Hepatocellular carcinoma (HCC) is the fastest growing cause of cancer-related mortality and there are limited therapies1. Although endoplasmic reticulum (ER) stress and the unfolded protein response (UPR) are implicated in HCC, the involvement of the UPR transducer ATF6α remains unclear2. Here we demonstrate the function of ATF6α as an ER-stress-inducing tumour driver and metabolic master regulator restricting cancer immunosurveillance for HCC, in contrast to its well-characterized role as an adaptive response to ER stress3. ATF6α activation in human HCC is significantly correlated with an aggressive tumour phenotype, characterized by reduced patient survival, enhanced tumour progression and local immunosuppression. Hepatocyte-specific ATF6α activation in mice induced progressive hepatitis with ER stress, immunosuppression and hepatocyte proliferation. Concomitantly, activated ATF6α increased glycolysis and directly repressed the gluconeogenic enzyme FBP1 by binding to gene regulatory elements. Restoring FBP1 expression limited ATF6α-activation-related pathologies. Prolonged ATF6α activation in hepatocytes triggered hepatocarcinogenesis, intratumoural T cell infiltration and nutrient-deprived immune exhaustion. Immune checkpoint blockade (ICB)4 restored immunosurveillance and reduced HCC. Consistently, patients with HCC who achieved a complete response to immunotherapy displayed significantly increased ATF6α activation compared with those with a weaker response. Targeting Atf6 through germline ablation, hepatocyte-specific ablation or therapeutic hepatocyte delivery of antisense oligonucleotides dampened HCC in preclinical liver cancer models. Thus, prolonged ATF6α activation drives ER stress, leading to glycolysis-dependent immunosuppression in liver cancer and sensitizing to ICB. Our findings suggest that persistently activated ATF6α is a tumour driver, a potential stratification marker for ICB response and a therapeutic target for HCC.

Anthropogenic emissions of hydrogen (H2) are expected to rise if H2 energy technology is widely implemented as part of the green energy transition1,2. Although atmospheric H2 is not radiatively active, it warms the Earth's climate through chemical effects on methane, ozone and water vapour1-6. Predicting the atmospheric response to anthropogenic perturbations is challenging, in part because of the limited duration of the modern instrumental record7. Ice core measurements of H2 can extend the observational record, providing information about anthropogenic and natural perturbations and the biogeochemical controls on H2 levels over long timescales. However, ice core measurements of H2 are challenging because of the high permeability of H2 in ice8,9. Here we present an ice core record of atmospheric H2 recovered from a Greenland ice core, spanning the past millennium. The record shows a 70-111% (2σ) rise in atmospheric H2 from the pre-industrial to the modern era, consistent with increasing direct emissions from fossil fuel burning and increased atmospheric concentrations of H2 precursors. The pre-industrial record also shows a 4-25% (2σ) decrease in H2 levels during the Little Ice Age (LIA), indicating that H2 biogeochemistry may be sensitive to climate change. The findings suggest that the sensitivity of H2 sources and sinks to climate warming should be considered in estimates of the radiative consequences of rising anthropogenic H2 emissions.

Body-brain communication has emerged as a key regulator of tissue homeostasis1-5. Solid tumours are innervated by different branches of the peripheral nervous system and increased tumour innervation is associated with poor cancer outcomes6-8. However, it remains unclear how the brain senses and responds to tumours in peripheral organs, and how tumour-brain communication influences cancer immunity. Here we identify a tumour-brain axis that promotes oncogenesis by establishing an immune-suppressive tumour microenvironment. Combining genetically engineered mouse models with neural tracing, tissue imaging and single-cell transcriptomics, we demonstrate that lung adenocarcinoma induces innervation and functional engagement of vagal sensory neurons, a major interoceptive system connecting visceral organs to the brain. Mechanistically, Npy2r-expressing vagal sensory nerves transmit signals from lung tumours to brainstem nuclei, driving elevated sympathetic efferent activity in the tumour microenvironment. This, in turn, suppresses anti-tumour immunity via β2 adrenergic signalling in alveolar macrophages. Disruption of this sensory-to-sympathetic pathway through genetic, pharmacological or chemogenetic approaches significantly inhibited lung tumour growth by enhancing immune responses against cancer. Collectively, these results reveal a bidirectional tumour-brain communication involving vagal sensory input and sympathetic output that cooperatively regulate anti-cancer immunity; targeting this tumour-brain circuit may provide new treatments for visceral organ cancers.

Scientific progress depends on the ability of researchers to synthesize the growing body of literature. Can large language models (LLMs) assist scientists in this task? Here we introduce OpenScholar, a specialized retrieval-augmented language model (LM)1 that answers scientific queries by identifying relevant passages from 45 million open-access papers and synthesizing citation-backed responses. To evaluate OpenScholar, we develop ScholarQABench, the first large-scale multi-domain benchmark for literature search, comprising 2,967 expert-written queries and 208 long-form answers across computer science, physics, neuroscience and biomedicine. Despite being a smaller open model, OpenScholar-8B outperforms GPT-4o by 6.1% and PaperQA2 by 5.5% in correctness on a challenging multi-paper synthesis task from the new ScholarQABench. Although GPT-4o hallucinates citations 78-90% of the time, OpenScholar achieves citation accuracy on par with human experts. OpenScholar's data store, retriever and self-feedback inference loop improve off-the-shelf LMs: for instance, OpenScholar-GPT-4o improves the correctness of GPT-4o by 12%. In human evaluations, experts preferred OpenScholar-8B and OpenScholar-GPT-4o responses over expert-written ones 51% and 70% of the time, respectively, compared with 32% for GPT-4o. We open-source all artefacts, including our code, models, data store, datasets and a public demo.

Bacteria have evolved a wide array of defence systems to combat phage infection, many of which rely on complex signalling systems and large protein complexes to function1. Here we describe a 164-residue prophage-encoded protein that defends bacteria by sensing conserved oligomeric components of phage assembly. This protein, called ring interacting pore 1 (Rip1), is activated by the portal or small terminase proteins of infecting phages-oligomeric ring-shaped complexes that are essential for virion maturation. Rip1 uses these phage protein ring complexes as a template to assemble into membrane-disrupting pores that inhibit phage virion assembly and cause premature death of the host cell. Rip1 homologues are widely distributed across bacteria and provide robust defence against diverse phages. This study reveals a strategy by which a small defence protein integrates both sensing and effector activity by exploiting a conserved feature of viral assembly. The mechanism mirrors eukaryotic pore-forming immunity but is executed by a single protein, offering an evolutionarily streamlined solution to viral detection and defence.

Electrons in solids owe their properties to the periodic potential landscapes they experience. The advent of moiré lattices has revolutionized our ability to engineer such landscapes on nanometre scales, leading to numerous ground-breaking discoveries. Despite this progress, direct imaging of these electrostatic potential landscapes remains elusive. Here we introduce the atomic single electron transistor (SET), a new scanning probe that uses a single atomic defect in a van der Waals material as an ultrasensitive, high-resolution potential sensor. Built on the quantum twisting microscope (QTM) platform1, this probe leverages the capability of the QTM to form a pristine, scannable two-dimensional interface between vdW heterostructures. Using the atomic SET, we present the first direct images of the electrostatic potential in a canonical moiré interface: graphene aligned to hexagonal boron nitride2-10. The measured potential exhibits an approximate C6 symmetry, minimal dependence on carrier density and a substantial amplitude of approximately 60 mV, even in the absence of carriers. Theory indicates that this symmetry arises from a delicate interplay of physical mechanisms with competing symmetries. The measured amplitude significantly exceeds theoretical predictions, suggesting that current understanding may be incomplete. With 1 nm spatial resolution and sensitivity to detect the potential of even a few millionths of an electron charge, the atomic SET enables ultrasensitive imaging of charge order and thermodynamic properties across a wide range of quantum phenomena, including symmetry-broken phases, quantum crystals, vortex charges and fractionalized quasiparticles.

Chemically reactive microbial natural products have enabled therapeutic development1 via their well-established anticancer, antibiotic and antioxidant activities. However, discovery of reactive metabolites is particularly challenging because they may not tolerate traditional bioactivity-guided isolation workflows2. Diazo-containing natural products are a subset of highly reactive microbial metabolites that display potent bioactivity3 and enable powerful biosynthetic transformations4,5; however, instability of the diazo group to light6, heat7, mild acid8 and mechanical shock9 has precluded their efficient discovery and application. Here we develop a reactivity-based screening approach to capture diazo-containing metabolites and facilitate their discovery by mass spectrometry. This workflow revealed two novel diazo-containing natural products, 4-diazo-3-oxobutanoic acid (1) and diazoacetone (2), from the human lung pathogen Nocardia ninae. Biosynthetic investigations revealed a distinct enzymatic logic for diazo formation involving hydrazone oxidation catalysed by the metalloenzyme Dob3, and its biochemical characterization suggests promising future applications in biocatalysis. Overall, our work highlights the power of reactivity-guided strategies for identifying reactive metabolites and facilitating the discovery of unique enzymatic transformations.

Fractionalization of the electron charge e is one of the most striking phenomena arising from strong electron-electron interactions. A celebrated example is the emergence of anyons with fractional charges in fractional quantum Hall effect (FQHE) states1-13. Recently, zero-field fractional Chern insulators (FCIs)14-19, lattice analogues of the FQHE states that form without Landau levels, have been realized20,21. FCIs provide a unique platform to investigate anyons, yet their detection remains a challenge. Here we report the observation of anyon-trions, a new type of excitonic complex formed by binding a trion with a fractional charge in twisted MoTe2 bilayers. Photoluminescence spectroscopy of quantum-confined excitons reveals emergent peaks that appear only within slightly doped FCI states. The new spectral features are red-shifted relative to the trions in undoped FCIs, but share the same electric field, temperature and magnetic field dependence. These observations suggest their origin as trions binding with elementary quasi-particles, that is, anyon-trions. Crucially, the ratio of binding energies between the anyon-trions in the -2/3 and -3/5 FCI states matches the expected fractional charge ratio of e/3 to e/5. This provides strong evidence for fractional charges in FCI-an essential property of anyons. Our results address a fundamental question in FCI physics and establish trion spectroscopy as a powerful probe of fractionally charged excitations, complementary to transport- and tunnelling-based approaches.

The superconducting gap defines the fundamental energy scale for the emergence of dissipationless transport and collective phenomena in a superconductor1-3. In layered high-temperature cuprate superconductors, in which the Cooper pairs are confined to weakly coupled two-dimensional (2D) copper-oxygen (CuO2) planes4,5, terahertz (THz) spectroscopy at subgap millielectronvolt (meV) energies has provided crucial insights into the collective superfluid response perpendicular to the superconducting layers6-9. However, within the CuO2 planes, the collective superfluid response manifests as plasmonic charge oscillations at energies far exceeding the superconducting gap, obscured by strong dissipation2,6,9,10. Here we present spectroscopic evidence of a below-gap, 2D superfluid plasmon in few-layer Bi2Sr2CaCu2O8+x and spatially resolve its deeply subdiffractive THz electrodynamics. By placing the superconductor in the near field of a spintronic THz emitter, we reveal this distinct resonance-absent in bulk samples and observed only in the superconducting phase-and determine its plasmonic nature by mapping the geometric anisotropy and dispersion. Crucially, these measurements offer a direct view of the momentum-dependent and frequency-dependent superconducting transition in two dimensions.

Parkinson's disease (PD) is an incurable neurological disorder that often begins insidiously with sleep disturbances and somatic symptoms, progressing to whole-body motor and cognitive symptoms1-5. Dysfunction of the somato-cognitive action network (SCAN)-which is thought to control action execution6,7 by coordinating arousal, organ physiology and whole-body motor plans with behavioural motivation-is a potential contributor to the diverse clinical manifestations of PD. To investigate the role of the SCAN in PD pathophysiology and treatments (medications, deep-brain stimulation (DBS), transcranial magnetic stimulation (TMS) and MRI-guided focused ultrasound stimulation (MRgFUS)), we built a large (n = 863), multimodal, multi-intervention clinical imaging dataset. Resting-state functional connectivity revealed that the substantia nigra and all PD DBS targets (subthalamic nucleus, globus pallidus and ventral intermediate thalamus) are selectively connected to the SCAN rather than to effector-specific motor regions. Importantly, PD was characterized by specific hyperconnectivity between the SCAN and the subcortex. We therefore followed six PD cohorts undergoing DBS, TMS, MRgFUS and levodopa therapy using precision resting-state functional connectivity and electrocorticography recording. Efficacious treatments reduced SCAN-to-subcortex hyperconnectivity. Targeting the SCAN instead of effector regions doubled the efficacy of TMS treatments. Focused ultrasound treatment benefits increased when the target was closer to the thalamic SCAN sweet spot. Thus, SCAN hyperconnectivity is central to PD pathophysiology and its alleviation is a hallmark of successful neuromodulation. Targeting functionally defined subcortical SCAN nodes may improve existing therapies (DBS, MRgFUS), whereas cortical SCAN targets offer effective non-invasive or minimally invasive neuromodulation for PD.

The first billion years of Earth history witnessed the emergence of continental magmatism, oceans and life. Yet, the details of how continents formed remain unknown because of the absence of preserved rocks1-8. Two conflicting Hadean models predominate: early onset of subduction and plate tectonics2-4, compared with early stagnant-lid and plume processes with delayed (post-Hadean) plate tectonics5-7. Here we report trace-element ratios (including Nb-Sc-U-Yb) correlated with age and hafnium and oxygen isotope ratios for Hadean detrital zircons from the Jack Hills (JH), Western Australia, which record unprecedented insights into the timing and setting of early magmatism. More than 70% of Hadean JH detrital zircons have Sc/Yb > 0.1, and 47% have U/Nb > 20, fingerprints for continental-arc and subduction settings. The remainder are ocean-island-like with little evidence for ocean-ridge settings. Hadean JH zircons probably originated from distinct terranes with separate tectonic histories. Subduction-related magmatism in the Hadean, as documented by JH zircons, alternated with periods of magmatic quiescence. This contrasts with dominantly stagnant-lid-like signatures for most Barberton Hadean zircons. The diverse settings for Jack Hills and Barberton detrital zircons imply contemporaneous operation of different tectonic styles during the Hadean, as well as a broader diversity of early crustal origins than previously known.

Deep mantle melting marks the onset of Earth differentiation1, yet a unifying framework for how buoyancy-driven mantle upwellings initiate melting and how such incipient melts evolve within the asthenosphere has remained elusive. Here we show that the first melts generated in any solid-state mantle upwelling are kimberlitic CO2-rich silicate melts that form at about 250 km depth through oxidation of elemental carbon to CO2 (refs. 2,3). Our experiments force a range of surface melts, derived from mantle plumes4 or broad upwellings5 (kimberlites, ocean island basalts and mid-ocean ridge basalts), into equilibrium with fertile mantle at adiabatic and super-adiabatic conditions at 7 GPa. The results define a framework in which redox melting at depth universally yields kimberlitic melts, which, while ascending through the asthenosphere by reactive porous flow6,7, evolve to higher degrees of melting, lesser volatiles and incompatible elements, but higher SiO2. Channelized flow7 in the lithosphere may then enable direct extraction of these melts, leading to kimberlites, where the lithosphere commences just above the C → CO2 redox front, to alkaline Si-undersaturated intraplate magmas where lithospheric thicknesses are 150-100 km, and to tholeiitic basalts below mid-ocean ridges where voluminous 'dry' melting becomes overwhelming. This framework is consistent with the widespread seismic low-velocity zone at about 250 km beneath mid-ocean ridges8,9 and aligns with ocean island and mid-ocean ridge basalts sampling the various geochemical mantle components at different degrees of melting in different proportions10,11.

Thetis cells (TCs) are a recently identified lineage of RORγt+ antigen-presenting cells comprising four subsets, TC I to TC IV, including a tolerogenic subset (TC IV) that instructs tolerance to gut microbiota and food antigens1-6. A developmental wave of TCs during early life creates a crucial window of opportunity for establishing intestinal tolerance1,5. The ontogeny of TCs and the cues that shape their abundance and heterogeneity remain unknown, however, limiting efforts to harness their therapeutic potential. Here we identify a population of RORγt+ progenitors, termed Thetis-lymphoid tissue inducer progenitors (TLPs), that give rise to the immediate TC progenitor (TCP) and the lymphoid tissue inducer (LTi) progenitor (LTiP), and identify PU.1 as the transcription factor that governs TC fate. Despite transcriptional similarity to myeloid-derived conventional dendritic cells, we show that TCs descend from the common lymphoid progenitor. Deletion of the plasmacytoid dendritic cell (pDC) lineage-determining transcription factor TCF4 expands TLPs and TCs, indicating a shared developmental branch with pDCs. TLPs are enriched in fetal liver, but, unlike LTi cells, TCs emerge postnatally, indicating that developmentally timed environmental cues promote TCP differentiation. We identify one such cue, RANKL provision by lymphoid tissue organizer cells, that is essential for TC I differentiation. Together, these findings define the ontogeny of TCs and the transcription factors that promote TC differentiation and heterogeneity, facilitating future investigations of these enigmatic cells and their therapeutic potential for tolerance induction in food allergy and autoimmunity.