Search

Despite rapid development of model algorithms, more evidence is needed on the widespread clinical implementation of auto-segmentation tools. In this prospective, multicenter, observational trial, the authors find that AI-assisted delineation performs better than manual delineation while improving time efficiency and reducing performance metrics variability across centers and physicians with varying expertise.

Macrophages and lipid metabolism have an important function in atherosclerosis. Here, the authors characterise the function of chitinase-like protein 3 (CHIL3) or its human functional homologue chitinase-3-like 2 (CHI3L2), and show these proteins enhance atherogenesis in a macrophage dependent manner and function by hydrolysing glycans on CD36 which enhances lipid uptake, activates mTOR and inflammatory macrophage responses and suppresses PPARγ-ABCG1-mediated cholesterol efflux.

Converting syngas into high-value long-chain alcohols is a major industrial challenge due to poor product selectivity. This study achieves 80% selectivity through a modular catalytic design that boosts carbon efficiency while minimizing waste.

Molecular ferroelectrics are promising for human interfaced high frequency acoustic sensing, though existing ferroelectric materials are often brittle making them impractical for sensors. Here the authors describe a stretchable ferroelectric material, using trimethylchloromethyl ammonium cadmium chloride and polyurethane electrospun for high-accuracy speech recognition under stretching.

This study maps China’s rooftop PV potential across 349 cities through 2050. It reveals diverging trajectories where carbon mitigation peaks in 2035–2040, yet water and land savings rise continuously, highlighting PV system’s evolving systemic value.

Maternal obesity is known to cause systemic lipid dysregulation, yet its effect on lipid reprogramming during the maternal-to-zygotic transition (MZT) remains unknown. Here, they show that obesity disrupts very-long-chain fatty acids (VLCFAs) storage in oocytes, impairing peroxisomal protein import in 2-cell embryos.

Accurately mapping diverse cell types in complex tissues remains a challenge for spatial transcriptomics. Here, the authors present UCASpatial, an entropy-weighted algorithm that enables ultra-precision deconvolution to resolve fine-grained immune landscapes and explore intercellular mechanisms.

Metal–organic frameworks with dynamic metal–ligand bonds form reversible gates that admit enzymes larger than the pores, enabling stable encapsulation and efficient multi-enzyme cascade catalysis for glycan precursor synthesis.

Here, the authors present a machine-learning assisted single-molecule sensor that can precisely quantify the structurally diverse per- and polyfluoroalkyl carboxylic acids in the absence of standards, together with state-of-the-art detection limits.

Stacking XA48-mediated effector-triggered immunity with XA21-mediated pattern-triggered immunity in Oryza sativa japonica reconstitutes the broad-spectrum resistance from wild rice.

HorusEye is a foundation model for universal X-ray tomography restoration that learns realistic degradation directly from data. It supports imaging at substantially lower doses and reduces hardware requirements while improving expert analysis and downstream AI performance.

The CMS experiment at CERN reports one of the highest-precision measurements of the W boson mass, finding it in line with standard model predictions and at odds with recent anomalous measurements.

SnRK1β1A in rice promotes susceptibility to multiple fungal diseases, and disrupting this infection-inducible gene confers broad-spectrum resistance without compromising growth or yield under normal field conditions.

Wu et al. identify circulating proteins associated with cardiovascular and neurological imaging traits and diseases, and validate GDF15 as a shared determinant of heart and brain pathologies in mice.

The incorporation of alkali metals in kesterite solar cells enables high efficiency but is challenging to control. Xu et al. show that combining Na and Li ions yields optimal phase formation, enabling efficient solar cells and modules.

Measurement-based quantum computing implements quantum algorithms by performing sequences of measurements on special classes of entangled states, such as cluster states. This approach has now been demonstrated on a superconducting quantum processor.

Yu et al. identify HELDR as a lncRNA co-amplified with EGFR in glioblastoma to promote a KAT7-activated malignancy programme, which can be potentially targeted with anti-EGFR therapy for improved therapeutic outcomes.

Digital array modulation microscopy uses natural Gaussian illumination and virtual detection modulation, together with a single-spectrum reconstruction algorithm, for three-dimensional super-resolution imaging with 300-nm and 100-nm axial and lateral resolution, respectively.

Self-regulation is critical yet inconsistently defined across neurodevelopmental conditions (NDCs). This preregistered PRISMA-aligned overview of reviews aimed to (1) review and summarize definitions and models of self-regulation, (2) synthesize key self-regulation-related constructs and (3) identify common self-regulation measures in NDC populations.

Infrared information encryption is a promising technique for data transmission, though current strategies fail to transmit data in real-time. Here the authors report a multilayer relaxor ferroelectric polymer stack which exhibits rectangular temperature wave for real-time and programmable encryption.

The Taiwan Precision Medicine Initiative (TPMI) has built a cohort of more than half a million individuals with Han Chinese ancestry, providing a rich resource of genetic and medical data for this group.

Here the authors report that genetic or pharmacological inhibition of phosphoglycerate kinase 1, an enzyme involved in glycolysis, attenuates acute kidney injury in mouse models of acute kidney injury via inhibition of ferroptosis.

U.S. federal climate policies can reduce air pollutant emissions and associated health impacts from fine particulate matter. However, near-term CO₂ reductions alone are insufficient to address racial/ethnic disparities in pollution exposure.

Two-dimensional (2D) materials are promising for in situ encoding and optoelectronic sensing in computing technologies. This Review discusses integrated optoelectronic sensing based on 2D materials and future challenges and prospects for integrated sensing encoding information beyond optical signals.

Here authors show that isolating interconnected channels in metal-organic frameworks eliminates inefficient molecular shuttling, greatly improving diffusion and separation performance. The strategy enhances isomer separation and material stability.

Conversion-type batteries with high energy storage efficiency are important to minimize the energy loss during energy storage. Here, authors develop a rechargeable battery with a maximum energy storage efficiency of 99.5% based on S-Cl synergistic chemistry and Cl₂ mediating role.

SDG 6.1 on drinking water lags behind, while rainwater holds future potential to address this goal. A parametric framework proposes a comprehensive stepwise roadmap demonstrating rainwater could provide safe drinking water for 0.45–2.08 billion people.

The Taiwan Precision Medicine Initiative recruited and genotyped more than half a million Taiwanese participants, almost all of Han Chinese ancestry, and performed comprehensive genomic analyses and developed polygenic risk score prediction models for numerous health conditions.

Liu et al. report Chinese normative lifespan brain charts showing later neurodevelopmental milestones than those detected in Western cohorts. Individual deviations from these norms are valuable in assessing clinical risk and outcomes.

Mechanistic analysis reveals that OSBPL3 drives colorectal cancer progression and trametinib resistance by regulating Hippo-YAP1 signaling via 14-3-3 interaction.

This study reports that genomic signatures composed of the loss of pluripotency inhibitors, expansion of pluripotency activators and maintenance of an epigenetically permissive state contribute to the plantlet formation in Kalanchoe.

Dendrobium is a genus of mostly epiphytic and lithophytic orchids. Here, the authors conduct comparative genome analysis for 24 chromosome-level genome assemblies and genome skimming sequencing data for 204 species in this genus, and reveal genomic variation, biosynthesis and evolutionary adaptation in Dendrobium orchids.

In nasopharyngeal carcinoma (NPC), cisplatin is known to be associated with poor treatment compliance and notable side effects therefore more effective platinum drugs are expected. Here this group updates the 10-year survival analysis results of a multicenter, randomized, phase 3 trial evaluating the efficacy of lobaplatin and fluorouracil induction chemotherapy plus concurrent chemoradiotherapy on 502 NPC patients.

Seawater electrolysis for green hydrogen faces the critical challenge of preventing chloride oxidation. Here, the authors report that reconstructed sulfate layers on NiFeS anodes establish a hydrogen-bond network that selectively weakens chloride binding while allowing hydroxide access for electrolysis.

Socioeconomic status is multidimensional and can change over time. Xu, Lui, Ji et al. found dimension-specific, late-stage-specific and change-specific socioeconomic effects on brain, cognition, personality and emotion.

Gut microbiota influence bone health, but the genetic and metabolic mechanisms are unclear. Here, the authors show that specific bifidobacterial taxa causally reduce bone mineral density, partly via n-3 fatty acid metabolism, highlighting host-microbe interactions with potential therapeutic implications.

Variations in the valence states of nickel oxide hinder the growth of a uniform self-assembled monolayer (SAM). Here, authors develop a SAM with the oxidized product forming S−O−Ni bond at Ni³⁺-rich regions, achieving maximum device efficiency of 26.5% in operationally stable perovskite solar cells.

YTHDC2 is a promising therapeutic target, but lacks potent inhibitors. Yang et al. develop a deep learning-based molecule generator EPMolGen. Using this model, they discover a potent, selective, and cell-active small molecule inhibitor of YTHDC2.