Large-scale CRISPR screening in human T cells holds significant promise for identifying genetic modifications that enhance cellular immunotherapy. Yet, many regulators of T cell performance in solid tumours are not revealed in vitro1,2. In vivo screening in tumour-bearing mice is more physiological but has been limited by low intratumoural T cell recovery. Here we developed an in vivo model that efficiently recovers human T cells from solid tumours, permitting genome-wide CRISPR screens with few mice. Tumour-infiltrating T cells from this model exhibit hallmarks of dysfunction compared with splenic T cells, creating an ideal screening context. We performed two genome-wide CRISPR knockout screens to identify regulators of intratumoural T cell abundance and effector function. The abundance screen revealed the P2RY8-Gα13 GPCR signalling axis as a negative regulator of T cell tumour infiltration. The effector function screen identified GNAS as a key driver of T cell dysfunction in tumours, whose product, Gαs, acts as a convergent node downstream of multiple GPCRs sensing distinct suppressive ligands. Knockout of GNAS rendered T cells resistant to multiple suppressive cues and significantly improved efficacy across diverse solid tumour models in chimeric antigen receptor (CAR) and T cell receptor (TCR) systems. Combinatorial knockout of P2RY8-GNAS further enhanced tumour control, demonstrating that complementary in vivo screens can identify orthogonal targets whose combined editing improves therapeutic potency. This flexible, scalable platform can be adapted for systematic discovery of genetic strategies to improve solid tumour T cell therapies.

Europe has experienced extreme floods in recent decades. However, even larger floods are possible and must be considered in flood risk management1,2. Their characteristics can be clarified by analysing the largest documented historical floods. In Central Europe, the Magdalena Flood of July 1342 is usually considered the largest of the last millennium; however, knowledge of its characteristics is incomplete3-7. Here we show that 16 major flood events occurred across much of Europe between late 1341 and 1343. Four of these events had return periods of 500-1,000 years (the Magdalena, Bartholomew, Candlemas and Jacob Floods). Although Magdalena was thought previously to be the only extreme European flood in 13423,7, our new documentary dataset suggests that it formed part of a broader sequence. The year with the greatest number of extreme floods during the past 700 years was 1342, and 1343 ranks among the top ten. This highly unusual sequence of floods had substantial socio-economic impacts, including a paradigm shift in flood mitigation measures in Europe. A series of volcanic eruptions along with multi-annual Arctic sea ice retreat is a plausible cause of this flood sequence. Clusters of extreme floods occurring within a few months are rarely considered in risk management8. Quick and proactive risk strategies are needed that account for this eventuality.

Atomically thin van der Waals crystals epitomize ideal material systems in the two-dimensional (2D) limit. This reduction in dimensionality often leads to important consequences, best exemplified by the emergence of new physics in graphene and other 2D materials that can be readily tuned by gating1,2. Vast opportunities arise in extending this top-down approach to other material systems. Recent experiments have demonstrated that the essential physics of high-temperature superconductivity in cuprates is contained within just two CuO2 planes3. Here we push dimensionality reduction to the extreme by examining a single layer of Bi2Sr2CuO6+δ (Bi-2201), which comprises only one CuO2 plane. In this ultimate 2D limit, we observe a robust dimensionality effect that manifests as an approximately 10% reduction in the optimal superconducting transition temperature. Moreover, this reduction in dimensionality offers unprecedented tunability-we successfully extended the phase diagram of Bi-2201 into uncharted territories via finely controlled oxygenation of single-monolayer specimens. Leveraging this tunability, we discovered that an anomalous metal state emerges between the insulating and superconducting states as the temperature approaches zero. Concurrently, we observe an anomalous scaling behaviour characterized by a divergent critical exponent. These findings illuminate the nature of the superconductor-to-insulator quantum phase transition in cuprates.

Naturalistic goal-directed behaviour often involves continuous actions directed at dynamically changing goals1-3. Just as microeconomics serves as a rigorous foundation for discrete choices, control theory can serve as a foundation for understanding choice in continuous ones3,4. In continuous contexts, behaviour is composed of blends of goals, and the closest analogue to choice is a strategic reweighting of goal-specific control policies5,6. Here, to understand the algorithmic and neural bases of continuous choice, we examined behaviour and brain activity in humans performing a continuous prey-pursuit task7. Using a newly developed control-theoretic decomposition of behaviour, we find that pursuit strategies are well described by a meta-controller dictating a mixture of lower-level controllers, each linked to specific pursuit goals. Neurons in the anterior cingulate cortex predict major changes in policy blends, whereas hippocampal neurons encode and update the latent policy state supporting early planning. Meanwhile, orbitofrontal cortex activity is consistent with an encoding of the current value structure of the task, rather than policy switching. Together these results are consistent with a tripartite functional division in which hippocampus serves as a state-estimating controller, anterior cingulate cortex serves as a meta-controller, and orbitofrontal cortex provides a value context signal.

The establishment of the infant gut microbiome is critical for later health1,2, yet how it is shaped by maternal and early-life factors remains unclear. Here we metagenomically sequenced 4,526 longitudinal faecal samples from 714 mother-infant pairs in the Dutch birth cohort Lifelines NEXT, spanning 12 weeks of pregnancy to 1 year postpartum. We integrated these data with 474 clinical and exposure variables, and with ultra-deep sequencing of breast milk and vaginal microbiomes. We observe that the maternal gut microbiome undergoes only subtle changes during pregnancy and postpartum, influenced by diet, infections and pre-pregnancy smoking. The maternal gut microbiome is a major reservoir for infant gut strains, with only occasional transmission from vaginal and breast milk microbiomes. Mother-infant gut strain sharing is time dependent, and higher maternal gut species abundance increases the likelihood of strain transmission. We find that the maternal gut microbiome is a predictor of infant eczema. Mode of delivery and feeding mode primarily shaped the infant gut microbiome and its functional profiles, with maternal exposures also having a role. Of 585 vaginally delivered infants, 155 were born at home, but home delivery was only moderately associated with infant gut microbiome composition, similar to other birth parameters such as duration of pushing and ruptured membranes. Overall, we highlight the central role of the mother and her microbiome in shaping the infant gut ecosystem and early health outcomes.

The hippocampus forms memories of our experiences in populations of coactive pyramidal neurons (PNs)1-3. Fast-spiking parvalbumin-expressing inhibitory neurons (PV INs) in the dentate gyrus-CA3/CA2 circuit of the hippocampus precisely control PN activity through mossy fibre-dependent feedforward inhibition4-11. PV INs coordinate experience-dependent changes in their intrinsic excitability, synaptic connectivity, physiology and plasticity properties9,12-15-referred to here as experience-dependent PV IN plasticity-to regulate PN activity. PV IN impairments in early life, when neural circuitry is highly sensitive to experience, are thought to result in network hyperexcitability, seizures and impaired cognition, which are hallmarks of neurodevelopmental disorders (NDDs)16-18. Here we designed an input-specific translatome screen to identify regulators of experience-dependent PV IN plasticity genes (XPGs) in the CA3/CA2 subregion of adult hippocampus. We demonstrate that a substantial proportion of upregulated candidate XPGs exhibit haploinsufficiency in autism spectrum disorder, epilepsies, bipolar disorder and schizophrenia, which suggests that there is impaired experience-dependent PV IN plasticity in NDDs. In proof-of-concept experiments, targeted upregulation of a candidate XPG, the homeobox gene Meis2 (ref. 19), in CA3/CA2 PV INs in an NDD risk mouse model in adulthood is sufficient to restore experience-dependent PV IN plasticity. Moreover, ensemble and sharp-wave ripple properties and cognition were improved, and seizures were suppressed. Thus, experience-dependent PV IN plasticity is a convergent mechanism for NDD risk genes that can be re-instated in adulthood to reverse developmental deficits in circuitry, network excitability and cognition.

Perioperative nivolumab significantly improved event-free survival (EFS) compared with placebo in patients with resectable non-small cell lung cancer (NSCLC) in the CheckMate 77T study (ClinicalTrials.gov NCT04025879 )1. Here, after randomization, 98 out of 229 patients who received nivolumab and 92 out of 232 patients who received placebo had evaluable biomarkers (41% of randomized patients). Of the 98 patients receiving nivolumab, 83 (85%) had detectable circulating tumour DNA (ctDNA) before initiating neoadjuvant treatment and 90 (92%) at neoadjuvant treatment completion. Of the 92 placebo-treated patients, 75 (82%) had detectable ctDNA at the treatment start and 78 (85%) at completion. Among the 98 nivolumab-treated patients, 76 (78%) had detectable and evaluable ctDNA before and after neoadjuvant treatment, and 50 of them (66%) had pre-surgical ctDNA clearance, and 25 out of 50 (50%) had pathologic complete response (pCR). For the placebo-treated group, these values were 64 out 92 (70%) for detectable and evaluable ctDNA before and after neoadjuvant treatment, and 24 out of 64 (38%) had pre-surgical ctDNA clearance, and 3 out of 24 (12%) had pCR. Furthermore, 4 out of 48 (8%) patients in the nivolumab group and 9 out of 44 (20%) in the placebo group who were negative for molecular residual disease (MRD) after surgery and before adjuvant treatment initiation became positive during the adjuvant treatment period; all had disease recurrence. EFS seemed to be prolonged with nivolumab (n = 60) versus placebo (n = 45) in patients with single or co-alterations in any of the KEAP1, STK11, CDKN2A and/or SMARCA4 driver genes (hazard ratio, 0.48; 95% confidence interval, 0.28-0.83). In a machine-learning model trained using biomarker-evaluable patients, top predictors of prolonged EFS included pre-surgical ctDNA clearance, non-N2 NSCLC, pCR, squamous tumour histology and nivolumab treatment. These findings provide insights into predictive markers for outcomes with perioperative nivolumab in resectable NSCLC.

By breaking the optical diffraction limit, super-resolution fluorescence microscopy has advanced our understanding of biological complexity under the framework of light-excited luminescence1. The use of external light excitation remains a key factor that shapes the imaging capabilities and live-cell compatibility of fluorescence-based approaches2. An alternative is the reaction-excited luminescence, such as electrochemiluminescence (ECL)3, chemiluminescence (CL)4 and bioluminescence (BL)5, providing a chemically defined toolbox for enabling different imaging merits, from ultrasensitive analysis6,7 to biocompatible imaging8,9. Despite its light-free excitation and high sensitivity, conventional luminescent-reaction-enabled imaging is fundamentally limited in spatiotemporal resolution owing to low photon budget10,11. Here we develop a chemistry-based super-resolution imaging framework, luminescent-reaction-enabled super-resolution imaging via entropy-weighted correlation combined with deconvolution (RIED). As an experimental-computational concept, RIED introduces a spatiotemporal recording strategy to uncover specific luminescent-reaction-enabled imaging information content, which is efficiently collected and computed to achieve super resolution using a reconstruction strategy adapted to reaction-driven photon statistics. We achieve super-resolution ECL, CL and BL imaging of intracellular organelles, attaining approximately 100 nm resolution. This approach is used for highly sensitive imaging of surface proteins and 41-h ultralong-term continuous super-resolution live-cell imaging of mitochondrial transfer dynamics. Our work establishes an emerging class of chemistry-enabled, laser-free super-resolution microscopy with expanded biological imaging versatilities.

Sustained and controlled delivery of glucose-lowering agents using engineered designer cells is recognized as an effective strategy for diabetes therapy1. However, current technologies rely on external signal control or have been programmed into mammalian cells using synthetic gene networks, which pose safety concerns arising from transplantation2,3. Here we developed an engineered oral-deliverable glucose-sensing and functional response probiotic living drug for 'sense-and-respond'-based control of diabetic blood glucose. We created a glucose sensor based on a synthetic gene circuit that incorporates the glucose-responsive transcriptional regulator HexR, coupled with a synthetic promoter. Upon oral administration of the engineered probiotics carrying the sensor, the cells reside temporarily in the intestine and regulate the expression of therapeutic transgenes in response to glucose levels that exceed the normal threshold. We show efficacy from the engineered probiotics for glycaemic control in multiple diabetic mouse and non-human primate models, demonstrating that long-term oral administration drives clear improvements in lipid profiles, while also attenuating development of multiple diabetic complications. Our probiotics-based living drug enables therapeutic dosing in response to real-time blood glucose levels, providing a programmable, orally deliverable sense-and-respond platform for metabolic therapy without transplantation.

Radial glial progenitors (RGPs) produce all excitatory neurons in the developing cerebral cortex. Mosaic analysis with double markers (MADM)-based lineage tracing in vivo has revealed a quantitative framework of RGP lineage progression1. Here we established MADM technology2,3 in mouse embryonic stem cells to probe RGP lineage progression in a self-organizing cortical organoid system. We found that RGPs exhibit a high level of plasticity in proliferative potential in organoids rather than strict temporally stereotyped lineage progression as observed in vivo. RGPs in organoids showed increased lineage restriction, diminishing cell-type diversity in clones of cortical projection neurons, despite uniform single-cell transcriptional signatures of RGPs and a unitary lineage trajectory. Thus, critical non-cell-autonomous cues that are absent in self-organizing systems and/or the genuine stem cell niche are essential for faithful temporal control of RGP lineage progression and the generation of clonal cortical cell-type diversity.

The dark matter content of ultra-diffuse galaxies (UDGs) is the subject of considerable debate1-5. Stellar streams, which form when a host galaxy tidally strips stars from an orbiting stellar system, provide a powerful technique to constrain the dark matter content of external galaxies6. The stripped stars form long, thin leading and trailing tidal arms that persist for billions of years. Stellar streams from globular clusters (GCs) are particularly sensitive probes of dark matter halos and substructure7-10. GC streams are expected to exist in a variety of host galaxy types11,12 but, so far, they have only been observed in the Milky Way (MW). Here we present evidence for the first, to our knowledge, extragalactic GC stellar stream, identified in deep Hubble Space Telescope (HST) imaging of the UDG UGC 9050-Dw1. The stream's morphology, colour and apparent association with a compact source support the GC progenitor interpretation observationally and we reproduce the observed surface brightness with simulated GC stellar populations. We use generative stream modelling, which fits dynamical models directly to the stream morphology, to constrain the mass of the progenitor and present the first stream-based halo constraint for an UDG. The stream models point to a GC origin and suggest a massive dark matter host halo. By extending the reach of GC stream analysis to external galaxies, this work opens a new chapter in dark matter science.

The world is confronting an escalating crisis of burgeoning photovoltaic (PV) waste1. However, the effectiveness and scalability of prevailing PV waste management approaches remain unclear owing to considerable heterogeneity across regions and over time. Here we develop a comprehensive framework to evaluate the economic and climate benefits of local versus outsourced recycling, covering mainstream technologies. Considering supply-side material constraints, global PV waste will reach 297-402 million tonnes by 2060, with middle-income regions such as China becoming major contributors after 2040. Notwithstanding anticipated technological advancements, the break-even point for global PV waste recycling remains more than a decade away. Combining region-specific recycling technologies with outsourced recycling strategies yields the maximal global net benefits, reducing greenhouse gas emissions by up to 3.32 billion tonnes of CO2-equivalent and generating cumulative net benefits of US$529.1-935.5 billion by 2060. However, outsourced recycling raises inequality concerns for low-income regions. Our results show that a well-designed declining-subsidy scheme effectively mitigates these inequalities, particularly in the early stages. We suggest that regionally adapted recycling strategies and international cooperation, with a focus on technology transfer and funding for recycling capacity in low-income regions, provide effective ways to achieve equitable and scalable PV waste circularity.

Spatial light modulation is a cornerstone of modern photonics. Crucial advances in photonic information processing1-3 rely on the spatial manipulation of the optical phase4-12 and state of polarization (SOP)13-27. The degree of polarization (DOP)28-30 is a fundamental property that can serve as an extra resource. However, no technology exists at present that can spatially modulate the DOP in a programmable manner31-34. Here we demonstrate spatial DOP modulation, thereby achieving control over a new degree of freedom of light. By engineering the polarization statistics at the micrometre scale with a phase-only spatial light modulator, we realize more than 1,024 spatial modes with fully programmable SOP and DOP. This structured light, sculpted in its polarization content to arbitrary shapes, enables direct encoding of information in a high-dimensional space. We encode colour images into a single-wavelength laser by using a one-to-one mapping between the red-green-blue space and the volume of the Poincaré sphere. Polarization colours expand the dimensionality of optical computing and encryption schemes, as demonstrated by (1) fully parallel photonic classification of colour images by a high-dimensional photonic neural network and (2) high-security multidimensional optical encryption. These approaches to optical processing of high-dimensional data highlight the opportunities enabled by DOP modulation in photonics, cryptography and computing.

Bridge recombinases from the IS110 family of transposons, such as IS621, associate with a bridge RNA (bRNA) to mediate programmable recombination between donor DNA and target DNA1,2. Although insertion is mediated by the recombinase-bRNA complex, it remains unknown how IS621 elements are excised from host genomes to form the circular DNA intermediates required for transposition. Here we show that bRNA is weakly expressed from IS621 loci in the Escherichia coli genome and that the IS621 recombinase-bRNA complex mediates excision less efficiently than insertion. Furthermore, we present the cryo-electron microscopy structures of the IS621 recombinase-bRNA complex bound to excision DNA substrates, providing mechanistic insights into the excision reaction. Similar to the previously reported donor- and target-bound insertion complex2, the excision complex comprises two recombinase dimers, each accommodating the target- and donor-binding loops of the bRNA. However, DNA recognition differs notably between the two complexes. Although the donor and target DNAs form a bent U-shape during insertion2, the excision substrates adopt linear conformations and bind across both bRNA loops, forming an X-shaped structure. This geometry reduces the efficiency of top-strand exchange and contributes to the naturally observed bias favouring insertion over excision. Despite these differences, the efficiencies of both reactions are similarly modulated by base pairing between specific dinucleotides in the bRNA, termed handshake guides, and the top strands of the DNA substrates. Overall, this study provides mechanistic insights into the complete IS110 transposition cycle and facilitates the optimal design of programmable bridge-editing applications.

Severe droughts are increasingly driving tree mortality, yet predicting forest resilience remains limited by a lack of understanding of within-species variation in drought resistance1-4. Although hydraulic traits such as embolism resistance are central to drought survival, studies conducted primarily in temperate regions have reported little intraspecific variation5-12, implying that there are evolutionary constraints in the adaptive potential of tree species. Here we test this assumption using a dataset comprising 11 hydraulic traits for 290 trees representing 18 species collected along Puerto Rico's fourfold rainfall gradient (1,000 to 4,000 mm per year). We find that substantial intraspecific variation, together with species turnover, drives coordinated shifts in drought resistance across the gradient. Most species exhibit substantially more embolism resistance and wider stomatal safety margins in drier forests, demonstrating a considerable level of intraspecific variation in tropical trees and a mechanism for adaptation to increasingly dry conditions. Species lacking such trait variability may be highly vulnerable under a future drier climate. Incorporating both interspecific and intraspecific trait variation into predictive models will advance our ability to forecast forest resilience to intensifying droughts.

The retinoblastoma protein (Rb) is a tumour suppressor best known for repressing E2F transcription factors and halting cell cycle progression1. In hormone receptor-positive (HR+) breast cancer, CDK4/6 inhibitors activate Rb by preventing its phosphorylation, forming a key component of current endocrine therapy regimens2. How pharmacologically activated Rb remodels chromatin and influences transcription beyond cell cycle arrest remains poorly understood. Here we show that CDK4/6 inhibition induces redistribution of hypophosphorylated Rb to promoters and enhancers. Although Rb predictably binds to cell cycle gene promoters to repress transcription, at other sites, it unexpectedly promotes expression of oestrogen-responsive genes by integrating into oestrogen receptor (ER)-rich transcriptional hubs. CDK4/6 inhibition enhances ER target gene expression in breast cancer cells, patient-derived xenografts and clinical HR+ breast cancer samples in an Rb-dependent manner. This reprogramming is mediated in part by KDM5A, whose interaction with Rb contributes to gene regulation at these loci. Critically, components of this Rb-driven ER transcriptional program are pro-proliferative. In endocrine-sensitive tumours, this effect can be neutralized with anti-oestrogen therapy, explaining therapeutic synergy. In endocrine-resistant settings such as ESR1-mutant breast cancer, the program persists, limiting the therapeutic efficacy of CDK4/6 inhibition. These findings reframe Rb as a dual-function transcriptional regulator that, although enforcing cell cycle arrest, can also activate programs that counteract its tumour suppressor function.

Tropical-forest aboveground biomass (AGB) is a major component of the global carbon cycle and understanding its response to climate change is crucial for predicting future climate-carbon feedbacks. Yet debate continues as to whether differences between studies in observed associations with climate reflect true regional variation or methodological differences1-5. Here we analyse around 16 million spaceborne-LiDAR-derived estimates of AGB for the year 2020 across intact lowland forests in the Amazon, the Congo Basin and Southeast Asia to investigate how climatic variables differentially relate to AGB on pantropical scales. We show that climatic associations with AGB are heterogeneous and depend on environmental context. Temperature dominates in drier forests, with AGB in the Congo Basin the most sensitive to warming, whereas Southeast Asian forests have the strongest declines in AGB under increasing water limitation. Across regions, the effects of temperature and drought anomalies intensify with aridity and are further modified by soils and topography. In the tallest (above 70 m), carbon-dense forests6,7, storms (lightning and windthrow) emerge as a strong negative driver. These results reconcile previously conflicting findings and show that predicting tropical-forest carbon storage requires accounting for interactions among climate, disturbance, soil and topography in a variety of biogeographical contexts.

The creation of effective governance mechanisms for artificial intelligence (AI) agents requires a deeper understanding of their core properties and the implications they have for deployment. This paper provides a characterization of AI agents that focuses on four dimensions: autonomy, efficacy, goal complexity and generality. We propose different gradations for each dimension and argue that each dimension raises unique questions about the design, operation and governance of these systems. Moreover, we draw on this framework to construct 'agentic profiles' for different kinds of AI agent. These profiles help to illuminate cross-cutting technical and non-technical governance challenges posed by different classes of AI agents, ranging from narrow task-specific assistants to highly autonomous general-purpose systems. By mapping out key axes of variation and continuity across four dimensions, agentic profiles provide developers, policymakers and members of the public with guidance for effective AI governance.

The physical processes that led to the formation of billion-solar-mass black holes within the first 700 million years of cosmic time, a period known as cosmic dawn, remain a puzzle1. Several theoretical scenarios have been proposed to seed and rapidly grow black holes2-4, but direct observations of these mechanisms remain elusive. Here we present a source 660 million years after the Big Bang that exhibits singular properties: among the largest hydrogen Balmer breaks reported at any redshift, broad multi-peaked Hβ emission, and Balmer line absorption in several transitions. We model this source as an enshrouded black hole in which the Balmer break and absorption features are a result of extremely dense, turbulent gas forming a dust-free envelope around a supermassive black hole5,6. This source may provide evidence of an early black hole embedded in dense gas-a theoretical configuration proposed to rapidly grow black holes by super-Eddington accretion7,8. Radiation from the black hole seems to dominate almost all observed light, leaving limited room for contribution from its host galaxy. If the source merged with its brighter neighbour, it would resemble the recently discovered 'little red dots' with perplexing spectral energy distributions9-11. The redness of the black hole is due to gas, not dust12,13, and scattering, not kinematics, gives rise to the complex line shapes and luminosities-black hole masses of these sources may therefore be overestimated by orders of magnitude.

Defence against viral infection is a conserved feature of all cellular life. From single-cell bacteria to humans and complex multicellular animals, constitutive and inducible forms of immunity are required to inhibit viruses and safeguard cellular fitness. Recent studies reveal that the components of human antiviral immunity are surprisingly ancient, originating billions of years ago in bacteria as pathways that defend against phage replication. The unification of previously disparate fields of human and bacterial immunity creates a foundation to explain key features of host-virus interactions. This Review defines principles of pathogen recognition, signal amplification and immune effector function that shape mechanisms of antiviral immunity that are shared across kingdoms of life. Shared forms of immunity, including cGAS-STING, inflammasomes, argonautes and viperin, reveal ancient features of antiviral defence. Similarly, direct comparisons of pattern recognition receptors and interferon-stimulated genes in human cells with CRISPR immunity and anti-phage defence systems in bacteria explain prevalent strategies to effectively sense and inhibit viral replication. Cross-kingdom analysis reveals universal rules that control host-virus interactions and highlights open questions in understanding of antiviral immunity.