Browse Articles

MultistageOT: Multistage optimal transport infers trajectories from a snapshot of single-cell data

Single-cell RNA-sequencing captures a temporal slice, or a snapshot, of a cell differentiation process. A major bioinformatical challenge is the inference of differentiation trajectories from a single snapshot, and methods that account for outlier cells that are unrelated to the differentiation process have yet to be established. We present MultistageOT ( https://github.com/dahlinlab/MultistageOT ), a generalized optimal transport-based framework that models cell differentiation in a single snapshot as a series of intermediate cell transitions. MultistageOT employs multiple transport stages to establish temporal progression within the snapshot—overcoming limitations with the classic bimarginal formulation of optimal transport. Moreover, our multistage framework uses global information across all cells and differentiation stages to infer coherent trajectories from initial to terminal states. This allows MultistageOT to infer individual outlier cells that are unrelated to the analyzed differentiation process—an essential mechanism for preventing the inference of spurious or biologically implausible trajectories. We benchmark MultistageOT on snapshot data of cell differentiation, showing significantly improved fate prediction accuracy over state-of-the-art bimarginal optimal transport and demonstrating MultistageOT’s unique ability to detect outlier cells.

Fluoride triggers lysis in <i>Streptococcus mutans</i> by inhibition of the Clp protease complex, leading to an unabated competence cascade

Fluoride has long been known to possess antimicrobial properties. For many bacteria, the toxic effects of fluoride are reversible. However, fluoride has also been shown to trigger lysis and cell death in many other diverse bacterial species, including dental pathogens. The underlying molecular mechanisms responsible for fluoride-induced cell lysis have not been established. Using Streptococcus mutans as a model, we show that fluoride elicits an uncontrolled stress response characterized by upregulation of competence pathways and extensive cell wall degradation. While controlled and limited autolysis under stress is an adaptive response, we show that expression of the competence-associated alternative sigma factor ComX is prolonged under fluoride stress, relative to other stressors. Using in vitro and in vivo analyses, we show that fluoride disrupts the typical tight temporal control of ComX by specifically inhibiting assembly and activity of Clp ATPases responsible for its proteolytic degradation. Unchecked, ComX upregulates bacteriocins and autolysins, while simultaneously suppressing immunity peptide expression via a 6S RNA-mediated mechanism. Thus, fluoride subverts cellular mechanisms to turn off competence pathways that are induced under cellular stress, causing irreversible damage to the cell wall and ultimately cell death. Phenylalanine partially restores Clp protease assembly and activity, providing a rationale for the frequent presence of a gene encoding chorismate mutase in fluoride-responsive operons. Together, our findings reveal the molecular mechanism of fluoride-dependent lysis in bacteria, fifty years after this phenomenon was first reported. These pathways could be exploited to potentiate the antimicrobial effects of oral fluoride.

A balance between glycitein and glyceollins governed by isoflavone 6-hydroxylase confers soybean resistance to <i>Phytophthora sojae</i>

Isoflavonoids, predominantly found in legumes, are specialized metabolites with antioxidant properties that benefit both plant resilience and human health. Using metabolic genome-wide association studies (mGWAS), we identified the cytochrome P450 gene ( Glyma.11g108300 ), GmIF6H1 , as a key determinant of glycitein biosynthesis in soybean [ Glycine max (L.) Merr.]. Biochemical assays together with in planta stable-isotope tracing demonstrated that GmIF6H1 catalyzes the 6-hydroxylation of daidzein, establishing a previously unrecognized and predominant biosynthetic route for glycitein. A single amino acid substitution in GmIF6H1 accounts for the domestication-associated reduction of glycitein-type isoflavonoids. Upon Phytophthora sojae infection, (malonyl)glycitins undergo sustained deglycosylation to release glycitein aglycone, underscoring its defensive role. Strikingly, both loss- and gain-of-function alleles increase susceptibility to P. sojae , indicating that precise tuning of GmIF6H1 expression is essential for effective resistance. Metabolite profiling further reveals complementary daidzein-centered defense strategies: Glycitein-type isoflavonoids (via daidzein 6-hydroxylation) function as phytoanticipins, whereas glyceollins (via daidzein 2’-hydroxylation) act as inducible phytoalexins. Together, these findings clarify the biosynthetic origin of the glycitein and underscore the synergistic action of glycitein and glyceollins in pathogen resistance, offering opportunities for engineering disease-resilient soybean cultivars.

Biological fidelity: The engine driving the neuromorphic renaissance

Reward-driven adaptation of movements requires strong recurrent basal ganglia–cortical loops

The basal ganglia (BG) are a collection of subcortical nuclei involved in motor control, sensorimotor integration, and procedural learning. They play a key role in the acquisition and adaptation of movements, a process driven by dopamine-dependent plasticity at cortico-striatal projections, which serve as BG input. However, BG output is not necessary for executing many well-learned movements. This raises a fundamental question: How can plasticity at BG input contribute to the acquisition and adaptation of movements which execution does not require BG output? Existing models of BG function often neglect the feedback dynamics within cortico-BG-thalamo-cortical circuitry and do not capture the interaction between the cortex and BG in movement generation and adaptation. In this work, we address the above question in a theoretical model of the BG-thalamo-cortical multiregional network, incorporating anatomical, physiological, and behavioral evidence. We examine how its dynamics influence the execution and reward-based adaptation of reaching movements. We demonstrate how the BG-thalamo-cortical network can shape cortical motor output through the combination of three mechanisms: i) the diverse dynamics emerging from its closed-loop architecture, ii) attractor dynamics driven by recurrent cortical connections, and iii) reinforcement learning via dopamine-dependent cortico-striatal plasticity. Our study highlights the role of the cortico-BG-thalamo-cortical feedback in efficient visuomotor adaptation. It also suggests a mechanism for early-stage acquisition of reaching movements through motor babbling. More generally, our model explains how the BG-cortical network refines motor output through its intricate closed-loop dynamics and dopamine-dependent plasticity at cortico-striatal synapses.

MBNL loss of function in smooth muscle as a model for myotonic dystrophy associated gastrointestinal dysmotility

Myotonic dystrophy type 1 (DM1) is the most common adult-onset muscular dystrophy and severely affects multiple organ systems, including the brain, heart, skeletal muscle, and gastrointestinal (GI) tract. Despite 80% of individuals with DM1 experiencing GI dysfunction that affects their daily life, the mechanisms of GI dysmotility in DM1 remain an understudied aspect of the disease. DM1 is caused by a CTG repeat expansion in the DMPK gene that, when expressed as an expanded CUG repeat RNA, sequesters and reduces the activity of the muscleblind-like (MBNL) RNA-binding protein family. We developed a mouse line with conditional, smooth muscle-specific knockout of Mbnl1 and Mbnl2 to model and investigate myogenic mechanisms contributing to GI dysmotility in DM1. Mice with Mbnl knockout exhibited delayed GI transit of small and large bowel in vivo and increased smooth muscle contractile tone of jejunum and colon segments ex vivo. Smooth muscle from the jejunum and colon showed no histopathology suggesting an intrinsic defect and contained increased phosphorylation of the 20 kDa myosin light chain (Mlc20), consistent with increased contraction. RNA sequencing of mouse and human DM1 GI samples enriched for smooth muscle revealed conserved misregulated alternative splicing of transcripts associated with the regulation of Mlc20 phosphorylation and smooth muscle contraction. These findings demonstrate that Mbnl knockout disrupts the regulation of contraction dynamics and causes GI smooth muscle hyperactivity, suggesting that therapeutics that reduce GI contractile activity may improve DM1 GI symptoms.

The inferred functional connectome underlying circadian synchronization in the mouse suprachiasmatic nucleus

Circadian rhythms in mammals arise from the spatiotemporal synchronization of ~20,000 neuronal clocks in the suprachiasmatic nucleus (SCN). Although anatomical, molecular, and genetic approaches have revealed diverse SCN cell types, how network-level wiring enables their synchronization remains unclear. To overcome the challenges of inferring functional connectivity from fixed tissue, we developed Mutual Information &amp; Transfer Entropy (MITE), an information-theoretic framework to infer directed cell–cell connections with high fidelity from long-term live-cell imaging. Recording and analyzing 3,290 h of clock gene expression from 8,261 SCN neurons across 17 mice, we uncovered a highly conserved, sparse SCN network organized into two asymmetrically coupled modules: dorsal and ventral. Connectivity analyses revealed five functional SCN cell types independent of neurochemical identity. Notably, only ~30% of vasoactive intestinal peptide neurons exhibited Hub-like connectivity, classifying them as Generators and Broadcasters of synchrony signals. Other spatially stereotyped cell types consistently identified as Bridges, Receivers, or Sinks. Simulations based on MITE-inferred connectomes recapitulated emergent SCN dynamics, including recovery from desynchrony and the daily dorsal-to-ventral phase wave of gene expression. Together, these results demonstrate that MITE enables precise mapping of cellular network topology, revealing the circuit logic and key cell types that mediate circadian synchrony across space and time in the mammalian SCN.

Correction for Jang et al., Lipid raft proteomics identify endothelial myosin-9 (MYH9) as a regulator of low-density lipoprotein transcytosis and atherosclerosis

We need transparency standards for social media research that involves companies

CDCA7 facilitates MET1-mediated CG DNA methylation maintenance in centromeric heterochromatin via linker histone H1

DNA methylation is a conserved epigenetic modification essential for maintaining genome stability. However, how methyltransferases maintain CG methylation within compact chromatin, including centromeres, remains unclear. In humans, CDCA7 is necessary for the inheritance of DNA methylation at juxta-centromeres. Mutations that impair its ability to bind chromatin result in Immunodeficiency, Centromeric Instability, and Facial Anomalies (ICF) syndrome, characterized by centromeric instability. To investigate whether CDCA7 function is conserved, we identified two Arabidopsis thaliana orthologs, CDCA7α and CDCA7β . The loss of both copies results in CG hypomethylation at pericentromeric regions and centromeric satellite repeat arrays. Machine learning analysis suggested that heterochromatic nucleosomes, with enrichment of H1, H2A.W, and H3K9me2, depend heavily on CDCA7 proteins for CG methylation maintenance of the associated DNA. Loss of H1 restores heterochromatic DNA methylation in cdca7α cdca7β mutants, indicating that CDCA7α and CDCA7β mainly remodel H1-containing nucleosomes for methyltransferases to access DNA. Notably, in h1.1 h1.2 mutants, CG methylation shows a significant increase in centromeres, which reveals a new inhibitory role of H1 in DNA methylation maintenance within satellite repeat arrays. Centromeric DNA hypermethylation is lost in h1.1 h1.2 cdca7α cdca7β quadruple mutants, demonstrating that CDCA7α and CDCA7β can act independently of H1 to enhance MET1 activity at nucleosomes. Overall, these findings establish CDCA7α and CDCA7β as conserved regulators of DNA methylation within heterochromatin and centromeric satellite repeat arrays.

Endogenous forms of ATP–ATP ⁴⁻ and MgATP ²⁻ —orchestrate distinct pathophysiological signaling via biased activation of P2X3 receptors

P2X receptors, a family of ATP-activated ion channels, encompass subtypes P2X1–7, which are expressed in both homo- and heterotrimeric forms across various tissues. These receptors play crucial roles in pathophysiological processes such as synaptic transmission, nociception, cough, and taste perception. Extracellular ATP exists as both MgATP 2− and ATP 4− , with P2X3 responding to both. The evolutionary rationale for two nearly identical ligands and their distinct signaling potential remains unclear. While previous structural studies suggest a uniform ATP recognition mechanism for two endogenous ATP forms, we propose that MgATP 2− and ATP 4− activate P2X3 through distinct mechanisms, leading to differential physiological and pathological outcomes. Using mutagenesis, voltage-clamp fluorometry, and small molecule interventions, we identify divergent interactions of ATP 4− and MgATP 2− with P2X3, despite binding to the same orthosteric pocket. In P2rx3 D158A/D158A transgenic mice, which selectively impair MgATP 2− activation, we find that MgATP 2− modulates ammonia-induced cough frequency without affecting complete Freund’s adjuvant-induced inflammatory pain or sweet taste preference. P2rx3 −/− mice show deficits in all three responses. The allosteric inhibitor aurintricarboxylic acid selectively modulates ATP 4− and MgATP 2− effects, resulting in distinct antitussive and analgesic outcomes in vivo. These findings uncover a mechanism of P2X3 activation by its endogenous ligands, diverging from previous structural models and resembling the biased activation mechanisms observed in G-protein-coupled receptors, offering insights for P2X3-targeted therapeutics.

Correction for Lee, The effects of social media criticism against public health institutions on trust, emotions, and social media engagement

Distinct neural dynamics in the ventral hippocampus and medial prefrontal cortex during social information processing

Social information processing involves coordinated neural activity across distributed brain circuits, with the ventral hippocampus (vHPC) and medial prefrontal cortex (mPFC) playing pivotal roles. However, whether these regions employ distinct coding strategies for different social stimuli remains unclear. Using in vivo electrophysiological recordings in freely interacting mice, we show that although both regions respond to social cues, they engage divergent neural coding mechanisms. During social interaction, the mPFC predominately recruits high gamma oscillations with amplitudes modulated by the phase of theta oscillations, whereas the vHPC primarily depends on low gamma activity phase-locked to theta rhythms. Information-theoretic and machine learning analyses demonstrate that neural populations in the mPFC encode social information more robustly than those in the vHPC. Moreover, integrating spiking activity with local field potential oscillations enhances decoding accuracy compared to spike-only models. Neural manifold analysis showed greater signal-noise angle in the mPFC compared to the vHPC, indicating more discriminative and stable social representations in the mPFC. Our findings demonstrate distinct engagement of neuronal populations and gamma oscillations in the vHPC and mPFC during social information processing.

Potentiation of ryanodine receptor–mediated calcium release by MAPK is responsible for epidermal transformation and carcinogenesis

Epidermal growth factor (EGF) induces anchorage-independent growth in promotion-sensitive (P + ) mouse epidermal cell model JB6 primarily through activation of the MAPK/ERK signaling pathway. The β-blocker carvedilol inhibits EGF-promoted JB6 transformation, but the underlying mechanism is unknown. Since carvedilol suppresses overactivated ryanodine receptors (RyRs) independently of its adrenergic blocking effects, we hypothesized that EGF-promoted transformation requires RyR-mediated calcium (Ca 2+ ) release and that carvedilol inhibits transformation via targeting RyRs. All RyR subtypes were present in epidermis and strongly upregulated by ultraviolet (UV) radiation, as demonstrated in an RyR2-tdTomato reporter mouse model. In vitro, EGF induced ERK phosphorylation and RyR2 upregulation and increased RyR agonist 4-chloro-m-cresol (4-CMC)-evoked Ca 2+ release, which is inhibitable by structurally divergent RyR stabilizers and inhibitors of MAPK and PLC, but not by most β-blockers. Expression of constitutively active K-RAS and MEK-1 or UV also potentiated 4-CMC-evoked Ca 2+ release. RyR agonists and the Ca 2+ ionophore ionomycin promoted JB6 transformation while RyR stabilizers, the intracellular Ca 2+ chelator BAPTA/AM, and inhibitors of MAPK and PLC blocked transformation. The RyR shRNAs abolished the transformation-inhibitory effect of carvedilol. The IC 50 values of five carvedilol derivatives for suppressing RyR-mediated Ca 2+ release positively correlated with the IC 50 values for transformation inhibition. In vivo, UV-induced DNA damage and skin inflammation were enhanced by topical 4-CMC treatment but attenuated in the RyR2-E4872Q knock-in mice in which RyR2 activity is reduced. Human skin tissue microarray analysis confirmed spatial colocalization of phospho-ERK and RyR2 in the same tumor areas. Thus, potentiation of RyR-mediated Ca 2+ release by MAPK is an important pathway leading to carcinogenesis.

Coexistence of trapped and flow-transported nuclei enables fast pigeon post communication across multinucleated cell

Multi-nucleated cells exist in all domains of life, ranging from animals, plants, and fungi to single-celled organisms such as the slime mold Physarum polycephalum . The large cell size, in the case of Physarum reaching centimeters and more, challenges the coordination of nuclei activity as signals need to cross large distances. In search of a mechanism for fast long-ranged communication among nuclei, we quantify nuclei dynamics and cytoplasmic flows in Physarum ’s tubular network. We observe nuclei in two interchangeable, dynamic states: mobile, flowing within the cytoplasmic shuttle flow, or trapped in the tube’s porous cell cortex. As we find nuclei to accumulate at the tube’s inner fluid–porous interface we theoretically explore and confirm, with physiological parameters, that slowing down of mobile nuclei during flow is sufficient for diffusible signal exchange between mobile and trapped nuclei. We analytically derive that communication akin to pigeon post with mobile nuclei serving as pigeons shuttling between trapped nuclei acting as waypoints, gives rise to signaling velocities that account for the rapid intracellular reorganization observed in Physarum . Since signal transfer by flow-transported nuclei outcompetes the mere diffusion of signals encoded in cytosolic proteins, pigeon post communication surpasses alternative signaling mechanisms, even diffusive relay signaling up to 20-fold in velocity. The key ingredients of pigeon post communication, namely alternating flows and waypoints, exist in other multi-nucleated cells and may also be generalized beyond intracellular signaling.

Contrastive independent component analysis for salient patterns and dimensionality reduction

In recent years, there has been growing interest in jointly analyzing a foreground dataset, representing an experimental group, and a background dataset, representing a control group. The goal of such contrastive investigations is to identify salient features in the experimental group relative to the control. Independent component analysis (ICA) is a powerful tool for learning independent patterns in a dataset. We generalize it to contrastive ICA (cICA). For this purpose, we devise a linear algebra–based tensor decomposition algorithm, which is more expressive but just as efficient and identifiable as other linear algebra–based algorithms. We establish the identifiability of cICA and demonstrate its performance in finding patterns and visualizing data, using synthetic, semisynthetic, and real-world datasets, comparing the approach to existing methods.

Localized nutrient colimitation of phytoplankton growth rates across the subtropical South Pacific Ocean

The simultaneous depletion of multiple nutrients in seawater potentially leads to colimitation of phytoplankton growth across large oceanic extents. Single limitation versus colimitation carries implications for mathematically predicting growth, its response to environmental forcing, and evaluating biogeochemical feedbacks. However, identifying colimited growth has proved challenging due to a lack of appropriate methods. Here, we present the results of 12 experiments conducted across the South Pacific that used a matrix of nutrient additions to strongly diluted surface seawater. Dilution restricted both grazing rates and nutrient drawdown due to phytoplankton accumulation. We find that despite simultaneous depletion of nitrate, phosphate, and iron concentrations throughout the oligotrophic gyre, community-level phytoplankton growth rates were only constrained by nitrogen. In contrast, zones of colimitation and serial limitation by nitrogen and iron were found along the eastern gyre margin. At the nitrogen-iron co-/serially limited sites, growth response surfaces to nutrient additions varied, suggesting the need for dynamic models to accurately represent colimited phytoplankton growth in the ocean.

High-resolution lidar observations of sedimentation-induced size sorting of droplets near a laboratory cloud top

Cloud optical properties and precipitation, which are crucial to weather and climate, are strongly influenced by cloud microphysical properties that are still poorly understood. Here, we develop a high-resolution time-correlated single-photon-counting lidar and apply it to observe cloud microphysical properties at one-centimeter range resolution in a convection chamber under well-controlled conditions. Together with concurrent in-situ measurements and theoretical analysis, our lidar observations indicate that although turbulent mixing tends to homogenize the cloud in the bulk region, entrainment and sedimentation cause inhomogeneities in droplet concentrations near the cloud top. Specifically, the topmost region is directly affected by entrainment, and lidar profiles show clear evidence of entrained air and detrained cloud filament. The transition region below exhibits vertical size sorting of cloud droplets caused by sedimentation. Our results suggest that using a single sedimentation velocity for all cloud droplets, as is done in many atmospheric models, overlooks key physics relevant to the microphysical structure near the cloud top. Our conceptual model used to describe these measurements can serve as a step toward improving the current modeling of processes in the cloud top region.

A metabolic cell death program downstream of SARM1 couples NAD ⁺ depletion to BAX activation and APAF1 degradation

SARM1 is a neuronal Nicotinamide adenine dinucleotide (NAD + ) hydrolase that drives axonal degeneration and neuronal death by depleting NAD + , yet how NAD + loss triggers axon loss and cell death has remained unclear. Here, we define a nonapoptotic death program downstream of endogenous SARM1 activation and NAD + loss using a genetically tractable nonneuronal eHAP cell model. Upon NAD + depletion, BAX is activated but caspase activation is suppressed due to APAF1 degradation via the E3 ligase HERC4, effectively uncoupling mitochondrial outer membrane permeabilization from apoptosome formation. Mechanistically, NAD + depletion inhibits mTOR/AKT signaling, destabilizing MCL1 and relieving BAX from repression. We further identified Neurofibromatosis type II, NF2, as a regulator that promotes SARM1 transcription through the Hippo–YAP/TAZ pathway. The SARM1-dependent BAX activation and the role of NF2 in axon degradation were validated in neuronal models of axon degeneration. Together, these findings reveal how SARM1-driven metabolic collapse rewires cell death execution, positioning BAX, MCL1, APAF1, NF2, and HERC4 as core effectors in a nonapoptotic degenerative pathway linking metabolic stress to neurodegeneration

Globally aggregated biodiversity data impact predictive and descriptive research

Here, we present an analysis of the growth and use of the Global Biodiversity Information Facility (GBIF) over the last 5 y. GBIF is the world’s largest data integrator for biodiversity information and plays a central role in research across the biodiversity and evolutionary science community. With the help of a comprehensive bibliographic dataset comprising 12,193 studies that used GBIF-mediated data, we demonstrate how the global scientific community utilizes the continuously fast-growing amount of open and Findable, Accessible, Interoperable, and Reusable biodiversity data in their research. Overall, more researchers engage with GBIF data, a potential consequence of the rising demands of more global environmental assessments, where GBIF-mediated data are being used as a key resource for biodiversity research. Studies utilizing species distribution modeling were most prevalent and data used for topics related to challenges of the Anthropocene (conservation, climate change, invasive, and pest species). More studies used observational data records, a category that also includes a substantial amount of citizen science data. Our data show that a thematic diversification of GBIF-using literature is accompanied by a rapid diversification of both the additional datasets that GBIF data are analyzed with, as well as the new analytical approaches taken by researchers. This emphasizes the growing importance of GBIF’s data infrastructure and services which support global sciences and reflect major shifts in applied science which dictate the need for GBIF and similar data infrastructures to evolve rapidly in order to maintain relevance for research.