Topic: Artificial Intelligence

Artificial IntelligenceNovember 3, 2026

Protein resource allocation across subcellular compartments revealed by large-scale absolute quantitative proteomics in yeast.

Subcellular compartmentalization within complex membrane-bound organelles is a hallmark of eukaryal life, enabling metabolic specialization through spatial segregation. Here, we compiled a dataset encompassing 300 absolute quantitative proteomes, the largest to date, across two yeast species under diverse experimental conditions. Leveraging big data analysis, formula fitting, and machine learning, quantitative correlations among protein abundance, resource allocation, and cellular phenotypes wer

Artificial IntelligenceNovember 3, 2026

Quantifying ancient architecture: Stone-level machine learning reveals architectural variation and settlement growth at Shivta.

Built environments provide durable evidence for how ancient settlements were organized, expanded, and transformed over time. Yet quantitative approaches to documenting architectural variability are generally labor-intensive, limited to individual structures, or dependent on aggregate volumetric estimates. We present a scalable, machine-learning-based framework that shifts the unit of analysis to the individual building stone and combines two complementary datasets: low-altitude drone imagery of

Artificial IntelligenceNovember 3, 2026

In-context learning as a retraining-free paradigm for constitutive modeling.

Constitutive modeling is fundamental to computational solid mechanics, yet most existing approaches remain material-specific and require retraining or recalibration when materials change. Here we introduce In-Context Modeling (ICM), a retraining-free paradigm for constitutive modeling that infers constitutive behavior from physical context rather than material-specific optimization. Using hyperelasticity as a testbed, we demonstrate that ICM can be trained once on equilibrium deformation fields

Artificial IntelligenceOctober 29, 2026

Data-driven discovery and rapid, direct synthesis of MXenes.

MXenes are a rapidly expanding family of two-dimensional transition metal carbides and nitrides, yet scalable synthesis routes remain limited and historical reports suggest a broader, largely unexplored chemical space. We used machine learning-assisted database mining to uncover hidden multilayer MXenes. Screening reveals 33 previously synthesized but unrecognized MXene candidates. Guided by these findings, we synthesized seven multilayer MXenes using a rapid, scalable self-propagating high-temp

Artificial IntelligenceOctober 29, 2026

Intrinsic gestational timing governs human cerebellar development after preterm birth.

Intrinsic programs shape brain maturation; however, which of these programs are perturbed by prematurity as well as the molecular pathways involved remain unknown. The human cerebellum serves as a paradigm of extrauterine development-its accelerated growth and circuit formation align with the third trimester, a period disrupted by preterm delivery. Using multimodal datasets encompassing in vivo neuroimaging and neurodevelopmental outcomes with postmortem spatial-transcriptomic and histopathologi

Artificial IntelligenceOctober 27, 2026

CX3CL1 (fractalkine)-CX3CR1 interface drives CD8⁺ T cell-induced neurodegeneration in human brain organoids.

Immune-neuron interactions play a critical role in neuroinflammatory and neurodegenerative diseases, such as Alzheimer's disease (AD), but how CD8⁺ T cells contribute to brain pathology remains poorly understood. Amyloid-β (Aβ), a hallmark of AD, is known to trigger neuronal stress and inflammation. Here, we established a human cortical organoid (hCO) model to investigate CD8⁺ T cell responses to Aβ-induced neurodegenerative stress. Aβ exposure increased neu

Artificial IntelligenceOctober 27, 2026

Causality in liquid water as a hallmark of emergent glassy dynamics.

In molecular liquids such as water, time-delayed influences between microscopic or mesoscopic variables are typically probed using time-correlation functions, which are symmetric under detailed balance and therefore blind to dynamical asymmetries. Here, we characterize water's dynamics using a causal inference metric that captures asymmetric couplings between collective variables. Analyzing equilibrium molecular dynamics simulations at ambient conditions and in the high-density liquid (HDL) regi

Artificial IntelligenceOctober 23, 2026

Cell-state-specific metabolic networks govern ferroptosis versus apoptosis in small cell lung cancer.

Cellular heterogeneity and plasticity are hallmarks of cancer that contribute to tumor growth and therapy resistance. Here, we investigated metabolic heterogeneity in small cell lung cancer (SCLC), an aggressive cancer type with considerable plasticity between neuroendocrine (NE) and non-NE cellular states. Through integrated transcriptomic and metabolomic analyses, we identified a universal dependency on exogenous cysteine/cystine (Cys) across all NE/non-NE SCLC cell states. Notably, NE and non

Artificial IntelligenceOctober 23, 2026

Trastuzumab Rezetecan in ERBB2-Low Advanced Breast Cancer: The HORIZON-Breast06 Randomized Clinical Trial.

Trastuzumab rezetecan (SHR-A1811) has demonstrated encouraging antitumor activity in a global phase 1 trial involving heavily pretreated patients with erythroblastic oncogene B (ERBB2)-low metastatic breast cancer. To compare trastuzumab rezetecan vs physician's choice of chemotherapy in patients with ERBB2-low metastatic breast cancer. This open-label, randomized, multicenter phase 3 trial was conducted at 73 hospitals across China. Eligible patients were females aged 18 to 75 years with centra

Physical/TechOctober 21, 2026

The anatomy of a personal health agent

Abstract Large language models are enabling a new generation of health agents, yet their ability to address the diverse needs of individuals in everyday health settings remains underexplored 1,2 . Here we developed a comprehensive personal health agent that reasons over multimodal data from consumer wellness devices and medical records to provide personalized health support. Through analysis of web search and health forum queries, together with user-centred research involving end users and healt

Artificial IntelligenceOctober 21, 2026

Programmable design of functional proteins from natural language.

Programming biological function-designing bespoke proteins to perform specified molecular tasks-is a foundational goal of molecular engineering. However, current design models remain fundamentally limited 1-3 : they typically require either natural proteins as starting points for optimization or manual reformulation of functional goals as geometric and sequence-level constraints to guide candidate generation. Here we introduce Pinal, a 16-billion-parameter foundation model that designs candidate

Artificial IntelligenceOctober 21, 2026

A bioavailable GPX4 inhibitor for ferroptosis induction in vivo and tumor suppression.

Ferroptosis is implicated in multiple human diseases, but the lack of bioavailable inducers has limited its physiological study and therapeutic translation. Here, we report KY0418, a selective covalent inhibitor of glutathione peroxidase 4 (GPX4) featuring an isourea warhead, with high plasma stability and broad tissue distribution after oral dosing in mice. KY0418 specifically targets GPX4 SeCys46 to induce ferroptosis hallmarks in human and mouse cancer cell lines, with greater selectivity tha

Artificial IntelligenceOctober 20, 2026

Integrating cooperative breeding with reciprocity and culture extends Hamilton's rule to humans.

We study how kinship and reciprocity shape giving and child care support in societies with limited access to markets. Extending Hamilton's rule to incorporate reciprocity, we test it using behavioral data from dictator games and surveys in remote matrilocal and patrilocal villages of northeast India (N = 406 parents of young children). We replicate with data from urban Manila neighborhoods (N = 301) and the Solomon Islands (N = 624). Consistent with predictions, transfers to recipients increase

Artificial IntelligenceOctober 15, 2026

Multi-omics atlas of chronic exercise.

Exercise reduces all-cause mortality, yet its molecular effects across organs, sexes, and circadian phases remain incompletely understood. We subjected mice to 8 weeks of voluntary wheel running and collected 22 tissues and plasma during active and inactive phases for bulk RNA sequencing, metabolomics, proteomics, and single-nucleus RNA and methylation sequencing. Exercise induced canonical fiber-type switching and widespread, tissue-specific remodeling, with the largest transcriptomic changes i

Artificial IntelligenceOctober 14, 2026

A microexon regulatory code reveals autism-linked genetic variation.

Neuronal microexons are a highly conserved class of alternative splicing events enriched in genes with critical functions in nervous system biology 1,2 . Microexon splicing is often disrupted in the brains of autistic individuals, and mice with microexon deficiencies display autism-like phenotypes 1,3,4 . However, the full landscape of microexons is not known, and the impact of genetic variation on these exons has not been systematically investigated. Here, through extensive transcriptome profil

Artificial IntelligenceOctober 13, 2026

Deep learning of fossil pollen morphology reveals 25,000 y of ecological change in eastern African grasslands.

Grass (Poaceae) pollen is largely overlooked in investigations of grassland evolution because the pollen of most species cannot be differentiated using traditional optical microscopy. However, the combination of superresolution microscopy and deep learning enables the capture and quantification of distinct morphological variation across the pollen of grass species. Using a semisupervised deep-learning strategy, we trained convolutional neural networks (CNNs) on superresolution images of known mo

Artificial IntelligenceOctober 13, 2026

Tyrosine phosphorylation unfolds nucleophosmin and disrupts its integration into the nucleolus.

Nucleophosmin (NPM1) is a multifunctional nucleolar protein essential for ribosome biogenesis, genome stability, and stress responses. Its integration into the nucleolus depends on its oligomerization and multivalent interactions that enable liquid-liquid phase separation (LLPS). Here, we investigate how tyrosine phosphorylation at Tyr17, Tyr29, and Tyr67 located within the interface between monomers at the N-terminal oligomerization domain regulates NPM1 structure and function. Replacing tyrosi

Artificial IntelligenceOctober 13, 2026

Human-artificial intelligence collaboration for comanaging epilepsy in real time.

One-third of the world's 70 million people with epilepsy have seizures uncontrolled by medication, and implantable devices are a promising treatment. These devices can detect impending seizures, missed medication and impaired cognition, but, like most medical devices, they cannot share this information with patients in real time. This is a missed opportunity for devices to learn from clinical events and teach patients how their behaviors affect their health. Here we present a device platform tha

Artificial IntelligenceOctober 12, 2026

A health-system-scale, episode-resolved multimodal atlas of cancer cachexia.

Cancer cachexia is a wasting syndrome with outsized impact on morbidity and mortality, yet its drivers and systemic consequences remain poorly defined. In this study, we repurposed longitudinal clinicogenomic data from 52,236 patients with cancer profiled by MSK-IMPACT (2010-2024) at Memorial Sloan Kettering Cancer Center to define episode-resolved trajectories of consensus-defined cachectic weight loss and linked these events to outcomes, serology and tumor genotypes. Cachectic episodes cluster

Physical/TechOctober 12, 2026

Multidimensional vulnerabilities of informal settlements in the Urban South

Informal settlements are widespread and rapidly growing, often associated with high vulnerabilities. However, global-scale understanding on their spatial distribution and complexity of vulnerability is limited. Here, we delineate the informal settlements across a sample of 1,914 cities of the Global South by integrating Sentinel-2 imagery with deep learning models, and we explore the heterogeneity of vulnerability using Urban Informal Settlement Vulnerability Index that encompasses living condit

Artificial IntelligenceOctober 8, 2026

Shared patterns of human milk composition link mammary gland function to infant growth.

The mammary gland produces nutrient-rich human milk (HM), yet how mammary functional state shapes HM composition and infant outcomes remains poorly understood. We used HM multi-omics - metabolomics, proteomics, micronutrients, macronutrients, and HM oligosaccharides - across 1,543 samples from three cohorts, including two randomized trials, as a noninvasive readout of mammary functional state. Trajectory modeling and multi-omic integration showed that maternal supplementation improved recovery f

Artificial IntelligenceOctober 8, 2026

Conversational diagnostic artificial intelligence in ambulatory primary care: a prospective feasibility study.

Artificial intelligence (AI)-based systems show promise for assisting primary care providers (PCPs) with patient care. We aimed to evaluate the safety and quality of clinical conversations of a patient-facing conversational AI system, which engaged in real-world urgent primary care appointments. In this prospective, single-centre, single-arm feasibility study, English-speaking patients aged at least 18 years interacted with the Articulate Medical Intelligence Explorer (AMIE) up to 5 days before

Artificial IntelligenceOctober 8, 2026

Advancing mathematics research with AI-driven formal proof search

Large language models (LLMs) increasingly excel at mathematics tasks, but their unreliability limits their utility in mathematics research. A mitigation is to use LLMs to generate formal proofs in languages such as Lean, in which the compiler verifies every proof step. We present the first demonstration of this method’s value in solving open problems at scale. We built an artificial intelligence agent for formal proof search that autonomously resolved nine of 353 open Erdős problems, proved 44/4

Artificial IntelligenceOctober 6, 2026

Large-scale semidefinite programming with graphics processing units.

Semidefinite programming (SDP) provides a powerful framework in applied mathematics with applications spanning optimization, machine learning, quantum computing, and beyond. However, the computational cost of solving large-scale SDP problems remains a significant practical limitation. We break this long-standing computational bottleneck through a synergistic codesign of low-rank algorithms and graphics processing unit (GPU) architectures, developing accelerated first-order methods that leverage

Artificial IntelligenceOctober 6, 2026

Genome stability disrupted by the circRBM39(4,5,6)-RPA2 axis represses breast tumorigenesis.

Genome instability (GI) is a hallmark of cancer. GI is associated with accumulative DNA damage and tumor-specific defects in DNA repair. Multiple antitumor drugs have been developed to promote persistent GI and to drive DNA damage beyond a threshold that tumor cells can survive. However, regulatory roles of circular RNAs (circRNAs) in GI and DNA damage remain elusive. Through circRNA profiling and a small interfering RNA (siRNA)-mediated screen, we have identified a circRNA termed circRBM39(4,5,

Artificial IntelligenceOctober 6, 2026

α-aminobutyric acid promotes tumor progression by polarizing macrophages through the ASNS-mTORC2 axis.

Tumor-associated macrophages (TAMs), the most abundant immune cell subset in the tumor microenvironment (TME), exhibit phenotypic plasticity and exert critical roles in tumor progression and antitumor immunity. Targeting TAM polarization has emerged as a promising strategy for cancer immunotherapy, yet the key regulators governing this process remain incompletely defined. Here, we identified α-aminobutyric acid (AABA) as a pro-tumor metabolite that drives M2-like polarization of TAMs to pr

Artificial IntelligenceOctober 6, 2026

An open vision-language model for diverse medical applications

Abstract Artificial intelligence has high potential for impact in healthcare applications, but its training and deployment are challenging due to diverse data, a complex spectrum of possible tasks and important privacy needs. High-performing foundation models that enable data-efficient fine-tuning for diverse downstream tasks can meaningfully accelerate development in this domain. We introduce MedGemma, a collection of medical vision-language foundation models based on Gemma 3. MedGemma demonstr

Artificial IntelligenceOctober 1, 2026

Keep AI-augmented science contestable

Artificial intelligence (AI) is moving upstream in science. Systems such as AlphaFold have already transformed protein-structure prediction, while newer agentic systems can search literatures, generate hypotheses, write and execute code, interpret results, and automate much of the path from idea to paper. Recent debates have asked how such systems may change scientific progress, surprise, and human understanding. Yet they also change something more fundamental: who—or what—participates in decidi

Artificial IntelligenceSeptember 29, 2026

Plant fiber exploitation contributed to the emergence of ground-edged cutting tools in North China.

Ground-edged cutting tools are widely regarded as technological hallmarks of early agriculture in North China, commonly hypothesized to have been developed primarily for cereal harvesting. Yet the functional foundations of this innovation remain insufficiently tested. This study integrates use-wear and microfossil analyses of 140 stone tools from the Peiligang site, with experimental data, to reassess their roles within long-term trajectories of technological change from the Late Paleolithic to

Artificial IntelligenceSeptember 29, 2026

Transient chaos in the self-organization of dissipative optical solitons.

Transient chaos is a hallmark of complex dynamics in nonlinear systems. Governed by chaotic saddles, it manifests as short-lived yet rich chaotic behavior preceding an abrupt transition to a stable attractor. However, real-time capture and quantitative characterization of such inherently unpredictable and nonrepetitive dynamics remain challenging, obscuring their physical origins. Here, we experimentally demonstrate transient chaos during the self-organization of dissipative optical solitons in