Browse Articles
Discover research articles across all indexed journals
Design and application of an AI- and AR-enhanced serious game for interactive learning in the Blue Calico Museum in China
Abstract With the widespread application of Artificial Intelligence (AI) and Augmented Reality (AR) technologies in the field of cultural education, technology-integrated serious games are increasingly emerging as effective tools for the dissemination of intangible cultural heritage. This study targeted university students and developed and deployed an AI- and AR-based serious game titled Dye Verse at the China Blue Calico Museum, serving as a museum-based learning intervention for higher education audiences. The game integrates features such as character creation, semantic guidance, AR recognition, and navigation, aiming to enhance students’ immersion and learning motivation. An experimental study involving 60 participants (N = 60) was conducted using pre- and post-knowledge tests and the User Experience Questionnaire (UEQ). Statistical analysis indicates that the experimental group significantly outperformed the control group in dimensions such as cultural knowledge acquisition, interactive engagement, and emotional identification. These findings validate the educational potential and communicative advantages of integrating AI and AR technologies in serious games. The study offers a cost-effective and highly interactive digital solution for small and medium-sized museums and provides both theoretical and practical implications for the sustainable dissemination of cultural heritage through gamified education.
Selective membrane wetting of phase-separated giant unilamellar vesicles by coacervate droplets
Interfacial Ru/RuOx heterostructures on carbon support regulate selectivity in lignin hydrodeoxygenation
Abstract Constructing well-defined heterostructure interfaces in catalysts provides an approach to modulate scaling constraints and steer reaction pathways in biomass upgrading. Herein, we demonstrate that thermal restructuring of hydroxyl groups on carbon nanofibers (CNF) induces the formation of heterostructures of Ru/RuO x , which function as bifunctional active sites for the one-pot hydrodeoxygenation (HDO) of lignin to liquid hydrocarbons. The optimized 5 wt% Ru/CNF catalyst delivers promising performance, achieving a mass/carbon yield of 49.1%/67.7%, with high selectivity toward saturated cycloalkanes. X-ray absorption spectroscopy and near-ambient pressure X-ray photoelectron spectroscopy confirm that thermal treatment of CNF tunes the oxidation state of Ru. DFT calculations reveal that O-rich Ru/CNF forms interfacial heterostructures of Ru/RuO x polarized active sites, characterized by O δ ⁻···Ru δ++ ···Ru δ+ ensembles that heterolytically activate H 2 and strongly polarize C-O bonds in phenolic intermediates. The cooperative interplay between metallic Ru and partially oxidized RuO x interfacial sites lowers the energy barriers for hydrogenation and deoxygenation reactions, enabling a cooperative reaction pathway. These insights elucidate the molecular basis of tunable selectivity in lignin HDO and demonstrate that a polarized, oxygen-decorated metal-support interface provides general design principles for engineering next-generation catalysts for sustainable fuel production.
Deep visual proteomics uncovers nociceptor diversity and pain targets
Abstract The richness of our somatosensory experience is reflected in the functional diversity of somatic sensory neurons. Single-cell RNA sequencing of sensory neurons has revealed a molecular basis for such diversity 1–3 . However, sensory neuron diversity has yet to be captured at the level of the proteome. Here, we combined electrophysiology with deep visual proteomics 4 to quantify over 6000 proteins from phenotypically-defined sensory neurons in mice and identified proteomic markers of sensory neuron subtypes. Comparative analysis revealed both concordance and meaningful divergence between transcriptomes and proteomes. We further show that up to 3000 proteins can be quantified from one-fourth of a single neuron, demonstrating subset-specific protein signatures. In culture, nociceptive neurons can be acutely sensitized to mechanical stimuli by nerve growth factor (NGF) which normally drives inflammatory pain in vivo 5 . Indeed, overnight exposure of peptidergic nociceptors to NGF and a protein kinase C (PKC) activator produced functional sensitization associated with proteome changes. Functional knockdown experiments identified the up-regulated B3GNT2 enzyme as a potential effector of nociceptor sensitization. In summary, we present a high-resolution proteomic resource linking molecular identity to function, enabling the discovery of mechanisms underlying somatic sensation and pain sensitization.
Mach-Zehnder atom interferometry with non-interacting trapped Bose-Einstein condensates
Mechanical hysterons with tunable interactions of general sign
Planar, spiral, and concentric traveling waves distinguish behavioral states in human memory
Fe(III) complexes with prolonged luminescence lifetimes via excited-state equilibration promoted by reversible intercomponent electron transfer
Abstract Achieving long-lived luminescence in complexes of earth-abundant metals remains challenging because excited states in first-row transition-metal systems typically deactivate rapidly under ambient conditions. Strategies capable of prolonging emission lifetimes in such compounds are therefore of considerable interest. Here we show that iron(III) complexes incorporating pyrene-functionalized ligands display luminescence in fluid solution at room temperature with lifetimes up to 6.5 ns. Spectroscopic analysis indicates that excited-state equilibration occurs through reversible intramolecular electron transfer from the pyrene unit to the iron centre, generating a charge-separated state. Although the ligand-to-metal charge-transfer state can also undergo reversible energy transfer to nearby pyrene triplet states, intramolecular electron transfer dominates, leading to the formation of a pyrene + -iron(II) charge-separated state that acts as a long-lived excited-state reservoir. Equilibration involving this state produces biphasic emission from the iron centre. These findings identify reversible intercomponent electron transfer as a strategy for achieving prolonged luminescence lifetimes in complexes of earth-abundant metals.
An information content principle explains regulatory patterns of gene expression across human tissues
Abstract Gene expression ranges from broadly expressed to tissue-specific patterns, with many genes displaying intermediate specificity. Understanding how regulatory architecture scales with tissue specificity can reveal fundamental principles of genome regulation. By analyzing cis -regulatory element counts across human genes with varying tissue specificity, we identify a non-monotonic pattern: genes with intermediate specificity harbor the most regulatory elements, suggesting distinct regulatory strategies across the expression spectrum. We apply the Minimum Description Length principle from information theory, and maximum parsimony from phylogenetics, to quantify regulatory demand underlying expression patterns. This measure scales consistently with cis -regulatory element counts, transcription factors, microRNAs, and gene structure, and distinguishes switch-like regulation in selectively expressed genes from fine-tuning regulation in broadly expressed genes. Regulatory element abundance peaks in genes of intermediate evolutionary age. Regulatory architecture appears to scale with informational costs, suggesting that the genome operates as a decompression device, where regulation is dictated by minimally required complexity.
CBP/p300 is critical for the expansion and maintenance of functional pancreatic α cell mass
Mechanistic insights into E. coli recovery from growth arrest
Abstract Bacteria survive hostile conditions by shutting down protein synthesis, but how they restart growth remains poorly understood. Here, we use an E. coli Δ rimM strain, which exhibits a prolonged growth arrest, as a model to investigate how bacteria recover from this state and restore protein synthesis. RimM is a conserved ribosome maturation factor for the 3’-major (head) domain of the 16S rRNA within the bacterial 30S subunit. The loss of RimM causes a longer delay in recovery than other 30S maturation factors, including RbfA. Cryo-EM analysis of Δ rimM ribosomes suggests a delayed recruitment of ribosomal proteins to the 30S head domain and increased occupancy of the initiation factors IF1 and IF3, as well as recruitment of the silencing factor RsfS to the 50S subunit. These coordinated changes provide a safeguarding mechanism to block the assembly of premature 70S ribosomes. Notably, while the delayed 30S assembly in Δ rimM reduces the activity of global protein synthesis during the recovery phase, bacteria attempt to compensate for this deficiency by producing higher levels of the ribosomal machinery, indicating a programmatic change in energy allocation. These findings highlight the importance of the RimM-assisted assembly of the ribosomal head domain for bacterial recovery from growth arrest.
Spin-polarized edge modes between different magnet-superconductor-hybrids
Abstract The interplay of magnetism and superconductivity can lead to intriguing emergent phenomena. Here we combine two different two-dimensional antiferromagnetic magnet-superconductor hybrids (MSH) and study their properties using spin-polarized scanning tunneling microscopy. Both MSHs show the characteristics of a topological nodal point superconducting phase with edge modes to the trivial substrate superconductor. At the boundary between the two MSHs we find low-energy modes which are spin-polarized. Based on a tight-binding model we can explain the experimental observations by considering two different topological nodal point superconductors. At their boundary spin-polarized chiral edge modes emerge that connect topological nodal points of the two different MSH. We demonstrate via the complex band structure that due to an asymmetric lateral decay these edge modes are spin-polarized, regardless of the details of the spin structure at the boundary. This work shows how interfaces between two distinct topological nodal point superconductors can serve as a platform to engineer spin-polarized edge modes.
Blockage of autophagy causes severe skeletal muscle disruption in a mouse model for myofibrillar myopathy 6
Abstract Myofibrillar myopathy 6 is a rare, autosomal-dominant neuromuscular disorder caused by an amino acid exchange Pro209Leu in the co-chaperone BAG3 , which disrupts muscle protein turnover and causes severe muscle weakness and shortened lifespan. We generated transgenic mice overexpressing the human mutant BAG3 P209L -GFP, which rapidly develop skeletal muscle weakness unlike controls expressing BAG3 WT -GFP. Here we show that mutant mice exhibit sarcomere breakdown, inflammation, protein aggregates, centralized nuclei and mitochondrial defects in their skeletal muscles, thereby reducing contraction force by ~90%. Omics profiling uncovered impaired protein synthesis, blocked autophagy, impaired mitophagy and loss of sarcomere proteins. Pathway modulation in vitro and in vivo showed autophagy dysfunction as the primary driver for the pathology, while BAG3 knockdown gene therapy markedly restored muscle function in vivo. In summary, this model recapitulates core disease features, revealing how BAG3 aggregates and loss of BAG3 function impair autophagy to drive muscle degeneration.
Structural optimization of drug molecules with incrementally trained language models
Abstract Automating structural optimization of drug molecules for on-target potency by machine learning is an open challenge in chemistry. Here, we capitalize on the ability of chemical language models (CLMs) to learn from sequential data and design new molecules with desired properties. We establish a training strategy mimicking the learning trajectory of a drug discovery program. Incremental CLM fine-tuning with increasingly potent template molecules from a given structure-activity relationship (SAR) series successfully biases the model to design highly active analogues. Prospective application of this technique to ligand development enables the data-driven design of molecules exceeding known representatives of given bioactive chemotypes in potency without external scoring. Our results reveal an ability of CLMs to capture SAR patterns and long-range dependencies, and to exploit SAR knowledge in designing analogues with improved on-target activity de novo corroborating their applicability to structural optimization of drug molecules.
Adipocytes signal to recruit specific mRNAs from surrounding cells to restore expression deficits
CMOS compatible probabilistic computing hardware with cointegrated reconfigurable p-bits and synapse arrays
Transforming healthcare through in-body bioelectronic systems
Abstract Recent advances in the development of in-body bioelectronic systems are providing new opportunities for the clinical management of various diseases and disorders. These emerging technologies are tailored to specific organs and are beginning to blend both diagnostic sensing and therapeutic actuation. The aim of these systems is to seamlessly integrate with the physiological environment, as illustrated by the diverse device strategies discussed throughout this article. Next generation modalities, such as optogenetics combining gene therapy with devices for photostimulation, are gaining popularity and offer advantages over existing therapeutic strategies. In this perspective, we explore the current state of technological developments, key challenges in the field and potential pathways for translating these innovations into clinical practice.
Combining structural modeling and deep learning to calculate the E. coli protein interactome and functional networks
Abstract We report on the integration of three methods that predict, on a proteome-wide scale, whether two proteins are likely to form a binary complex. The methods include PrePPI, which uses three-dimensional structure information as a basis for predictions, Topsy-Turvy, which uses a protein language model, and ZEPPI, which uses evolutionary information to evaluate protein-protein interfaces. Testing on the high-quality HINT database of binary PPIs reveals that the integrated method has better performance and identifies more high-confidence interactions than any of the component methods. The AF3Complex algorithm is used to predict the structures of 374 PPIs with a large fraction having at least partially overlapping interfaces with PrePPI models of the same complex. Clustering of the high-confidence E. coli interactome yields 385 subnetworks which have high functional coherence. Biological insights derived from the subnetworks, including the annotation of proteins of unknown function, are discussed in detail.