Browse Articles
Discover research articles across all indexed journals
Mechanistic studies of the human NIPA2 transporter
Ion-Pair Mediation Enables Ambient Urea Formation from CO <sub>2</sub> and NH <sub>3</sub> in Microdroplets
Generalized graph foundation models as versatile data-driven digital twins for complex technological systems
Abstract Digital twins are comprised of computational models that mimic the ‘as built’ characteristics of devices, systems, and networks of systems whose performance in the real world warrants quantitative and critical assessment.The literature on constructing digital twins is historically focused around task-specific, physics-based models that seek to understand the device from first principles, thereby constructing an idealized digital twin, based on the known physics, of the system.However, these “physics-based digital twins” (pbDT) can be quite difficult to construct, as the level of detail required to accurately model the complex physics of many devices is often missing or expensive to obtain, especially for systems with widespread deployment.Additionally, pbDTs generally assume the device is working as intended; in practice, many systems experience some form of performance degradation, or derating, that causes them to operate off-specification, in manners such that the basic physics is undetermined.As it is often the goal of a digital twin model to quantify these departures from the idealized system, it is quite difficult to separate assumptions from the expected model output.In contrast, data-driven digital twins (ddDT) seek to model the system as it actually is based on real observations and datastreams arising from the device in question.ddDTs enable agility in responses because they learn system dynamics directly from data at operational timescales. As a result, complex physical phenomena of a physical system or process, which are often difficult to explicitly model, can be captured since their combined effects are implicitly reflected in the measured datastreams of sensors and control signals. Additionally, ddDTs generally utilize a flexible model architecture (typically an artificial neural network) to avoid injecting implicit bias into the system. This flexibility also lends itself to another advantage: modularity, that a single ddDT model architecture can be used to train ddDTs for multiple, quite different systems, and to answer multiple different questions about the real-world system.With a ddDT, it is possible to train the model in a self-supervised manner via a reconstruction objective to obtain a trained “encoder” module.This trained encoder module can then be used as a Foundation Model (FM) for the system to answer different task-specific questions without necessitating the training of a new, task-specific, pbDT or training another ddDT from scratch.This work presents a unified pipeline for constructing data-driven Foundation Models for three exemplifying cases: solar-photovoltaic fleets, direct-ink-write additive manufacturing, and laser-powder-bed-fusion additive manufacturing.Although these three systems are conceptually very different, the presented Foundation Model utilizes the flexibility of spatiotemporal graph neural networks (st-GNNs) to apply the same methodology to each case, allowing scientists to focus on their scientific objectives rather than troubleshooting an overwhelmingly detailed physics-based modeling pipeline.
On the nature of chemical short-range order evolution
Topology-Gated λ Exonuclease Enables Amplification-Free Signal Boosting
Health behaviors and health status among physicians and the general population in South Korea
Abstract Since physicians affect the health and disease management of patients, they should act as a model for the lifestyle related to health. This study compared the health behaviors and health status of physicians and the general population in Korea. We used data from the 2016 Korean Physician Survey (KPS), provided by the Korean Medical Association, and the 2016 Korea National Health and Nutrition Examination Survey (KNHANES), administered by the Korea Disease Control and Prevention Agency. Propensity-score matching (PSM) was applied to ensure sociodemographic comparability between the two groups, followed by chi-squared tests(or t-tests) and conditional logistic regression to compare health-related characteristics. Compared to the general population, physicians showed a higher health screening rate (83.7% vs. 78.2%), lower current smoking rate (19.5% vs. 30.4%), and higher rate of practicing moderate to vigorous physical activities (7.6% vs. 2.7%). Physicians had a lower average of sleeping hours per day (6.5 h vs. 6.9 h). Physicians had a lower subjective perception of good health compared to the general population (28.9% vs. 37.4%), while their perceived stress was higher (59.5% vs. 32.6%). Physicians showed better health behaviors than the general population but they had fewer sleeping hours per day, lower subjective perception of good health, and higher perception of stress. These aspects should be improved as they might affect negatively the quality of healthcare that they provide.
Comprehensive biophysical and structural profiling of alpha-actinin-2 variants reveals mechanistic diversity in hypertrophic cardiomyopathy
Abstract Hypertrophic cardiomyopathy (HCM) is a genetic disease associated with sudden cardiac death. Variants in alpha-actinin-2 (ACTN2), a Z-disc protein that anchors actin thin filaments have been implicated in HCM, yet their structural consequences remain poorly defined. Here, we characterise seventeen HCM-associated ACTN2 variants spanning multiple domains using an integrated and tiered workflow combining high-throughput assays, structural modelling and biophysical approaches. All variants display reduced solubility, with actin-binding domain (ABD) substitutions showing pronounced thermal instability by differential scanning fluorimetry. Modelling of nine variants predicts diverse pathogenic mechanisms including compromised actin-binding, impaired ABD regulatory conformations, disrupted dimerisation interfaces, and perturbed domain architecture. Crystal structures of two rod-domain variants reveal intact dimerisation despite modelling predictions. Actin-binding assays for ABD variants confirm altered actin engagement suggesting that binding dynamics may drive pathogenicity. Limited proteolysis indicates reduced structural stability across variants, while size-exclusion chromatography coupled with multi-angle light scattering or small-angle X-ray scattering (SEC-MALS/SAXS) shows a strong propensity for aggregation. Batch-mode SAXS further demonstrates early aggregation onset in selected ABD variants at elevated temperatures. Collectively, these findings establish that HCM-linked ACTN2 variants compromise protein integrity through multiple mechanisms, highlight the ABD as a hotspot of vulnerability and provide a potential framework for interpreting cardiomyopathy-associated variants.
Real-Time Probing of the Multistep/Multichannel Photoisomerization and Dissociation of Propenal
Antifungal activity and defense response activation by Co- and Ni-doped ZnO nanoparticles against tomato root rot disease
Shared cloud interactions unveil a candidate binary-system supernova pair with no known analogue
Abstract IC 443 is one of the most extensively studied supernova remnants in the Galaxy, yet the surrounding region remains shrouded in mystery. Here we show that 16 years of Fermi -LAT observations uncover extended gigaelectronvolt gamma-ray emission from G189.6+3.3, a source long hidden in the shadow of the much brighter IC 443. The gamma-ray morphology is consistent with the X-ray shell detected by eROSITA, indicating an origin of the emission in the remnant. Spectral and spatial analyses identify distinct hadronic and leptonic gamma-ray components associated with regions containing and lacking molecular gas, respectively, revealing a clear spatial segregation of these emission mechanisms. In the northern shell boundary, hadronic gamma rays coincide with an H α filament linked to the S249 ionized hydrogen cloud. Morphological studies, crushed-cloud modelling and ultraviolet observations of this region support gamma-ray production through proton re-acceleration in compressed post-shock gas. The shared-cloud interaction positions G189.6+3.3 at the same distance as IC 443, with a linear separation between the two remnants and a time delay in their explosion events, bolstering the hypothesis that these two remnants are part of a binary system, both resulting from separate supernova events.
Carbon-Mediated Rechargeable Operation for Light-Driven Ammonia Production Using Quantum Dot- <i>Azotobacter vinelandii</i> Hybrids
A multidimensional benchmarking framework for large language models in oncologic decision making
Ultrafast multi-level control of sub-50 nm skyrmions in a Pd-intercalated van der Waals magnet
Abstract Achieving ultrafast, multi-level control of nanoscale skyrmions offers a transformative route for advancing van der Waals spintronics towards high-speed, scalable neuromorphic computing applications. However, progress has been impeded by the relatively large skyrmion size (~100 nm) in existing van der Waals magnets and the lack of efficient control strategies. Here, we simultaneously address both challenges by combining atomic intercalation with femtosecond laser manipulation. Through Pd atomic intercalation into the van der Waals magnet Fe 3-δ GaTe 2 , we realize magnetic field-stabilized skyrmions with an average diameter of ~43 nm at room temperature, the smallest skyrmions reported in the van der Waals magnets to date. Mechanism analysis reveals that this size reduction arises from enhanced Dzyaloshinskii-Moriya interaction and suppressed Heisenberg exchange coupling. On this tailored platform, we further demonstrate femtosecond laser-induced ultrafast generation of 43 nm skyrmions with an ultra-low energy consumption of 0.6 pJ per skyrmion. Most importantly, by tuning the laser pulse number, we achieve deterministic, multi-level modulation of skyrmion density, enabling skyrmion-based optical neuromorphic computing with a simulated training accuracy of ~91%.
General Synthesis of γ-Arylamino Alcohols via Difunctionalization of Terminal Alkenes Using a Supported Cobalt Single-Atom Catalyst
Whole exome sequencing and genotype-phenotype correlation in homocystinuria in an Iranian population: a multicenter study
Abstract Homocystinuria is an uncommon metabolic disorder characterized by increased homocysteine concentrations. The condition may arise from mutations in the cystathionine beta-synthase (CBS) gene (classic) or in other genes associated with the cobalamin and folate metabolic pathways (non-classic). This study investigated the genetic diversity and clinical outcomes of homocystinuria in an Iranian pediatric population, exploring genotype-phenotype correlations. A multicenter cross-sectional study carried out from January 2024 to August 2025 at three principal referral centers in Iran. A total of 48 pediatric patients diagnosed with homocystinuria and possessing whole-exome sequencing (WES) results were included. Clinical data, including neurological, ocular, and vascular manifestations, were extracted. The majority (52.1%) of patients in the cohort had B12-related homocystinuria, 25% had CBS-related, and 22.9% had B9-related problems. In most patients (93.7%) at least one neurological symptom was identified, with seizures and developmental delay being the most common. Nine patients (18.7%) had ocular symptoms, 7 patients (14.5%) had skeletal symptoms and skin/hair manifestations were seen in 5 patients (10.4%). Classical homocystinuria was associated with significantly elevated plasma homocysteine levels in comparison to non-classical subtypes (p = 0.011). Novel variants were identified in 10 patients (20.8%), affecting MMACHC , ABCD4 (associated with Cobalamin J type), HCFC1 (associated with Cobalamin X type), CBS , TCN2 and MTHFS genes. This study illustrates the genetic diversity and clinical variability of homocystinuria in Iran. The results highlight the clinical importance of regional genetic databases and whole-exome sequencing in the diagnosis and management of this rare metabolic disorder.
Reversible layered/non-layered phase transition in a topological semimetal
Ion-Modulated Ostwald Ripening Dynamics of Nitrogen Nanobubble Pairs
Bimodal social capital associated with psychosocial profiles and brain network topology in a Japanese population
Proton trap engineered electric swing adsorption for scalable and cost-effective direct air capture
Abstract Direct air capture (DAC) is critical to achieve carbon neutrality, yet current technologies face significant barriers to widespread, cost-effective deployment. Amine-based electric swing adsorption (ESA) offers a promising low-energy, steam-free pathway, but its efficiency is fundamentally limited by an inherent 2:1 amine-to-CO 2 stoichiometric penalty. Here, we overcome this bottleneck by engineering a point defect-mediated proton trapping network into ESA sorbents, enabling a 1:1 amine-CO 2 stoichiometry. Our engineered sorbent achieves a CO 2 uptake of 6.57 mmol g −1 from 400 ppm CO 2 , a 28.8% improvement over the state-of-the-art sorbents. Regeneration is achieved with a low energy input of 3.4 GJ t −1 and exhibits a CO 2 release rate 48% faster than conventional thermal methods. N 5- d GA remains stable under 0-80% relative humidity fluctuations and at a gas velocity of 1 m s −1 . Techno-economic analysis projects DAC operating costs of $48-62 t −1 using renewable electricity, up to 78% lower than temperature swing adsorption DAC and below the $100 t −1 CO 2 target. This work presents a sorbent design and ESA process, establishing a scientifically rigorous and economically viable pathway towards gigaton-scale DAC deployment.