek Debnath, Barbara A Semar, Cayleigh P O'Hare, Riley M Alvarez, Srikanth Singamaneni, Baranidharan Raman, Alexandra L Rutz.
      Conducting hydrogels are promising materials for forming biomimetic bioelectronic interfaces to monitor and stimulate biological activity. However, most developed materials possess fixed shapes, which limit their application to specific types of device interfaces. In non-conducting biomaterials, granular hydrogels have enabled encapsulating, conformal, and injectable biointerfaces due to unique possession of both microporosity and dynamic mechanical properties. Bringing this adaptability to conducting hydrogels would be promising for enhancing bioelectronic interfaces. However, granular hydrogels remain largely unexplored as conducting biomaterials. Methods for fabricating spherical hydrogel microparticles from the conducting polymer poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) are presented. When densely packed, these microparticles form a conducting granular hydrogel with microporosity as well as shear-thinning and self-healing dynamic mechanical properties. The PEDOT:PSS granular hydrogel can be extruded and maintain structure post-3D printing. Modulating microparticle PSS content achieves high granular hydrogel conductivity (137 S m-1), and microparticles exhibit excellent cytocompatibility (>98% viability). Finally, utility is demonstrated as bioencapsulating electrodes for electrophysiological monitoring. These results highlight the functionality of the PEDOT:PSS conducting granular hydrogel, which in the future may be realized as 3D printed bioencapsulating electrodes, 3D tissue engineering scaffolds for monitoring encapsulated cells, and injectable therapies for enhanced cell recruitment and tissue regeneration combined with electronic stimulation.
    Keywords:  PEDOT:PSS; bioelectronic devices; conducting granular hydrogels; conducting microparticles; electrophysiological monitoring
    DOI:  https://doi.org/10.1002/smll.202506438
  20. Nat Biotechnol. 2025 Oct 08.
    Durable epigenetic silencing with a TALE-based editor.
    Tassilo L A Wachsmann, Lei S Qi.
      
    DOI:  https://doi.org/10.1038/s41587-025-02871-x
  21. Nature. 2025 Oct 08.
    Chemistry Nobel for scientists who developed massively porous 'super sponge' materials.
    Davide Castelvecchi, Miryam Naddaf.
      
    Keywords:  Chemistry
    DOI:  https://doi.org/10.1038/d41586-025-03195-1
  22. Biomacromolecules. 2025 Oct 06.
    Combinatorial Discovery of RAFT Cationic Polymers for mRNA Delivery: Structure-Function Insights from High-Throughput Screening and Machine Learning.
    Wenting Yang, Shangqian Li, Siqi Yu, Alex G C de Sá, Tanmayee Sai Sivani Ita, Tian Liang, Helen Forgham, Ruirui Qiao, Jiulong Li, Patrick S Stayton, David B Ascher, Huan Meng, Andrew K Whittaker, Changkui Fu.
      Cationic polymers are considered promising delivery systems for mRNA, offering potential advantages over lipid nanoparticles. Here a library of tertiary amine-containing, methacrylate-based cationic polymers with diverse molecular characteristics and properties were prepared by combinatorial RAFT polymerization for mRNA delivery. The ability of the synthesized cationic polymers to complex with mRNA was thoroughly investigated. The biological responses, including cellular uptake, cytotoxicity, and mRNA transfection efficiency, of the formed mRNA-polymer polyplexes were systemically investigated. Through high-throughput screening assays, we identified several lead polymers that showed superior effectiveness in delivering mRNA, with performance significantly outperforming other synthesized cationic polymers as well as polyethylenimine (PEI) and Lipofectamine, two benchmark gene delivery materials. To unravel the complex structure-function relationships between the chemical and physical properties of cationic polymers/mRNA polyplexes and their biological responses, machine learning analyses were conducted. These in silico studies identified several key attributes that are predictive of cellular uptake, cytotoxicity, and mRNA transfection efficiency, providing valuable insights for the future design of more potent cationic polymers for mRNA delivery.
    DOI:  https://doi.org/10.1021/acs.biomac.5c01236
  23. Phys Rev Lett. 2025 Sep 19. 135(12): 128404
    Shape Switching and Tunable Oscillations of Adaptive Droplets.
    Tim Dullweber, Roman Belousov, Camilla Autorino, Nicoletta I Petridou, Anna Erzberger.
      Living materials adapt their shape to signals from the environment, yet the impact of shape changes on signal processing and associated feedback dynamics remains unclear. We find that droplets with signal-responsive interfacial tensions exhibit shape bistability, excitable dynamics, and oscillations. The underlying critical points reveal novel mechanisms for physical signal processing through shape adaptation in soft active materials. We recover signatures of one such critical point in experimental data from zebrafish embryos, where it supports boundary formation.
    DOI:  https://doi.org/10.1103/1cq4-x499
  24. Nat Commun. 2025 Oct 09. 16(1): 8987
    Programmable material via thiol-ene polymerization initiated by electric-field induced thiyl radical on piezoelectric ZnO.
    Jun Wang, Zhao Wang, Jorge Ayarza, Ian Frankel, Chao-Wei Huang, Kai Qian, Yixiao Dong, Pin-Ruei Huang, Katie Kloska, Chao Zhang, Siqi Zou, Matthew Mason, Chong Liu, Nicholas Boechler, Aaron P Esser-Kahn.
      The spatial and temporal control of material properties at a distance has yielded many unique innovations including photo-patterning, 3D-printing, and architected material design. To date, most of these innovations have relied on light, heat, sound, or electric current as stimuli for controlling the material properties. Here, we demonstrate that an electric field can induce chemical reactions and subsequent polymerization in composites via piezoelectrically-mediated transduction. The response to an electric field rather than through direct contact with an electrode is mediated by a nanoparticle transducer, i.e., piezoelectric ZnO, which mediates reactions between thiol and alkene monomers, resulting in tunable moduli as a function of voltage, time, and the frequency of the applied AC power. The reactivity of the mixture and the modulus of a naïve material containing these elements can be programmed based on the distribution of the electric field strength. This programmability results in multi-stiffness gels. Additionally, the system can be adjusted for the formation of an electro-adhesive. This simple and generalizable design opens avenues for facile application in adaptive damping and variable-rigidity materials, adhesive, soft robotics, and potentially tissue engineering.
    DOI:  https://doi.org/10.1038/s41467-025-64011-y
  25. ACS Nano. 2025 Oct 06.
    Transformative Mechanical and Piezoelectric Properties in Organic-Inorganic Hybrid Materials for Biomedical Applications.
    Jaewon Cho, Cheol Hyun Park, Priyannth Ramasami Sundharbaabu, Seong-Uk Ju, Davin Shin, Jun Hyuk Heo, Miso Kim, Jung Heon Lee.
      Organic-inorganic hybrid materials are multifunctional composites that combine distinct organic and inorganic components to yield synergistic enhancements in properties such as mechanical strength and piezoelectric performance, resulting in performance that exceeds that of the individual materials. This review begins by examining organic-inorganic hybrid structures found in nature, particularly within biological systems, and highlights the historical development of synthetic hybrids inspired by these natural designs. These materials are categorized based on their dimensional architectures, ranging from zero-dimensional (0D) to three-dimensional (3D) systems. A focused discussion follows on the mechanisms through which the organic and inorganic phases interact to influence the mechanical and piezoelectric properties, with a particular emphasis on interfacial interactions, structural hierarchy, and functional tunability. The biomedical applications of these hybrid materials are then summarized, detailing their roles in biomechanical energy harvesting, sensing, and regenerative medicine. This review concludes by addressing emerging challenges and outlining future directions for the rational design of organic-inorganic hybrid materials to advance biomedical technologies.
    Keywords:  DNApatite; bioinspired materials; biomedical applications; electromechanical coupling; energy harvesting; hierarchical structures; hybrid composite; piezoelectric
    DOI:  https://doi.org/10.1021/acsnano.5c07220
  26. Appl Environ Microbiol. 2025 Oct 07. e0157925
    Enhancing Escherichia coli production of material proteins using circular mRNAs.
    Alden Filko, Fuzhong Zhang.
      Industrial bioproduction of proteins, particularly protein-based materials (PBMs) like spider silk and elastin proteins, is rapidly expanding. PBMs often have high molecular weights and are highly repetitive, transcribed from long and repetitive mRNAs that are prone to degradation in microbial hosts. As a result, recombinant expression of PBMs often has low protein yields. In this study, we engineered a circular mRNA expression system to enhance mRNA stability and protein expression. The system uses self-cleaving ribozymes to form circular mRNA structures and a pair of insulation RNA loops to improve protein translation. When tested using a green fluorescent protein (GFP) reporter, the engineered circular mRNA enhanced GFP expression by 1.5-fold compared to expression from a linear construct. mRNA circularization was further confirmed using reverse transcription followed by DNA amplification and sequencing. We also demonstrate the effectiveness of circular mRNA in enhancing the expression of various material proteins, including a 96-mer repeat of Nephila clavipes dragline silk, a titin repeat, a mussel foot protein oligomer, and an silk-amyloid repeat, resulting in up to 2.5-fold increase in protein yield. Additionally, the circular mRNA system also improved the stability of the PBM-encoding plasmid. Overall, the circular RNA expression system enhances both the expression level and plasmid stability and is suitable for various protein production applications.IMPORTANCEIndustrial bioproduction of complex proteins is limited by unstable expression. Long and repetitive proteins have unstable expression and often yield truncated products that will change the properties of the final materials. We show that by using a self-circularizing mRNA system, the expression is stabilized to not only increase yields but also prevent truncated products. The ability to produce full-length proteins consistently and control their size offers precise control over protein properties, making it highly relevant for products with specific mechanical properties. The study showcases the potential for scaling up protein production in industrial bioreactors under challenging conditions. The findings contribute to synthetic biology tools and offer new avenues for manufacturing bioproducts at an industrial scale.
    Keywords:  biomaterials; circular mRNA; protein expression
    DOI:  https://doi.org/10.1128/aem.01579-25
  27. Nature. 2025 Oct 08.
    'Google for DNA' brings order to biology's big data.
    Elie Dolgin.
      
    Keywords:  Bioinformatics; DNA sequencing; Databases
    DOI:  https://doi.org/10.1038/d41586-025-03219-w
  28. Nat Commun. 2025 Oct 09. 16(1): 9007
    Cellulose-mediated ionic liquid crystallization enables tough-stiff switchable ionogels.
    Siheng Wang, Huayu Liu, Zhengyang Yu, Xinle Ren, Qi Hua, Mahyar Panahi-Sarmad, Pu Yang, Chuhang Liu, Scott Renneckar, He Liu, Feng Jiang.
      Nature has inspired to fabricate mechanically switchable materials for applications in various aspects, which is, however, unique but challenging to achieve reversible phase transitions using common ionic liquids in ionogels with ambient temperature-triggered crystallization feature. Here, we develop a tough-stiff switchable ionogel through a reversible solvent crystallization design. Cellulose acts as a chemical regulator, competitively binding with polymers to promote the formation of ionic liquid crystals. This results in a tough ionogel with a bulk toughness of 25.7 MJ m-3 and a fracture toughness of 47.1 kJ m-2, which can switch into a stiff ionogel with a tensile modulus of 134.6 MPa and a compressive modulus of 48.9 MPa. Upon heating, the crystallized ionogel reverts to its unconfined as ionic liquid crystals melt. This phase-driven structural and rigidity transition enables dynamical programming, with rapid, reversible and repeatable shape recovery through heating. Our study demonstrates solvent crystallization in ionogels, offering a strategy for creating intelligent, reconfigurable, and performance-switchable materials with customizable functions.
    DOI:  https://doi.org/10.1038/s41467-025-64061-2
  29. J Mater Chem B. 2025 Oct 06.
    Hydrolytic degradation of PEG-based thiol-norbornene hydrogels enables multi-modal controlled release.
    Nathan H Dimmitt, Chien-Chi Lin.
      Poly(ethylene glycol) (PEG) hydrogels crosslinked by orthogonal thiol-norbornene click chemistry have emerged as an ideal platform for tissue engineering and drug delivery applications due to their rapid crosslinking kinetics and excellent biocompatibility. Norbornene-functionalized PEG (PEGNB) is routinely synthesized through the Steglich esterification of 5-norbornene-2-carboxylic acid with hydroxyl-terminated PEG. When crosslinked with thiol-bearing macromers, PEGNB hydrogels undergo slow hydrolytic degradation due to hydrolysis of ester bonds connecting a PEG backbone and a NB moiety. In prior work, we replaced the pungent and nauseous 5-norbornene-2-carboxylic acid with odorless carbic anhydride (CA) for synthesizing PEG-norbornene-carboxylate (PEGNBCA), a new macromer that could be readily photo-crosslinked into thiol-norbornene hydrogels with faster hydrolytic degradation than the PEGNB counterparts. In this contribution, we employed a modular approach to tune the hydrolytic degradation of PEGNBCA hydrogels over days to months. We first demonstrated the diverse crosslinking of PEGNBCA hydrogels using either photopolymerization or enzymatic crosslinking. We characterized the hydrolytic degradation of these hydrogels under different solution pH values and temperatures. Via adjusting crosslinker functionality and the ratio of fast-degrading PEGNBCA to slow-degrading PEGNB, tunable hydrolytic degradation of PEGNBCA hydrogels was achieved from under 2 days to over 3 months. Finally, we designed the highly tunable PEGNBCA hydrogels with varying mesh sizes, degradation rates, and covalent tethering of degradable linkers to afford long-term controlled release of model drugs.
    DOI:  https://doi.org/10.1039/d5tb01524c
  30. Nature. 2025 Oct;646(8084): 286-287
    Customizable AI systems that anyone can adapt bring big opportunities - and even bigger risks.
    Yarin Gal, Stephen Casper.
      
    Keywords:  Computer science; Machine learning; Policy
    DOI:  https://doi.org/10.1038/d41586-025-03228-9
  31. Nat Chem. 2025 Oct 09.
    Thermochemical heterolytic hydrogenation catalysis proceeds through polarization-driven hydride transfer.
    Hai-Xu Wang, Yogesh Surendranath.
      Heterolytic hydrogenations, which split H2 across a hydride acceptor and proton acceptor, comprise a key reaction class that spans the chemical value chain, including CO2 hydrogenation to formate and NADH regeneration from nicotinamide adenine dinucleotide (NAD+). The dominant mechanistic models for heterogeneous catalysis of these reactions invoke classical surface reaction steps, largely ignoring the role of interfacial charge separation. Here we quantify the electrochemical potential of the catalyst during turnover and uncover evidence supporting an interfacial electrochemical hydride transfer mechanism for this overall thermochemical reaction class. We find that the proton acceptor induces spontaneous electrochemical polarization of the metal catalyst surface, thereby controlling the thermodynamic hydricity of the surface M-H intermediates and driving rate-determining electrochemical hydride transfer to the hydride acceptor substrate. This mechanistic framework, which applies across diverse reaction media and for the hydrogenation of CO2 to formate and NAD+ to NADH, enables the determination of intrinsic reaction kinetics and exposes design principles for the future development of sustainable hydrogenation reactivity.
    DOI:  https://doi.org/10.1038/s41557-025-01939-0
  32. Small. 2025 Oct 07. e09922
    Stiff and Fracture-Resistant Ion Gels Enabled by Synergetic Physical Entanglement and Hydrogen Bonding.
    Ryota Tamate, Yuji Kamiyama, Ken Kojio.
      In this study, ion gels are developed that simultaneously exhibit exceptional stiffness and fracture resistance through the synergistic effects of physical entanglements and hydrogen bonding between polymer chains within an ionic liquid matrix. Through radical copolymerization conducted in an ionic liquid under extremely low initiator concentrations, ultrahigh molecular weight polymers in situ with nearly complete monomer conversion are successfully synthesized. This strategy enabled the one-pot synthesis of physically crosslinked polymer gels composed of abundant entanglements and hydrogen bonds between polymer chains. Notably, it is demonstrated that the synergy between physical entanglements arising from ultrahigh molecular weight polymer chains and noncovalent hydrogen bonding enables the simultaneous enhancement of mechanical properties that typically exhibit trade-off relationships, such as stiffness, toughness, and fracture resistance. Consequently, the synthesized ion gels exhibited outstanding mechanical performances, ranking among the best previously reported tough polymer gels, while maintaining a favorable balance between ionic conductivity and mechanical strength. These findings underscore the broader significance of the approach, indicating that the integration of physical entanglements and reversible interactions offers a generalized pathway to mechanically robust materials across various polymer systems.
    Keywords:  entanglements; fracture resistance; hydrogen bonds; ion gels; ultrahigh molecular weight polymers
    DOI:  https://doi.org/10.1002/smll.202509922
  33. Cell Syst. 2025 Oct 03. pii: S2405-4712(25)00239-X. [Epub ahead of print] 101406
    Comprehensive genetic interaction analysis of the Bacillus subtilis envelope using double-CRISPRi.
    Byoung-Mo Koo, Horia Todor, Jiawei Sun, Jordi van Gestel, John S Hawkins, Cameron C Hearne, Amy B Banta, Kerwyn Casey Huang, Jason M Peters, Carol A Gross.
      Understanding bacterial gene function remains a major challenge. Double-mutant genetic interaction analysis addresses this challenge by uncovering the functional partners of targeted genes, enabling association of genes of unknown function with known pathways and unraveling of connections among well-studied pathways, but such approaches are difficult to implement at the genome scale. Here, we use double-CRISPR interference (CRISPRi) to systematically quantify genetic interactions at scale for the Bacillus subtilis cell envelope, including essential genes. We discover >1,000 genetic interactions, some known and others novel. Our analysis pipeline and experimental follow-ups reveal the shared and distinct roles of paralogous genes such as mreB and mbl in peptidoglycan and teichoic acid synthesis and identify additional genes involved in the well-studied process of cell division. Overall, our study provides valuable insights into gene function and demonstrates the utility of double-CRISPRi for high-throughput dissection of bacterial gene networks, providing a blueprint for future studies in diverse species. A record of this paper's transparent peer review process is included in the supplemental information.
    Keywords:  Mbl; MreB; cell division; cell envelope; cell shape; functional genomics; genetic interaction mapping; peptidoglycan biosynthesis; synthetic lethality; teichoic acid biosynthesis
    DOI:  https://doi.org/10.1016/j.cels.2025.101406
  34. ACS Appl Mater Interfaces. 2025 Oct 10.
    Water-Modulated Construction of Nanoclay/g-C3N4 Heterostructures for Selective Generation of Singlet Oxygen in Peroxymonosulfate Activation Processes.
    Qian Zhang, Menghan Yu, Aidong Tang, Liangjie Fu, Xiongbo Dong, Huaming Yang.
      Singlet oxygen (1O2), a highly selective oxidant in advanced oxidation processes, remains challenging to generate efficiently and exclusively due to competing radical pathways. Here, we report a water-modulated interlayer confinement strategy for constructing nanoclay/g-C3N4 heterostructures (RT/CN) that promote selective 1O2 production through tailored interfacial electronic modulation. By precisely tuning the rectorite-to-water ratio during precursor grinding, urea undergoes confined polymerization within the RT interlayers, forming ultrathin g-C3N4 nanosheets covalently anchored via Si-N-Al linkages. This asymmetric interfacial architecture induces localized electron redistribution, enabling peroxymonosulfate (PMS) activation through an oxidative, nonradical pathway while fully suppressing radical generation. The optimized RT/CN-H3 catalyst achieves dual-mode 1O2 production: interfacial electron transfer under dark conditions and hole-mediated enhancement under visible light. Remarkably, it delivers 96% degradation of Orange II in 30 min (k = 0.102 min-1) under dark, with near-exclusive (∼100%) 1O2 selectivity. This scalable, metal-free platform demonstrates robust reactivity across diverse pollutants and environmental conditions. The work establishes a generalizable strategy for harnessing the interlayer confinement of natural minerals, with broader implications for sustainable oxidation chemistry, environmental remediation, and the rational design of green catalytic systems.
    Keywords:  dual-mode oxidation; g-C3N4; interlayer confinement; peroxymonosulfate activation; singlet oxygen
    DOI:  https://doi.org/10.1021/acsami.5c15587
  35. Nano Lett. 2025 Oct 08.
    Transient Emulsion-Assisted Fabrication of Liquid Metal Microsphere Arrays for Mechanically Adaptive Anisotropic Conductive Packaging.
    Rui Pan, Dilong Liu, An Cao, Wei Zhang, Yunle Yao, Yi Gong, Lin Chen, Yue Li, Xingyou Tian.
      Liquid metal (LM) microsphere arrays hold great promise as adaptive conductive frameworks for next-generation flexible electronics. However, the spatial arrangement of micronano LM droplets poses a longstanding challenge to their practical application due to their inherent ultrahigh surface tension. Here, we introduce a transient emulsion-assisted self-assembly and fusion strategy that turns surface tension from a barrier into a driving force for ordered LM microsphere formation and precise positioning. The resulting LM microspheres exhibit dynamic interfacial conduction, enabling excellent mechanical adaptability under deformation. Embedded into a thermally responsive polymer matrix, the fabricated LM microsphere-arrayed anisotropic conductive film (ACF) achieves an ultralow contact resistance (0.303 mΩ/mm2, 96% lower than conventional ACFs) and stable performance under cyclic loading. Demonstrated in packaging flexible chip-LED arrays with stretchable circuits, this approach ensures both mechanical resilience and electrical reliability. This work offers a scalable pathway toward high-performance, compliant interconnects for future soft electronic systems.
    Keywords:  anisotropic conductive film; liquid metal microsphere array; soft electronic packaging; surface-tension-driven fusion; transient emulsion-assisted self-assembly
    DOI:  https://doi.org/10.1021/acs.nanolett.5c03578
  36. Proc Natl Acad Sci U S A. 2025 Oct 14. 122(41): e2415658122
    Graph neural networks for predicting metal-ligand coordination of transition metal complexes.
    Jacob W Toney, Roland G St Michel, Aaron G Garrison, Ilia Kevlishvili, Heather J Kulik.
      High-throughput virtual screening campaigns are invaluable for surveying the combinatorial space of possible transition metal complexes (TMCs), but they rely on accurate metal-ligand connectivity for meaningful results. Here, we curate a dataset of 70,069 unique ligands of known coordination from experimental structures of TMCs deposited in the Cambridge Structural Database. Using this dataset, we train separate graph neural network models to predict the total number and individual identities of ligand coordinating atoms with high accuracy and precision. Interpreting each model in terms of the learned molecular representations uncovers trends aligned with our understanding of coordination chemistry as well as chemical insights. Next, we integrate the trained models with the high-throughput screening software molSimplify and illustrate their utility by generating 1,175 TMCs and validating their geometries with density functional theory calculations. We anticipate these models will accelerate computational screening of TMCs with de novo combinations of metals and ligands in physically realistic coordination.
    Keywords:  graph neural networks; machine learning; transition metal chemistry
    DOI:  https://doi.org/10.1073/pnas.2415658122
This page is being maintained by Thomas Krichel. It was last updated on 2025‒10‒12 at 9:27.