Search

A genome language model, Protein Set Transformer, trained on viral datasets, uncovers evolutionary rules of protein content and organization driving precise virus identification, host prediction, and protein annotation for viral genomics and ecology.

For single-cell RNA-seq experiments the sequencing budget is limited, and how it should be optimally allocated to maximize information is not clear. Here the authors develop a mathematical framework to show that, for estimating many gene properties, the optimal allocation is to sequence at the depth of one read per cell per gene.

Neural mechanisms underlying reinforcement learning in naturalistic environments are not fully understood. Here authors show that reinforcement learning (RL) agents with hippocampal-like recurrence, unlike feedforward networks, match animal behaviour and neural data in navigation tasks, revealing that hippocampal circuits support RL in naturalistic environments.

Typical quantum error correcting codes assign fixed roles to the underlying physical qubits. Now the performance benefits of alternative, dynamic error correction schemes have been demonstrated on a superconducting quantum processor.

In many machine learning applications, one uses pre-trained neural networks, having limited access to training and test data. Martin et al. show how to predict trends in the quality of such neural networks without access to this information, relevant for reproducibility, diagnostics, and validation.

Imaging pancreatic islets in living animals is often limited by poor optical access and motion instability. Here, the authors show that transplanting islets onto the brain’s dura mater enables stable, long-term, and functional imaging in awake mice.

The spontaneous emergence of autocatalytic RNAs is central to origin of life. Here, the authors use machine learning, high-throughput screening and statistical physics to explore large neutral space of catalytic RNAs.

In superconducting thin films of a few atomic layers, surface and bulk orders may compete. Here, the authors measure how the Pearl length, which characterizes the distribution of magnetic field in vortices, varies with thickness in NbSe₂, showing that surface superconductivity dominates at low thickness.

Peer review information: Nature Communications thanks Martino Poggio, and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

An analysis of habitat fragmentation using a dataset of more than 4,000 species worldwide shows that fragmentation reduces biodiversity at all scales, and that increases in β diversity do not compensate for the loss of α diversity.

PRS-CSx is a polygenic risk score construction method that improves cross-population polygenic prediction by integrating GWAS summary statistics from multiple populations.

Chen et al. report that TGF-β signalling, although largely considered anti-inflammatory, has proinflammatory effects on endothelial cells. Inhibition of endothelial TGF-β signalling decreases atherosclerosis in mice and reverts established plaques, in part by decreasing endothelial-to-mesenchymal transitions.

Tensor networks exploit the structure of turbulence to offer a compressed description of flows, which leads to efficient fluid simulation algorithms that can be implemented on both classical and quantum computers.

The authors collate literature on the responses of bird assemblages to forest loss and show that locations with a more variable natural environment and a longer history of agricultural land use have bird assemblages that are more tolerant to forest loss.

Nuclear organization impacts allelic exclusion. Here, the authors describe a self-limiting protein bridge that connects two chromosomes, and two transcription and splicing compartments, to maintain monogenic VSG expression in the African trypanosome.

A hybrid analogue–digital quantum simulator is used to demonstrate beyond-classical performance in benchmarking experiments and to study thermalization phenomena in an XY quantum magnet, including the breakdown of Kibble–Zurek scaling predictions and signatures of the Kosterlitz–Thouless phase transition.

Stabl selects sparse and reliable biomarker candidates from predictive models.

VirtualMultiplexer is a generative AI tool that produces realistic multiplexed immunohistochemistry images from tissue biopsies. The generated images could be used to improve clinical predictions, enhancing histopathology workflows and accelerating cancer research.

Inspired by insects in nature, the authors develop a neuromorphic robotic system with obstacle avoidance, tunnel centering and gap crossing capabilities. Their robotic system accomplishes these multiple capabilities by steering towards regions of low apparent motion.

During mitosis, mitochondria partition into daughter cells through microtubule-based transport. Here the authors show that the mitochondrial protein Miro and the cytoskeletal-associated protein Cenp-F interact in a cell-cycle dependent manner to promote microtubule-directed movement of mitochondria.

SAMHD1 is a cellular dNTPase proposed to inhibit HIV-1 reverse transcription in non-cycling immune cells by limiting dNTP substrate supply; its anti-viral but not dNTPase function is downregulated by phosphorylation of T592. Here, Martinat et al. describe an additional SUMOylation at residue K595, which promotes the dNTPase-independent restriction activity.

Entanglement between single photons and solid-state emitters is a key component for photonic quantum computing and networks. Here, using a single electron spin in a quantum dot, the authors present a deterministic photon source achieving three-qubit entanglement of one electron spin and two photons.

Lamkanfi, Shirasaki, Izawa and colleagues show that IL-1β release in human monocytes stimulated with lipopolysaccharide or carrying NLRP3 activatory mutations is mostly restricted to cells undergoing lytic cell death.

Rheumatoid arthritis (RA) mainly impacts the joints, which are difficult to sample, so identifying peripheral biomarkers may aid in diagnosis and treatment response predictions. Here the authors pursue a longitudinal study profiling plasma protein and use machine learning to train an algorithm for predicting RA onsets and responses to antirheumatic drugs.

A new version of nanorate DNA sequencing, with an error rate lower than five errors per billion base pairs and compatible with whole-exome and targeted capture, enables epidemiological-scale studies of somatic mutation and selection and the generation of high-resolution selection maps across coding and non-coding sites for many genes.

Tandem mass spectroscopy is a useful tool to identify metabolites but is limited by the capability of computational methods to annotate peaks with chemical structures when spectra are dissimilar to previously observed spectra. Goldman and colleagues use a transformer-based method to annotate chemical structure fragments, thereby incorporating domain insights into its architecture, and to simultaneously predict the structure of the metabolite and its fragments from the spectrum.

This protocol describes a detailed method for superresolution imaging of plant tissues by structured illumination microscopy (SIM). Details include microscope calibration, tissue preparation, image acquisition and evaluation of SIM images.

Super-resolution microscopy is a valuable tool in bioimaging, but often requires complex systems or post-processing. Here, the authors present super-linear excitation-emission (SEE) microscopy, which overcomes these limitations by taking advantage of markers with super-linear dependence between emission and excitation power.

Viruses exhibit complex interactions and physical properties, yet their behavior in confined spaces remains largely unexplored. Authors develop a real-time detection method for single virus transport through synthetic nanopores, revealing a jamming phenomenon due to viral interactions and confinement

An analysis of 24,202 critical cases of COVID-19 identifies potentially druggable targets in inflammatory signalling (JAK1), monocyte–macrophage activation and endothelial permeability (PDE4A), immunometabolism (SLC2A5 and AK5), and host factors required for viral entry and replication (TMPRSS2 and RAB2A).

Several unusual features have been reported for bacteria of the phylum Planctomycetes, such as cytosolic compartmentalization and an endocytosis-like process. Here, Boedekeret al. provide evidence supporting a Gram-negative cell plan and the absence of endocytosis-like processes in these organisms.

Neural mechanisms underlying state-dependent flexible selection are not fully understood. Here authors show that NPY homologues in Drosophila larva differentially modulate reciprocally connected inhibitory neurons to bias non-feeding decisions, favoring escape-type actions (Head Cast), over protective-type actions (Hunch), in response to a mechanical cue.

Lavdovskaia, Hanitsch, Linden et al. provide a comprehensive roadmap of mitoribosome biogenesis and establish that mitochondria use a unique pathway for the assembly of their translation machinery.

In this work, the authors report NMR lipids Databank to promote decentralised sharing of biomolecular molecular dynamics (MD) simulation data with an overlay design. Programmatic access enables analyses of rare phenomena and advances the training of machine learning models.

Active fluids exhibit regimes with a complex spatio-temporal structure reminiscent of inertial turbulence. Now, inertial and active turbulence are theoretically shown to be closely related indeed.

Individuals with schizophrenia show reduced structural similarity in temporal, cingulate, and insular lobes, especially those with worse cognition and symptoms, affecting late maturing association areas with low metabolism and high neurotransmission.

This updated analysis of the Cell Tracking Challenge explores how algorithms for cell segmentation and tracking in both 2D and 3D have advanced in recent years, pointing users to high-performing tools and developers to open challenges.

The relationship between new greenspaces and gentrification is an important one for urbanization. Here the authors show a positive relationship for at least one decade between greening in the 1990s–2000s and gentrification that occurred between 2000–2016 in 17 of 28 studied cities in North America and Europe.

H2B.8 is identified as a histone variant that mediates a newly described mechanism of transcription-compatible chromatin condensation in flowering plant sperm cells.

Understanding deregulation of biological pathways in cancer can provide insight into disease etiology and potential therapies. Here, as part of the PanCancer Analysis of Whole Genomes (PCAWG) consortium, the authors present pathway and network analysis of 2583 whole cancer genomes from 27 tumour types.

The authors present SVclone, a computational method for inferring the cancer cell fraction of structural variants from whole-genome sequencing data.

Cancers evolve as they progress under differing selective pressures. Here, as part of the ICGC/TCGA Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium, the authors present the method TrackSig the estimates evolutionary trajectories of somatic mutational processes from single bulk tumour data.

In genome-wide association meta-analysis, it is often difficult to find an independent dataset of sufficient size to replicate associations. Here, the authors have developed MAMBA to calculate the probability of replicability based on consistency between datasets within the meta-analysis.

STAARpipeline is a comprehensive framework for flexible and scalable rare-variant association analysis using whole-genome sequencing data and annotation information.

Entanglement was observed in top–antitop quark events by the ATLAS experiment produced at the Large Hadron Collider at CERN using a proton–proton collision dataset with a centre-of-mass energy of √s  = 13 TeV and an integrated luminosity of 140 fb⁻¹.

Although progress in the coverage of routine measles vaccination in children in low- and middle-income countries was made during 2000–2019, many countries remain far from the goal of 80% coverage in all districts by 2019.

A microphysiological cartilage on a chip that enables the application of strain-controlled compression recapitulates the mechanical factors involved in the pathogenesis of osteoarthritis.