Biology subjects

Zhu, X.

Publications and source records attributed to Zhu, X..

At least 37 records · Page 2Linked to original sources

bioRxiv2026

Polystyrene microplastics uptake drives Inflammatory, Epitranscriptomic, and Metabolic Reprogramming in Human aortic endothelial cells

Microplastics (MPLs) are pervasive environmental pollutants increasingly linked to adverse human health outcomes, including atherosclerosis. However, the underlying mechanisms remain poorly understood. Human aortic endothelial cells (HAECs), which line the inner surface of blood vessels, play a critical role in maintaining vascular homeostasis and in the development of atherosclerosis. This study demonstrates that polystyrene microplastics enter HAECs through clathrin-mediated endocytosis and macropinocytosis and subsequently co-localize with mitochondria and lysosomes. Exposure to MPLs induced coordinated transcriptional, epitranscriptomic, and metabolomic reprogramming in HAECs. Transcriptomic analysis revealed disruption of mitochondrial genes and activation of inflammatory pathways with the response of the NF-{kappa}B pathway being particularly prominent. Mass spectrometry analysis of RNA modification further identified significant remodeling of the epitranscriptomic landscape, highlighted by increased 1-methyladenosine (m1A) modification and reciprocal regulation of its associated enzymes (TRMT61A upregulation and ALKBH3 suppression), along with alterations in other RNA modifications such as m3C, pseudouridine ({Psi}), m5C, and m7G. Comparative analysis of transcriptomic profiles from human atherosclerotic plaques revealed shared dysregulated pathways in vascular regulation and cellular signaling. Metabolomic profiling further showed extensive remodeling of lipid metabolic networks associated with oxidative stress and inflammation. Together, these findings suggest that MPLs exposure may disrupt endothelial function and pose a potential risk to human cardiovascular health. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=141 SRC="FIGDIR/small/737624v2_ufig1.gif" ALT="Figure 1"> View larger version (34K): org.highwire.dtl.DTLVardef@2ad601org.highwire.dtl.DTLVardef@47f041org.highwire.dtl.DTLVardef@ee45cdorg.highwire.dtl.DTLVardef@128d6d5_HPS_FORMAT_FIGEXP M_FIG C_FIG

molecular biology↗

bioRxiv2026

UTRGen: A unified framework for full-spectrum design of mRNA 5' UTRs

AO_SCPLOWBSTRACTC_SCPLOWThe 5' untranslated region (5' UTR) is a key regulatory element that governs mRNA translation and protein output. However, existing computational methods typically address isolated tasks such as functional prediction or sequence optimization, limiting their ability to support rational design across the full 5' UTR engineering workflow. Here, we present UTRGen, a unified modeling framework for 5' UTRs that integrates sequence generation, multi-property prediction, and constrained function-guided design. UTRGen is pre-trained autoregressively on large-scale 5' UTR datasets from multiple species and subsequently adapted to diverse downstream regulatory tasks. Across systematic evaluations, UTRGen generates novel and diverse 5' UTRs while preserving sequence, structural, and functional characteristics of natural UTRs. After task-specific fine-tuning, UTRGen achieves state-of-the-art performance across 14 benchmark datasets, improving translation efficiency prediction by up to 11.1%, expression level prediction by up to 13.2%, and mean ribosome load prediction by up to 3.0% relative to the strongest baselines. It also achieved the best overall performance for internal ribosome entry site identification. To enable controllable design, we formulate function-guided 5' UTR design as a GRPO-based refinement process over a pre-trained autoregressive sequence prior, using composite rewards to encode functional objectives and biological constraints while regularizing toward the natural 5' UTR distribution. The resulting sequences show consistently improved predicted translation efficiency and expression levels across cellular contexts, and reveal interpretable sequence features associated with high activity, including reduced C content, fewer upstream AUGs, and depletion of inhibitory motifs. Together, our results establish a unified modeling strategy for 5' UTR design and lay a foundation for programmable control of translation.

bioinformatics↗

bioRxiv2026

Computational reconstruction of hierarchical cis-regulatory networks reveals synergistic transcription control and disease-associated rewiring

Gene regulation emerges from coordinated interactions among dispersed cis-regulatory elements, yet how these elements integrate into functional regulatory networks and collectively regulate gene transcription remains poorly understood. Here, we present ORIGAMI, a multi-omics, gene-centric deep learning framework that reconstructs functional cis-regulatory networks constrained by transcriptional output. ORIGAMI formulates cis-regulatory modeling as a latent graph inference task, which integrates DNA sequence, epigenomic signals, and three-dimensional chromatin priors to infer denoised regulatory graphs that capture functional interactions rather than structural proximity alone. The inferred regulatory graphs exhibit distinct topological regimes, where hierarchical and modular organization encodes cell-state-specific functional demands and enables synergistic transcriptional control. Furthermore, we show that these regulatory architectures undergo measurable state-dependent rewiring across disease contexts. Finally, ORIGAMI accurately predicts the transcriptional consequences of both cis- and trans-regulatory perturbations and links the rearrangement of regulatory architecture to perturbation response. Together, ORIGAMI advances a network-based view of gene regulation and establishes a foundation for virtual cell modeling of regulatory dynamics.

bioinformatics↗

bioRxiv2026

Physics-driven self-supervised learning for quantitative high-fidelity structured illumination microscopy

Structured illumination microscopy (SIM) enables rapid, long-term super-resolution (SR) imaging of live-cell dynamics. However, although current state-of-the-art (SOTA) SIM reconstruction methods achieve high-fidelity structural SR, they consistently lack reliable intensity quantification, restricting their use in quantitative biology. Here, we develop qHiFi-SIM, a physics-driven self-supervised learning framework for quantitative high-fidelity SR-SIM imaging. By leveraging the wide-field image as a physical intensity reference, our approach enables self-supervised training without reliance on SR data with ground-truth intensity. qHiFi-SIM achieves high structural fidelity (structural similarity, SSIM > 0.95) with a twofold resolution enhancement, while maintaining excellent intensity linearity (coefficient of determination, R{superscript 2} > 0.99). It also exhibits strong transferability across diverse SIM setups and typical samples, and is compatible with SOTA SIM algorithms, enabling direct quantitative correction of their intensity deviations without retraining. We demonstrate the unique advantages of qHiFi-SIM for live-cell quantitative visualization of mitochondrial structure, intensity, and membrane potential dynamics, as well as for high-fidelity SIM-FRET (Forster resonance energy transfer) functional imaging. We anticipate that qHiFi-SIM will serve as a practical tool for SR structural visualization and quantitative functional imaging in live cells.

Cell Biology↗

bioRxiv2026

A U1-U3 snRNA-snoRNA interaction couples SF3B1 mutation to chromatin-state rewiring and genome instability

Mutations in spliceosome factors such as SF3B1 are recurrent across human diseases, including myelodysplastic syndromes and leukemia1-4, yet splicing defects alone do not fully explain the widespread chromatin alterations and genome instability in mutant cells5. Here, by comprehensively mapping snRNA-directed RNA-RNA interactions, we identify two previously unrecognized interaction motifs in U1 snRNA beyond canonical 5' splice-site pairing6,7. These motifs enable U1 RNA to i) bind intronic and other chromatin-associated RNA (caRNA) regions outside of splice sites, and ii) base pair specifically with snoRNA. We uncover a U1-U3 snRNA-snoRNA interaction that recruits the H3K36 methyltransferase SETD2 to caRNA, promoting gene-body H3K36me3 and antagonizing H3K27me3 to modulate chromatin accessibility. The snRNA-snoRNA interface is essential for this previously unrecognized layer of chromatin and transcriptional regulation mediated through SETD2. SF3B1 mutation enhances U1-U3 binding and increases the association of the U1-U3 complex with caRNA, driving chromatin-accessibility remolding, R-loop formation, DNA damage, and copy-number abnormalities that promote tumorigenesis. A U1-specific 2'-O-methoxyethyl antisense oligonucleotide that selectively blocks U1-U3 pairing suppresses these genomic abnormalities, reduces leukemic infiltration, and prolongs survival in xenograft and patient-derived models, establishing pathological snRNA-snoRNA rewiring as a critical driver of SF3B1-mutant leukemogenesis.

cell biology↗

bioRxiv2026

Structure and functional diversity of antibodies targeting the P. falciparum circumsporozoite protein C-terminal domain

The Plasmodium falciparum circumsporozoite protein (PfCSP) is the major surface antigen on Pf sporozoites. WHO-recommended vaccines RTS,S/AS01E and R21/Matrix-M target the PfCSP major repeat region and C-terminal domain (ctCSP). Although multiple studies associated protection with antibody responses to ctCSP, only a few ctCSP-specific monoclonal antibodies (mAbs) have been characterized. Here, crystal structures of 11 Fab-ctCSP complexes reveal how mAbs against the conserved {beta}-epitope region achieve diverse modes of strain-transcending recognition, in contrast to mAbs to the hypervariable -epitope. Consistent with previous studies, ctCSP on sporozoites could be unmasked by mAbs that bind CSP repeats, with unmasking dependent on the mAb fine-specificity and binding mode. In vitro, ctCSP mAbs promoted stronger Fc-receptor signaling, cellular cytotoxicity, and phagocytosis than repeat region mAbs, while mAb combinations targeting distinct PfCSP epitopes modulated Fc-signaling and cellular cytotoxicity. This study provides a rationale for optimization of PfCSP-based immunogens to enhance Fc-mediated contributions to malaria vaccine efficacy.

microbiology↗

bioRxiv2026

Long-read RNA-seq resolves isoform-level and context-specific regulatory architecture of complex traits in cattle

Traditional short-read RNA-seq lacks the resolution to fully capture full-length isoforms and complex alternative splicing, leaving a "missing regulation" gap in existing molecular atlases. Here, we present a population-scale, multi-omics dissection of bovine complex traits using matched deep whole-genome sequencing (30x), long-read (ONT) and short-read RNA-seq, as well as metabolome data from 432 dairy cows across four distinct lactation stages. We expanded the bovine transcript atlas with 13,177 novel isoforms and fine-mapped 1,088,900 regulatory effects across 11 diverse molecular phenotypes. Notably, long-read mapping uncovered 606 genes with isoform expression quantitative trait loci (12.8%, eQTL) and 1,116 genes with detailed splicing QTL (sQTL) (31.8%) that were missed by conventional short-read mapping. We further showed that lactation-specific regulation was profoundly mediated by the metabolic microenvironment (46.8%) and cellular composition (4.3%). Across genome-wide associations (GWAS) of 90 bovine complex traits, isoform-level and context-specific regulatory effects captured GWAS loci missed by short-read eQTL and sQTL analyses, increasing colocalized loci from 35.2% to 60.3%. These regulatory variants also exhibited evolutionary constraints with human immune and metabolic loci. Our study establishes an invaluable resource (https://cattleblr.farmgtex.org) demonstrating that long-read transcriptomics is essential for resolving the complexity of regulatory mechanisms underlying mammalian physiology.

genetics↗

bioRxiv2026

Identification of Heterogeneous Cortical Thickness Patterns Associated with Prenatal Gestational Diabetes Exposure: A SuStaIn-Based Subtyping Study

ImportancePrenatal exposure to gestational diabetes mellitus (GDM) has been associated with adverse metabolic, neurodevelopmental, and psychiatric outcomes in offspring. However, whether GDM-exposed youth exhibit heterogeneous neuroanatomical patterns remains unclear. ObjectiveTo identify distinct cortical thickness subtypes among GDM-exposed youth and examine their associations with anthropometric, neurocognitive, psychiatric/behavioral and neuroimaging measures both cross-sectionally and longitudinally. Design, Setting, and ParticipantsThis cohort study used the Adolescent Brain Cognitive Development (ABCD)(R)data, a multisite longitudinal population study. Subtype and Stage Inference (SuStaIn), an unsupervised machine learning framework, was applied to cross-sectional structural MRI data to identify cortical thickness patterns in 573 GDM-exposed youth and 2854 healthy controls. Posthoc longitudinal analyses included 1,853 observations from a subset of GDM-exposed youth with 1-, 2-, and 4-year follow-up visits to examine subtype differences in developmental trajectories over time. Exposure(s)Prenatal exposure to GDM. Main Outcome(s) and Measure(s)The primary outcomes included identification of cortical thickness subtypes and their inferred regional ordering patterns. Secondary outcomes included subtype-specific anthropometric, neurocognitive, psychiatric/behavioral and neuroimaging measures. ResultsThe GDM-exposed sample had a mean age of 119.02 {+/-} 7.34 months and was 47.5% female. Two cortical thickness subtypes were identified. Between subtypes, Subtype 1 (63.2%) was characterized by earlier inferred insula involvement and was associated with greater height (d = 0.36, pFDR < 0.001) and weight (d = 0.26, pFDR = 0.007), whereas Subtype 2 exhibited earlier inferred frontal involvement and nominally higher Attention-Deficit/Hyperactivity Disorder (ADHD) prevalence (d = 0.08, p = 0.036), steeper longitudinal cortical thinning across all six cortical regions of interest ({beta} range: -0.05 to -0.13, all pFDR < 0.05), and a smaller decline in Obsessive-Compulsive Disorder (OCD) prevalence over time ({beta} = -1.02, pFDR = 0.049). Conclusions and RelevanceGDM exposure was associated with two distinct offspring cortical thickness subtypes, each showing different inferred regional ordering patterns and clinical associations. One subtype showed an insula-cingulate-predominant pattern associated with anthropometric measures, whereas the other showed a frontal-predominant pattern associated with nominally higher psychiatric measures and faster cortical thinning over time.

neuroscience↗

bioRxiv2026

ACLY integrates metabolism and chromatin accessibility to enable B Cell activation and humoral immunity

B cell activation and differentiation into antibody-secreting cells require extensive metabolic and epigenetic remodeling, yet the molecular mechanisms that integrate these programs remain incompletely understood. ATP-citrate lyase (ACLY) links glucose metabolism to acetyl-CoA production, supporting lipid biosynthesis and protein acetylation. However, its role in humoral immunity has not been fully defined. Here, using genetic and integrated multi-omics approaches, we show that B cell activation is accompanied by coordinated metabolic, transcriptional, and epigenetic reprograming. Although ACLY is dispensable for B cell development and homeostasis, it is required to establish chromatin accessibility programs in activated B cells, with a more pronounced impact on the epigenetic landscape than on transcriptional output. ACLY-deficient B cells exhibit profound defects in TLR and BCR elicited activation, survival and metabolic fitness ex vivo. In vivo, B cell-intrinsic loss of ACLY results in impaired antigen-specific antibody production, associated with reduced germinal center and plasmablast formation, but normal homeostatic proliferation. Deletion of ACLY after B cell activation reduces plasmablast generation in vivo, indicating a continued requirement for ACLY beyond the initial activation phase. Together, these findings identify ACLY as a central regulator that links metabolism to epigenetic programing that supports B cell activation and humoral immunity.

immunology↗

bioRxiv2026

Integrated optimization of experimental and computational workflows improves genome recovery in long-read gut metagenomics

Short-read metagenomic sequencing has been widely applied in microbial research due to its high quality and decreasing cost. However, short reads are inherently fragmented, which limits assembly contiguity and the recovery of complete microbial genomes. In contrast, long-read sequencing, with significantly longer read lengths, helps overcome these limitations. Achieving complete and accurate genome recovery is the core goal of metagenomics. To address this challenge, we systematically evaluated and optimized the long-read metagenomic workflow, from sample processing to computational assembly, using the CycloneSEQ platform.

bioinformatics↗

bioRxiv2026

Metabolic control of smooth muscle cell phenotype switching in atherosclerosis

The loss of smooth muscle cell (SMC) contractile phenotype contributes to various diseases including atherosclerosis. However, its metabolic basis is not entirely elucidated. Since the transforming growth factor beta (TGF{beta}) signaling is among principal regulators of SMC contractility, we studied metabolic regulation of TGF{beta} signaling in SMCs in vitro and atherosclerotic mouse models and human lesions. We found that TGF{beta} induced Ac-CoA synthetase 2 (ACSS2)-dependent Ac-CoA production, by suppressing pyruvate dehydrogenase kinase 4 (PDK4). This stabilized R-SMADs and TGF{beta} receptor 1, preserving SMC contractile phenotype. SMC-specific PDK4 knockout mimicked the effect of TGF{beta} signaling both metabolically and phenotypically, increasing glucose-derived synthesis of Ac-CoA and SMC contractile phenotype. SMC-specific Pdk4 knockout in ApoE knockout mice reduced atherosclerosis. Furthermore, human specimens demonstrated a strong correlation between PDK4 level and atherosclerosis severity. These findings indicate that continuous TGF{beta} signaling, critical to the maintenance of the normal SMC contractile state and is regulated by PDK4 and carbohydrate metabolism. TeaserReducing PDK4 metabolically restricts aortic plaque growth via TGF{beta}-dependent SMC contractility.

cell biology↗

bioRxiv2026

HESTIA: Scalable Multimodal Integration of Histology and High-Resolution Spatial Transcriptomics for Robust Spatial Domain Identification

Spatial omics has revolutionized molecular biology by providing invaluable insights into how native tissue microenvironments regulate cellular functions and disease mechanisms. Accurately capturing this structural complexity and decoding the underlying biological processes requires effectively integrating data from multiple modalities. However, transitioning to subcellular resolutions introduces massive data scales and severe transcriptomic sparsity, which challenge current analytical frameworks. To address this, we present HESTIA (Histology-Enhanced Scalable cross-Resolution inTegration for spatial trAnscriptomics), a highly efficient multimodal algorithm designed for identifying spatial domains in large-scale, high-resolution spatial omics data. By circumventing memory-intensive computations, HESTIA efficiently processes massive datasets on which existing algorithms fail due to memory constraints. HESTIA outperforms current multimodal methods in clustering accuracy and spatial continuity, accurately delineating fine structural boundaries. Furthermore, applying HESTIA to large-scale pathological samples successfully dissects clinically relevant intratumoral heterogeneity and maps distinct immune microenvironments in lung and colorectal cancers.

bioinformatics↗

bioRxiv2026

Predictable adaptive evolution of a phage endolysin through substrate recognition optimization

Bacteriophages release progeny by producing endolysins that degrade the bacterial cell wall. While the frequent horizontal transfer of endolysins suggests substantial evolutionary plasticity, the mechanisms by which these enzymes adapt to new phage-host contexts remain poorly understood. Here, we investigated the evolutionary dynamics and structural mechanisms governing endolysin adaptation following experimental evolution of a chimeric phage generated via heterologous endolysin exchange between phages infecting different hosts. Using replicate experimental evolution and time-resolved PacBio sequencing, we identified a dominant, highly reproducible adaptive trajectory characterized by the stepwise fixation of three key mutations. This constrained mutational order correlated with incremental gains in enzymatic activity, reflecting a rugged yet predictable fitness landscape. High-resolution structural analyses revealed that these substitutions lie exclusively outside the catalytic site; instead, they enhance substrate recognition through electrostatic tuning, optimized hydrophobic packing, and local conformational refinement, resulting in significantly higher binding affinity. While the adaptive trajectory was largely conserved, one replicate followed an alternative path, highlighting the interplay between selection and historical contingency. Adaptation was further shaped by a functional trade-off, whereby increased lytic activity on the novel host was accompanied by reduced activity on the ancestral host, consistent with antagonistic pleiotropy. Genome-wide sequencing additionally identified a compensatory mutation in a lytic transglycosylase, suggesting coordinated evolution of the broader lysis machinery. Together, these results demonstrate that endolysins evolve through reproducible adaptive walks constrained by structure, selection, and trade-offs, providing a mechanistic framework for understanding enzyme evolution and informing rational protein engineering.

evolutionary biology↗

bioRxiv2026

An aPKC rheostat induces apical contraction in response to epithelial stretching

Apical-basal polarity in epithelial cells is controlled by a conserved set of polarity factors that define the apical, junctional and basolateral domains of the cell, but how these factors adapt to or control changes in domain sizes during cell shape changes remains unclear. Atypical protein kinase C (aPKC) is the main effector of apical identity, phosphorylating the lateral factors, Bazooka/Par-3, Lgl, Par-1 and Yurt to exclude them from the apical domain. Using analogue-sensitive aPKC in Drosophila follicle cells, we found that aPKC substrates differ over 100-fold in their sensitivity to inhibition, revealing a hierarchy of substrates, that is conserved in mammals, in which high-affinity substrates out-compete low-affinity substrates when aPKC activity is limiting. Mild aPKC inhibition prevents the phosphorylation of its lowest affinity substrate, Yurt. Yurt then accumulates apically by binding to Crumbs, where it activates apical constriction through Shroom, Cysts/Dp114RhoGEF, Rho kinase and Myosin. Yurt localises apically in cells that are stretched, either by morphogenesis or artificially, indicating that stretching reduces aPKC activity to trigger an antagonistic contraction. By contrast, yurt- cells fail to resist stretching. Thus, the aPKC/Yurt pathway functions as a homeostatic stretch response, in which apical and lateral epithelial polarity factors collaborate to mechanically regulate apical domain size.

cell biology↗

bioRxiv2026

Hex-MASP for Mapping the Whole-tissue Spatial Proteome and the Intra-brain Distribution of Monoclonal Antibodies

Whole-tissue spatial proteomics level provides critical insights into region-specific biological regulations but remains challenging. Previously, we introduced the Micro-scaffold Assisted Spatial Proteomics (MASP) concept for whole-tissue mapping. However, this prototype required substantial development in spatial resolution, practicality, and throughput for practical application. Here we present a next-generation MASP technique (hex-MASP) featuring i) a new design of hexagonal-micro-wells fabricated with optimized Projection Micro-Stereolithography (P{micro}SL) 3D-printing, achieving high spatial resolution, sampling robustness and mechanical strength for reproducibly compartmentalizing even tough tissues; ii) enhanced throughput/effectiveness in sample preparation and LC-MS analysis with high quantitative quality. Applied to mouse brain, hex-MASP for the first time achieved in-depth, whole-tissue mapping for >6,000 proteins in mouse brains, with high spatial accuracy and excellent data quality. The substantially improved resolution revealed critical regional details across the entire brain, that were not previously captured, enabling precise depiction of protein distribution heterogeneity. This technique enabled the discovery of many unreported regionally-enriched proteins across brain structures. We further applied hex-MASP to investigate the intra-brain distribution of intracerebroventricularly-dosed antibody therapeutics and related proteins, which to our knowledge, enabled whole tissue mapping of protein drugs for the first time and revealed novel mechanistic insights into antibody distribution and localized treatment effects. Hex-MASP represent a robust, scalable platform for whole-tissue spatial proteomics.

pharmacology and toxicology↗

bioRxiv2026

Transdiagnostic Neurobiological Biotypes of Trauma Timing: Data-driven approach of Childhood and Adulthood-Onset Trauma

BackgroundThe developmental timing of trauma exposure may critically shape neurobiological outcomes, yet distinctions between childhood-onset trauma (CT) and adulthood-onset trauma (AT) remain poorly understood. AimThis study explores whether trauma onset timing is associated with distinct resting-state functional connectivity (rsFC) pattern using data-driven approach. MethodsSeventy-seven trauma-exposed individuals (Mage=36.74 years) with post-traumatic stress disorder (PTSD), PTSD with major depressive disorder (MDD), and trauma-exposed healthy controls (TEHC) underwent resting-state fMRI. Of these participants, 15 with CT only, 17 with both CT and AT, and 47 with AT only. RsFC was calculated across the amygdala, hippocampus, nucleus accumbens (NAcc), the salience (SN), default mode (DMN), and frontoparietal networks (FPN). K-means clustering identified subgroups based on rsFC, with robustness assessed via bootstrapping, cross-validation, and replication using Gaussian Mixture Modeling. The identified clusters were compared on trauma timing, type, cumulative exposure, and clinical measures. ResultsA two-cluster solution provided the most stable fit. The two generated clusters were significantly different in CT-only prevalence (p < 0.05; Cramers V = 0.26, 95% CI). The CT cluster was marked by hyperconnectivity between amygdala-FPN, DMN-SN, NAcc-SN, and hippocampus-FPN relative to the AT cluster. Individuals with both CT and AT were evenly distributed across clusters. Clusters did not differ in PTSD or comorbid diagnoses, trauma type, or cumulative exposure. ConclusionData-driven clustering revealed distinct neurobiological profiles differentiating CT and AT. CT was associated with hyperconnectivity across salience, reward, and regulatory circuits, supporting developmental timing as a determinant of brain network organization in trauma-exposed populations.

neuroscience↗

bioRxiv2026

Connections across regional glymphatic clearance, neural activity and amyloid-β deposition in cortex

Neural activity inevitably produces waste, which promotes neurodegeneration with topographic features. The glymphatic system is important for waste clearance. However, the spatial characteristics of glymphatic clearance across cortex and whether it interplays with neural activity in contribution to amyloidosis in human remain unexplored. Here, by intrathecal administration of gadolinium-based contrast agents, glymphatic influx and clearance patterns across cortex in 96 participants are depicted via Glymphatic MRI. Analyses integrating post-mortem transcriptomic profiles from Allen Human Brain Atlas indicate that, genes related with excitatory and inhibitory neurons, and pathways engaging in synaptic function were enriched in regions with faster glymphatic clearance. FALFF was calculated from resting-state fMRI to represent neural activity. At the regional level, based on a subgroup with rs-fMRI (N = 15), regional glymphatic clearance was positively coupled with spontaneous neural activity. Mismatch index, reflecting decoupling between spontaneous neural activity and glymphatic clearance function, turned out to be positively associated with regional severity of amyloidosis using open-source 11C-PiB dataset. Together, this study for the first time demonstrates the intricate interplays between neural activity and glymphatic dynamics from transcriptional to physiological level. The mismatch between these two processes may serve as an undescribed comprehensive mechanism promoting regional vulnerability to proteopathy and subsequent neurodegeneration in cortex.

neuroscience↗

bioRxiv2026

CellPulse: A Foundation Model of Coordinated Gene Dynamics Simulating Viral Infectious Diseases

Understanding how cells respond to perturbations like viral infections requires models capturing coordinated gene dynamics. However, current gene expression foundation models are predominantly reliant on single-cell data and static gene expression, limiting their applicability in real clinical scenarios. We present CellPulse, a direction-aware foundation model trained on the Virus Stimulated Atlas (VISTA), a newly curated atlas of over 23 million bulk RNA-sequencing differential expression profiles from viral infections. CellPulse models the direction and magnitude of gene expression changes via a structured representation of differential expression and a direction-aware attention mechanism, enabling the learning of coherent regulatory programs. It shows powerful diagnosing capability by accurately classifying 31 distinct virus types across diverse clinical and laboratory samples, solely from host transcriptional signatures. Crucially, without prior knowledge injection, CellPulses interpretability reveals virus-associated host factors that mediate infection. Using a selection of host factors for in silico drug screening yielded numerous compounds with confirmed efficacies in wet-lab assays, while cell-based and animal experiments further verified the causal relationship between host targets and viral infections. Overall, CellPulse represents a generalizable foundation model for deciphering coordinated gene dynamics from bulk transcriptomics, bridging host response modeling with clinical relevance and therapeutic discovery for infectious diseases and beyond.

bioinformatics↗