Biology subjects

Wen, Z.

Publications and source records attributed to Wen, Z..

At least 19 recordsLinked to original sources

bioRxiv2026

Stimulating microglial neuroinflammation with amyloid-beta oligomers compatible with solid-state NMR structural characterization

Soluble amyloid-{beta} (A{beta}) oligomers are considered the primary neurotoxic species in Alzheimer's disease, yet the relationship between oligomer structure and biological activity remains poorly understood. Oligomer sample preparations amenable to high-resolution structural characterization have not been tested in cell-based assays, mostly because detergents used to drive oligomer formation are toxic to cells. Here, we describe a detergent-free protocol in which size-exclusion chromatography-isolated A{beta}42 monomers spontaneously oligomerize at 37 {degrees}C in a concentration- and temperature-dependent manner. Biophysical characterization shows a {beta}-sheet structure by circular dichroism, a self-associating population above 100 kDa by silver-stained SDS-PAGE, and bead-like strings by negative-stain TEM resembling the previously characterized SDS-induced 150 kDa oligomer. Solid-state NMR spectroscopy of selectively 13C/15N-labeled 37 {degrees}C oligomers yields well-resolved spectra with {beta}-strand chemical shifts, and two-dimensional 13C-13C DARR and CHHC experiments provide initial inter-residue distance constraints that reveal both similarities to and differences from the 150 kDa oligomer. Treatment of mouse microglial cells (MMCs) with 50 nM 37 {degrees}C oligomers elicited a robust pro-inflammatory cytokine response that exceeded that of monomeric A{beta}42, as demonstrated by Luminex multiplex profiling and principal component analysis. hiPSC-derived microglia (iMGs) treated with 1 {micro}M oligomers also separated from vehicle, establishing activity in a human cellular model. Monomeric A{beta}42 was inactive in both systems, indicating that inflammatory activity depends on the assembly state of A{beta}42 rather than on the presence of peptide alone. These results establish a detergent-free A{beta}42 oligomer preparation that is simultaneously compatible with solid-state NMR structural characterization and biologically active in neuroinflammation assays, providing a platform for studying microglial immune responses to oligomers.

neuroscience↗

bioRxiv2026

Synaptic microRNA dysregulation orchestrates aberrant glutamatergic transmission in a human neuronal model of 22q11.2 deletion syndrome

22q11.2 Deletion Syndrome (22q11DS), the most prevalent chromosomal microdeletion disorder in humans (incidence ~1 in 4,000 live births), arises from hemizygous deletions of a 1.5-3.0 Mb region on chromosome 22. Notably, 20-30% of individuals with 22q11DS develop schizophrenia (SCZ) by early adulthood, making 22q11DS one of the strongest genetic risk factors for SCZ. However, the mechanisms by which the 22q11.2 deletion perturbs neuronal function and drive psychotic manifestations in 22q11DS remain poorly defined. In this study, we generated human 2D cortical neurons and 3D cortical organoids from induced pluripotent stem cells (iPSCs) derived from clinically well-characterized individuals with 22q11DS. Cellular and electrophysiological analyses revealed that 22q11DS neurons exhibit hyperexcitability, reduced glutamatergic synaptic transmission, and impaired homeostatic synaptic plasticity, which are pathophysiological hallmarks that may be relevant to the clinical, cognitive, and psychophysiological profiles of patients. Transcriptomic profiling of neurons and synaptosomes from 22q11DS cortical organoids uncovered dysregulated immune pathways and synaptic gene networks. Integrative mRNA-microRNA network analysis identified synergistic downregulation of hsa-miR-185-5p and hsa-miR-128-3p at synapses, driving aberrant regulation of immune and synaptic transcripts. Experimental knockdown of these microRNAs in control neurons recapitulated the molecular and functional phenotypes of 22q11DS neurons, thereby establishing a mechanistic link to synaptic dysfunction and impaired plasticity. Collectively, these findings delineate a multimodal mechanistic framework that bridges genetic vulnerability, synaptic pathology, and psychosis penetrance in 22q11DS, positioning dysregulated synaptic microRNAs as nodal regulators of neurodevelopmental networks and potential therapeutic targets.

neuroscience↗

bioRxiv2026

P2Y14 receptor agonist impairs, and antagonist improves whole-body glucose homeostasis and liver steatosis

Knockout of the UDP-glucose-activated P2Y14 receptor (P2Y14R) in adipocytes or whole-body has been reported to provide metabolic benefits in obese mice. We hypothesized that selective P2Y14R activation would lead to metabolic impairments, whereas pharmacological antagonism would improve metabolic deficits in diet-induced obese (DIO) mice. Here, we investigated the metabolic effects of a synthetic P2Y14R agonist, UDP-like methylene-bridged MRS2905. Acute P2Y14R activation with MRS2905 triggered a robust and prolonged hyperglycemic effect in lean and obese mice with impaired glucose homeostasis. Moreover, MRS2905 treatment of obese mice lowered fasting plasma insulin and increased glucagon levels, along with upregulation of liver JNK phosphorylation and the expression of rate-limiting gluconeogenic genes. The MRS2905-induced hyperglycemic effect was blunted in whole-body P2Y14R knockout mice compared with wild-type control. In contrast, a potent P2Y14R antagonist mono-ester prodrug (MRS4779) partially reversed the agonist-induced hyperglycemia and restored proper glucose homeostasis after acute treatment. Chronic MRS4779 administration in DIO mice reduced fat mass and improved various metabolic parameters including liver steatosis. Additionally, we examined the roles of P2Y14R in hepatocytes of DIO mice. Here, we report that P2Y14R was upregulated in liver and hepatocytes from obese mice compared to lean mice. Overnight fasting also upregulated hepatic P2Y14R expression. P2Y14R deletion from hepatocytes in DIO mice improved fasting blood glucose level and lipid metabolism without improving glucose homeostasis. These results suggest a novel P2Y14R function in hepatic lipid metabolism, and P2Y14R antagonists may prove useful for the treatment of obesity and obesity-related metabolic disorders.

physiology↗

bioRxiv2026

CNVeil resolves haplotype-specific copy number and uncovers subclonal architecture hidden from total copy number profiling in single-cell cancergenomes

Single-cell DNA sequencing (scDNA-seq) resolves copy number variation (CNV) at single-cell resolution, revealing tumor heterogeneity and subclonal structure. Most existing methods, however, infer only total copy number. Haplotype-resolved copy number, which captures allelic imbalance and clonal evolution, remains far less developed, largely because low coverage, allelic dropout, and technical noise in scDNA-seq make phased allelic inference substantially harder than total copy number estimation. We present CNVeil, a haplotype-aware framework that infers total, allele-specific, and chromosome-scale haplotype-resolved copy number from scDNA-seq data. CNVeil first builds robust total copy number profiles through highly variable bin selection, hierarchical clustering, subclone-aware ploidy estimation, and cross-cell consensus segmentation. Using this profile as a stable scaffold, it infers allele-specific copy number with an expectation-maximization algorithm applied to heterozygous SNP allele counts, then reconstructs haplotype-specific copy number by enforcing coherent haplotype orientation across adjacent segments via dynamic programming. We benchmarked CNVeil against 12 state-of-the-art methods, including eight total copy number callers, two allele-specific callers, and two haplotype-resolved callers, across 20 simulated and real datasets spanning six experimental settings, including high-multiplexed single-nucleus sequencing, Acoustic Cell Tagmentation (ACT), and 10x Chromium. This constitutes the largest comparative evaluation of single-cell copy number inference methods to date. CNVeil consistently outperformed existing tools in segmentation accuracy, ploidy inference, subclone identification, and allele-specific copy number estimation. In a breast cancer multi-omics (wellDR-seq) cohort, CNVeil uncovered haplotype-specific subclonal diversification invisible to total copy number analysis alone and linked allele-specific copy number states to transcriptional variation. By transforming sparse single-cell allelic signals into chromosome-scale haplotype-resolved profiles, CNVeil closes a major methodological gap and provides a scalable framework for studying tumor evolution and functional genomic heterogeneity at single-cell resolution.

bioinformatics↗

bioRxiv2026

Tissue geometry encodes a surface tension gradient that drives epithelial renewal

The mammalian intestinal epithelium renews itself every few days through coordinated cell movement and extrusion along the crypt-villus axis, yet how these processes are physically organized across intact tissue remains unknown. Using long-term live imaging of intact intestinal tissue, nuclear strain mapping, and targeted perturbations of cell-cell adhesion, actomyosin contractility, and extracellular matrix adhesion, we find that epithelial cells experience a graded mechanical landscape encoded by villus geometry; tension is highest at the villus tip and decays toward the base. This tissue-scale tension gradient, generated by E-cadherin-mediated adhesion and actomyosin contractility, is the dominant predictor of collective cell velocity, outperforming force magnitude and cell density. Artificial re-establishment of a tension gradient is sufficient to reorient epithelial migration, whereas disruption of the gradient arrests both cell movement and extrusion. Epithelial renewal further requires a precise balance between this tension gradient and extracellular matrix-derived friction; excessive friction uncouples migration from extrusion, leading to pathological cell accumulation. Together, these findings reveal that intestinal epithelial homeostasis is organized by a geometry-encoded mechanical gradient that coordinates collective cell motion, extrusion, and tissue repair through a single tissue-scale framework.

cell biology↗

bioRxiv2026

Extracellular signalling regulates gastrin transcription through site-specific phosphorylation and nuclear redistribution of Menin

The multiple endocrine neoplasia type 1 (MEN1) gene encodes Menin, a nuclear scaffold protein and tumor suppressor that regulates transcription. It is frequently mutated in endocrine neoplasia. MEN1-gastrinomas are aggressive neuroendocrine tumors (NETs) that arise predominantly in the submucosal Brunners glands of the duodenum, an organelle rich in extracellular growth factors. Many duodenal NETs retain wild-type MEN1 allele and nuclear Menin, suggesting post-translational inactivation of its tumor-suppressor function. The Menin C-terminal domain (CTD) contains a conserved phosphorylation site at Ser487 within the first of three nuclear localization signals (NLS1-3). We hypothesized that extracellular signaling regulates Menin by phosphorylating the CTD at Ser487 blocking its nuclear localization. Using CTD deletion mapping, site-directed mutagenesis, and kinase activation in gastric cell lines, we show that loss of NLS1-3 reduces Menins nuclear localization, stability, and repression of GASTRIN. Cell stimulation by epiregulin, forskolin, or phorbol ester induced Menin Ser487 phosphorylation and its nuclear translocation, relieving repression of GASTRIN. The phospho-mimetic S487D mutant remained cytoplasmic and phenocopied CTD deletion of NLS1-3 sustaining de-repression of GASTRIN. These findings showed that Ser487 phosphorylation restricts nuclear accumulation of Menin and functionally links extracellular signaling to post-translational modification of Menin that ultimately contributes to transcriptional derepression and neuroendocrine tumorigenesis. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=127 HEIGHT=200 SRC="FIGDIR/small/717082v1_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@1f96df4org.highwire.dtl.DTLVardef@a1db4borg.highwire.dtl.DTLVardef@4435f9org.highwire.dtl.DTLVardef@3373b3_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗

bioRxiv2026

Fear Learning Induced Brain Dynamics Predict Individual Extinction Memory Expression following Transcranial Magnetic Stimulation

Fear learning and extinction unfold as time-dependent processes. Herein, we examined how fear learning dynamically reorganizes brain activity immediately after learning, and whether such reorganization can be modulated with TMS application during extinction learning to prospectively predict extinction-memory expression. Eighty-seven healthy adults completed a three-day Pavlovian threat-learning protocol with resting-state fMRI acquired before and after conditioning (Day 1), dorsolateral prefrontal cortex (DLPFC) transcranial magnetic stimulation (TMS) applied during extinction learning (Day 2), and fMRI during extinction recall and renewal (Day 3). Using coactivation pattern analysis with a hidden Markov model within a 24-nodes threat-circuit parcellation, we identified a fear-learning-induced brain state characterized by global threat-circuit coactivation with heightened engagement and transition uncertainty post conditioning, and a progressive increase in engagement across post-conditioning. Critically, conditioning-induced functional connectivity reorganization within this state predicted individual differences in extinction recall- and renewal-related brain activation under TMS-modulated extinction (cross-validated; recall r = 0.47, p = 0.001; renewal r = 0.37, p = 0.01; permutation-tested), but not under natural extinction. Similar associations were observed between neural features and behavioral expression. These findings demonstrate that fear learning reshapes spontaneous brain-state dynamics and that such learning-induced reorganization serves as an interpretable biomarker for neuromodulation-linked extinction-memory expression.

neuroscience↗

bioRxiv2026

Differentiable Gene Set Enrichment Analysis for Pathway-Level Supervision in Transcriptomic Learning

In transcriptomics-driven drug discovery, upstream predictors of chemical-induced transcriptional profiles (CTPs) are typically trained with gene-wise objectives, whereas downstream interpretation relies on pathway-level, rank-based statistics such as Gene Set Enrichment Analysis (GSEA). This objective mismatch destabilizes pathway conclusions under prediction errors: small ranking perturbations can flip enrichment direction or distort pathway ordering. To bridge this gap, we present differentiable GSEA (dGSEA), a training-compatible surrogate that maps predicted gene-level scores to pathway enrichment with well-behaved gradients. Technically, dGSEA replaces discrete ranking operations with temperature-controlled soft sorting, smooth prefix accumulation, and differentiable extremum aggregation. Critically, to preserve the statistical semantics of classical GSEA, we introduce sign-specific robust permutation normalization (dNES) with optional{kappa} -calibration. For computational efficiency, a scalable Nystrom-window approximation (nyswin) reduces the quadratic bottleneck to near-linear complexity, enabling genome-scale evaluation. Empirically, across synthetic benchmarks and LINCS L1000 signatures, dGSEA matches classical GSEA accuracy with improved numerical stability. When incorporated as an auxiliary objective for SMILES-to-transcriptome prediction, dGSEA improves pathway-level agreement (macro correlation 0.257 [->] 0.306; sign accuracy 0.620 [->] 0.641) without compromising gene-level performance, providing a practical mechanism for pathway-aware optimization in transcriptomic prediction pipelines.

bioinformatics↗

bioRxiv2026

The Spatiotemporal Proteome Landscape of Aging: Structural determinants of age-sensitive proteome remodeling

Aging is marked by a decline in cellular functions accompanied by widespread changes in mRNA and protein abundance, yet whether aging broadly remodels subcellular protein localization and concentration--and why some proteins change while others remain stable--remains unclear. This gap matters because cellular function depends not only on expression levels but also on correct spatial organization. Using yeast replicative aging as a model, we built a robotic pipeline to enrich old cells from 5,661 strains, acquired 90 million single-cell 3D images, and applied machine learning to map proteome-wide changes in localization, concentration, and aggregation throughout aging. This age-resolved single-cell atlas uncovers widespread proteome remodeling and rewiring of protein interaction networks. Moreover, structural analysis reveals biophysical determinants of age-sensitive proteome remodeling across ages and species. Together, these results reveal a structure-encoded intrinsic principle underlying spatial proteome breakdown during aging and provide a resource to dissect mechanistic links among aging hallmarks.

systems biology↗

bioRxiv2026

Convergent effects of neurodevelopmental disorder-associated variants at mitochondria

Investigations into the molecular pathogenesis of clinically defined neurodevelopmental disorders (NDDs) including autism spectrum disorders (ASD) and schizophrenia (SCZ) have produced evidence implicating dysfunctional mitochondrial metabolism. However, the functional connection between risk variants and mitochondrial proteins largely remains unclear. We tested the hypothesis that proteins encoded by NDD-associated copy number variants (CNVs) and SCZ risk genes are enriched within the mitochondrial interactome. We found that NDD- and SCZ-associated genes exhibit mitochondrial association comparable to their overlap with synaptic proteins, with interaction networks converging most strongly on mitochondrial translation. Two high-risk CNVs, the 3q29 deletion (3q29Del) and the 22q11.2 deletion (22q11Del), confer similar risks for ASD and SCZ and have independently been linked to mitochondrial phenotypes. To test whether these CNVs produce convergent effects on mitochondrial proteins in developing human neural tissue, we generated an isogenic series of 3q29Del and 22q11Del induced-pluripotent stem cells (iPSCs) and differentiated them into forebrain cortical organoids. Quantitative proteomic analysis showed high similarity in the profiles of dysregulated proteins in 3q29Del and 22q11Del compared to isogenic controls. Enrichment analysis of proteins altered in both variants revealed significant convergence on the mitochondrial ribosome and translation machinery. Furthermore, manipulation of mitochondrial translation elicited similar proteomic and functional responses in organoids and neural progenitor cells across both CNVs. These findings indicate that NDD-associated genes have rich interactions with mitochondrial proteins and that two of the strongest risk factors for NDDs may similarly disrupt neural mitochondrial metabolism through impaired mitochondrial translation.

neuroscience↗

bioRxiv2026

Cell type-resolved proteomics reveals intra- and intercellular signalingin Alzheimers disease

Alzheimers disease (AD) arises from pathological interactions among diverse brain cell types, but cell-specific proteomic changes remain underexplored. Here, we present deep proteomic profiling of sorted or proximity-labeled brain cells from AD mouse models (5xFAD and AppNL-G-F) at multiple ages, quantifying 13,411 proteins in microglia (three subtypes), astrocytes, oligodendrocyte precursor cells, and neurons. We identified 3,028 differentially abundant proteins across these cell types, the majority of which were not detected in bulk proteomic datasets, and constructed cell type-specific networks to define functional modules and hub proteins. Comparison with transcriptomic data revealed that [~]30% of proteomic changes are RNA-independent. Further analyses uncovered cross-cell type signaling proteins conserved in human AD brains, such as pleiotrophin (Ptn), which is transcriptionally enriched in astrocytes but accumulates in microglia. Importantly, recombinant PTN directly activates induced microglia-like (iMG) human cells. Thus, these findings provide a comprehensive cell type-resolved proteomic atlas of AD models, highlighting novel intra- and intercellular signaling events. HighlightsO_LIA high-resolution cell type-resolved proteomic atlas of Alzheimers disease mouse models C_LIO_LI[~]3,000 cell type-specific protein alterations identified beyond bulk tissue analyses C_LIO_LIProteomic profiling of microglial subtypes reveals subtype-specific changes in Alzheimers disease C_LIO_LIAstrocyte-microglia signaling is highlighted and validated through PTN-mediated interactions C_LI

neuroscience↗

bioRxiv2026

Mapping Gene Drive Dynamics onto Mendelian Models

CRISPR-based gene drives bias their own transmission and can spread even when deleterious, giving rise to evolutionary dynamics that can be substantially more complex than those governed by standard Mendelian inheritance. Identifying conditions under which gene-drive dynamics can be faithfully approximated by Mendelian models would therefore enable the extensive theoretical toolkit of classical population genetics to be applied to gene-drive systems. Here, we develop a general mapping framework that translates gene-drive models into dynamically equivalent Mendelian models, allowing their behavior to be analyzed using classical theory. By deriving both haploid and diploid effective-parameter mappings, we identify Mendelian models that closely reproduce allele-frequency trajectories of gene drives across a wide range of conversion rates, fitness costs, and dominance effects. We delineate the regions of the parameter space where a one-parameter haploid approximation provides an accurate first-order representation, and where incorporating dominance in a diploid mapping substantially improves fidelity and recovers internal equilibria and threshold behavior. Analytic approximations yield efficient mappings across most of the drive parameter space, while a trajectory-based grid search further improves accuracy near nonlinear regime boundaries. To demonstrate the utility of this framework, we apply it to predicting gene swamping in a two-deme migration-selection model and show that the mapped Mendelian system accurately forecasts transitions between fixation and loss under three relevant release scenarios: environmental variation in fitness, engineered fitness asymmetries, and environment-dependent conversion. Together, these results establish a theoretical bridge between non-Mendelian gene drives and classical population genetic models, providing an interpretable and computationally efficient foundation for predicting gene-drive outcomes and guiding the design of gene drive systems and deployment strategies.

evolutionary biology↗

bioRxiv2025

scCotag: Diagonal integration of single-cell multi-omics data via prior-informed co-optimal transport and regularized barycentric mapping

Recent advances in high-throughput single-cell technologies have enabled characterization of cellular states across distinct omics layers, yielding complementary insights into the organization of biological systems. To jointly leverage these disparate modalities, computational methods have been developed to align non-overlapping cell populations profiled in different omics layers, a task known as diagonal integration, thereby facilitating more fine-grained biological interpretations. However, existing integration approaches unrealistically assume that all cells are alignable and treat prior-knowledge-derived feature correspondences as fully reliable, retaining them unrefined throughout the alignment process. In this work, we introduce scCotag, a co-optimal transport (COOT)-based deep learning frame-work for diagonal integration of single-cell multi-omics data, to address these limitations. scCotag first infers cell alignment and feature correspondence with prior-informed COOT in an iterative manner, and then leverages the resulting transport plans to jointly learn cell and feature embeddings via regularized barycentric mapping and graph reweighting. Systematic benchmarking on human brain, bone marrow, and blood single-cell RNA-seq and ATAC-seq datasets demonstrates the overall superior performance of scCotag over state-of-the-art methods in both cell alignment accuracy and embedding accuracy. In particular, scCotag excels in simulated imbalance scenarios where not all cells are alignable. Applying scCotag to data in postmortem brains of Alzheimers disease (AD) and non-AD (NoAD) patients further yields more fine-grained biological interpretation of regulatory mechanisms. Together, these results demonstrate that scCotag provides a robust framework for single-cell diagonal integration and regulatory inference.

bioinformatics↗

bioRxiv2025

Reinforcement learning for adaptive control of phenotypically heterogeneous bacterial populations

Bacterial populations display extraordinary resilience to antibiotic stress, driven by diverse physiological states that allow some cells to persist and later repopulate. This phenotypic heterogeneity, amplified by environmental fluctuations, undermines the effectiveness of conventional fixed-dose treatment regimens. To address this challenge, we introduce a reinforcement learning (RL) framework that discovers adaptive treatment strategies using only experimentally accessible, population-level measurements. The RL agent learns to infer the hidden physiological state of the population and leverages this knowledge to maintain control even under conditions not encountered during training. Moreover, when granted control over nutrient availability, an important driver of physiological change often overlooked in antibiotic treatment protocols, the agent consistently drives population extinction, surpassing adaptive protocols based solely on drug dynamics. This computational framework offers a powerful, data-driven approach for designing adaptive treatment strategies to counter the growing threat of antimicrobial resistance.

bioinformatics↗

bioRxiv2025

Neural Representation of Associative Threat Learning in Pulvinar Divisions, Lateral Geniculate Nucleus, and Mediodorsal Thalamus in Humans

Understanding the neural mechanisms underlying associative threat learning is essential for advancing behavioral models of threat and adaptation. We investigated distinct activation patterns across thalamic pulvinar divisions, the lateral geniculate nucleus (LGN), and the mediodorsal thalamus (MD) during associative threat learning using fMRI. The anterior pulvinar and MD exhibited parallel activation patterns, that may reflect distinct contributions to automatic and more deliberative learning processes. Additionally, our findings suggest a hierarchical functional organization of pulvinar activation during fear conditioning, in which coordinated activation among inferior, lateral, medial, and anterior divisions may support the integration of threat-related information. Pulvinar divisions and the MD showed activation during extinction learning and exhibited activation patterns consistent with salience processing during extinction recall and threat renewal. LGN activation patterns were consistent with feedforward visual processing. These findings extend current models of threat learning and memory by suggesting distinct patterns of thalamic involvement across associative threat learning and memory and providing a functional framework for investigating thalamic contributions to these processes.

neuroscience↗

bioRxiv2025

Epidermal Stem Cells Control Periderm Injury Repair via Matrix-Driven Specialization of Intercellular Junctions

Epidermal stem cells interact with the extracellular matrix (ECM) to regulate their differentiation and maintain skin architecture. Here, we demonstrate a novel role for basal epidermal stem cells (BECs)-ECM interaction in regulating adhesion molecules expressed by the periderm--the superficial epidermal cells (SECs) of the embryonic bilayered skin. Using the developing zebrafish fin fold, we identify BECs form distinct regions of collagen-versus laminin-enriched basement membranes through integrin-mediated adhesions. Mechanistically, collagen-associated BECs form desmosomes and adherens junctions (AJs) with SECs while laminin-associated BECs display reduced desmosomes but sustain AJs and actomyosin expression with SECs. Notably, we show both in vivo and in a bilayered human keratinocyte model, that laminin, compared to collagen, is sufficient to repress desmosome formation while sustaining AJs specifically at the interlayer cell contacts. In vivo, laminin deficiency enhances desmosome expression across layers and impairs the wound-healing capacity of SECs. This defect was partially rescued by genetic reduction of the desmosome protein Desmoplakin-1a, highlighting the role of ECM-dependent junctional specialization in mediating differences in SEC injury response. Overall, our findings identify that stem cells, through their matrix, establish specialized junctions in the overlying stratified epithelium, which contribute to skin healing properties.

developmental biology↗

bioRxiv2025

A Deep Learning Framework for Predicting Functional Visual Performance in Bionic Eye Users

AO_SCPLOWBSTRACTC_SCPLOWEfforts to restore vision via neural implants have outpaced the ability to predict what users will perceive, leaving patients and clinicians without reliable tools for surgical planning or device selection. To bridge this critical gap, we introduce a computational virtual patient (CVP) pipeline that integrates anatomically grounded phosphene simulation with task-optimized deep neural networks (DNNs) to forecast patient perceptual capabilities across diverse prosthetic designs and tasks. We evaluate performance across six visual tasks, six electrode configurations, and two artificial vision models, establishing our CVP approach as a scalable pre-implantation method. Several chosen tasks align with the Functional Low-Vision Observer Rated Assessment (FLORA), revealing correspondence between model-predicted difficulty and real-world patient outcomes. Further, the CVP paradigm exhibited strong correspondence with psychophysical data collected from normally sighted subjects viewing phosphene simulations, capturing both overall task difficulty and performance variation across implant configurations. Comparing frozen-feature linear probing with full end-to-end fine-tuning reveals that preserving natural-image representations--rather than adapting them to phosphene-specific statistics--better reproduces human perceptual behavior, consistent with the constrained plasticity of adult visual cortex. The findings position CVP as a scientific tool for probing perception under prosthetic vision, an engine to inform device development, and a clinically relevant framework for pre-surgical forecasting.

neuroscience↗

bioRxiv2025

Bona fide hematopoietic stem cells in zebrafish originate from the supra-intestinal artery

Hematopoietic stem cells (HSCs) in vertebrates are widely believed to originate from hemogenic endothelial cells (HECs) in the dorsal aorta (DA) through endothelial-to-hematopoietic transition. However, conclusive in vivo DA-specific lineage-tracing studies to confirm this origin are lacking. Here, leveraging infrared light-mediated temporospatial-resolution lineage-tracing technology, we demonstrate that in zebrafish, HSCs do not originate from the DA. Instead, following the DA hematopoiesis, which produces hematopoietic precursors with limited renewal potential, lifelong renewable HSCs emerge from the supra-intestinal artery (SIA). Remarkably, while the DA produces hematopoietic cells from unipotent HECs, the SIA generates hematopoietic cells from both unipotent HECs and bipotent vessel-resident hemangioblasts (VRHs). These VRHs undergo multiple rounds of division and budding, generating both hematopoietic cells and SIA-resident endothelial cells. The distinct characteristics of DA and SIA cells are further reinforced by single-cell transcriptomic profiling. Our findings challenge the current paradigm of HSC development, underscoring the complexity of vertebrate hematopoiesis.

developmental biology↗