ABSTRACTBiomolecular condensates are membrane‐less assemblies that selectively concentrate proteins, RNAs, and metabolites to integrate developmental and environmental cues. The remarkable diversity of plant condensates reflects the constraints of sessile organisms that must coordinate postembryonic organ development with continuous environmental adaptation. We review how plants employ condensates to integrate temperature, light, redox, and nutrient signals. We provide physicochemical foundations, including phase diagram behavior, critical solution temperature properties, and sticker‐and‐spacer models, as a framework for interpreting how environmental stimuli are transduced into condensate assembly/disassembly. We organize each biological system through a unified scaffold–client–RNA–metabolite framework, distinguishing experimentally validated conclusions from open mechanistic questions. Applying this framework across nuclear, cytoplasmic, chloroplastic, and membrane‐associated condensates, we evaluate how temperature shifts, redox changes, post‐translational modifications, and metabolite fluctuations drive reversible phase transitions. We highlight how saturation concentration thresholds function as nonlinear filters buffering environmental noise, how membrane‐associated phase separation may nucleate cytoplasmic condensates, and where current evidence remains insufficient to distinguish bona fide liquid–liquid phase separation from alternative assembly mechanisms. By grounding plant condensate biology in physicochemical principles and comparative evidential analysis, we identify both well‐supported mechanisms and critical gaps that must be addressed to translate condensate‐biology into strategies for crop‐resilience.

ABSTRACT Bionic visual processing hardware serves as the core technology for mimicking the efficient information processing of biological vision systems and forms the foundation for perceiving and recognizing both static and dynamic scenes. However, most current bionic visual hardware remains at the level of individual devices and simple arrays. Full hardware implementation of bionic vision chips, as well as comprehensive demonstrations in practical application scenarios, is still scarce. This work proposes a complete optoelectronic insect‐inspired visual sensor. We realize a Si QDs/ReS 2 heterogeneous integrated neuromorphic chip fabricated using a 180‐nm CMOS process technology and demonstrate its capabilities on both static and dynamic visual tasks. Based on the wavelength‐dependent synaptic properties of ReS 2 , this neuromorphic photoreceptor chip exhibits high‐fidelity discrimination between purple and red features. Through light pulse sequence modulation, it achieves temporal encoding of dynamic visual information and can precisely identify leaf movement trajectories in eight directions. We further achieved 3D perception capabilities, attaining 99.4% accuracy in classification tasks. This work establishes a novel 1D/2D/3D heterogeneously integrated material platform and a CMOS‐compatible integration strategy for a low‐power, multifunctional brain‐inspired neuromorphic sensor chip, advancing the application of optoelectronic devices in artificial intelligence and machine vision.

ABSTRACT Tumor immune escape is a major barrier to durable cancer immunotherapy, as advanced malignancies create a tumor microenvironment (TME) that preferentially exhausts and disables T cell responses. While most approved cell therapies are T cell‐based, this limitation motivates the exploration of an alternative effector cell platform. Natural killer (NK) cells, innate cytotoxic lymphocytes capable of antigen‐independent recognition and killing, offer a compelling foundation for next‐generation therapies with an improved safety profile. In this review, we first outline the cellular, molecular, and metabolic features of the immunosuppressive TME that restrict cytotoxic lymphocyte function, emphasizing mechanisms that limit immune cell‐mediated responses. We then summarize key aspects of NK cell biology that can circumvent these barriers and critically evaluate current NK‐based strategies, including engineered chimeric antigen receptor (CAR)‐NK products and metabolic and trafficking interventions. Finally, we highlight emerging in vivo viral and mRNA/lipid nanoparticle platforms for CAR‐NK generation and their potential to enhance scalability and therapeutic durability.

ABSTRACTFoundation Models (FMs) have made significant strides in both industrial and scientific domains. In this paper, we evaluate the performance of FMs for single‐cell sequencing data analysis through comprehensive experiments across eight downstream tasks pertinent to single‐cell data. Overall, the top FMs include scGPT, Geneformer, and CellFM by considering model performances and user accessibility among ten single‐cell FMs. However, by comparing these FMs with task‐specific methods, we found that single‐cell FMs may not consistently excel than task‐specific methods in all tasks, which challenges the necessity of developing foundation models for single‐cell analysis. In addition, we evaluated the effects of hyperparameters, initial settings, and stability for training single‐cell FMs based on a proposed scEval framework, and provide guidelines for pre‐training and fine‐tuning to enhance the performances of single‐cell FMs. Our work summarizes the current state of single‐cell FMs, points to their constraints and avenues for future development, and offers a freely available evaluation pipeline to benchmark new models and improve method development.

ABSTRACTGrowing evidence suggests a role for the gut microbiome in progression of cholangiocarcinoma (CCA), however, its diagnostic and therapeutic potential remains incompletely characterized. Here, metagenomic sequencing was performed on fecal samples (n = 785) from individuals across East, Central, and Northwestern China. Gut microbial dysbiosis in CCA was characterized by depletion of short‐chain fatty acids‐producing species and enrichment of potential pathobionts (Klebsiella aerogenes, Clostridium symbiosum). Diagnostic models built using species‐level markers demonstrated superior performance, compared to pathway‐based models, achieving area under the curve (AUC) values of 98.63% and 99.42% in the discovery cohort, with robust cross‐regional validation (AUC = 80.89% and 80.43%). The model effectively distinguished CCA from hepatocellular carcinoma (AUC = 97.86%) and liver fibrosis (AUC = 98.73%) and nonalcoholic fatty liver disease (mean AUC = 96.86%). Analysis of public datasets encompassing 6847 samples across 31 studies and 11 disease states revealed moderate disease specificity influenced by biomarker overlap across conditions. Mechanistically, depleted Bifidobacterium pseudocatenulatum suppressed CCA progression, associated with inhibition of the PI3K‐AKT‐mTOR pathway. Collectively, this study supports the potential of fecal metagenomic signatures as a complementary noninvasive aid for CCA detection, and provides functional evidence for a candidate protective microbe.

ABSTRACTThe key to achieving high power density in piezoelectric materials lies in the synergistic optimization of both high piezoelectric constant (d33) and piezoelectric voltage coefficient (g33). Traditional inorganic ceramics offer high d33 but suffer from low g33, poor flexibility, and toxic‐Pb content. Polymers, while flexible, are limited by their relatively low d33 and insufficient stability. Molecular ferroelectrics can simultaneously offer high piezoelectricity, flexibility, and low loss, yet most reported systems contain toxic‐Cd components, restricting their use in biomedical and wearable applications. Metal‐free molecular ferroelectrics, replaced toxic metals while retaining the molecular ferroelectric attributes, combine environmental friendliness, and biocompatibility, making them ideal candidates for piezoelectrics. However, their piezoelectric output has traditionally been weak. Here, a novel metal‐free molecular ferroelectric, [(CH3C6H10NH3)(18‐crown‐6)][ClO4] (CH3C6H10NH3 = trans‐4‐methylcyclohexylaminium), was designed and synthesized via a methylation‐assisted ring‐expansion strategy, enhancing the piezoelectric performance. The material exhibits a d33 of 67.42 ± 1.87 pC/N and a g33 of approximately 2768.9 × 10−3 V m/N. When composited with thermoplastic polyurethane (TPU) into a porous architecture, the material delivers an output voltage of 120 V and a power density of 432.1 µW/cm2, representing the highest performance reported among metal‐free molecular piezoelectrics. This work offers a practical way to design high‐performance, biocompatible piezoelectrics, creating a new platform for flexible sensors and self‐powered devices.

ABSTRACTAndrogen receptor (AR) signaling is a primary oncogenic driver of castration‐resistant prostate cancer (CRPC), yet the mechanism remains incompletely understood. Through proteomic profiling of CRPC and primary PCa cells, we identify G Protein Nucleolar 3 (GNL3) as a novel AR coregulator. GNL3 physically interacts with AR, enhances its chromatin occupancy, and directly coactivates transcriptional programs that promote cell proliferation, including NEK2 and CDC20. Concurrently, GNL3 functions as a corepressor of immune‐responsive genes such as CXCL10 and TAP1 via class I histone deacetylases (HDACs), thereby facilitating CD8+ T cell elimination and establishing an immunosuppressive tumor microenvironment. GNL3 expression and AR‐GNL3 complex formation progressively increase from normal prostate to CRPC and correlate with poor clinical outcomes. Functionally, GNL3 knockdown sensitizes CRPC cells to AR antagonists and impairs tumor growth and metastasis. Furthermore, we demonstrate that combinatorial inhibition of NEK2, class I HDACs, and AR signaling can be a potential therapeutic strategy for CRPC. Overall, these findings establish GNL3 as a dual‐function AR coregulator and therapeutic target, providing mechanistic insights into transcriptional regulation and immune evasion in advanced PCa.