Biology subjects

Zhao, W.

Publications and source records attributed to Zhao, W..

At least 37 records · Page 2Linked to original sources

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

Targeting NAT10 Activates Tumor-Intrinsic Immunity and Suppresses Tumor Progression in Head and Neck Squamos Cell Carcinoma

Head and neck squamous cell carcinoma (HNSCC) remain a major clinical challenge due to its high heterogeneity and limited therapeutic response, resulting in a 5-year overall survival rate of only [~]50%. Identifying molecular pathways that drive tumor progression while suppressing anti-tumor immunity is therefore critical for developing more effective therapies. N-acetyltransferase 10 (NAT10) is the only known enzyme responsible for catalyzing the RNA modification N4-acetylcytidine (ac4C) on rRNA, tRNA, and mRNA, and has been implicated in tumor progression in several cancers. In this study, we identify NAT10 as a key suppressor of tumor-intrinsic immune signaling in HNSCC. NAT10 expression was significantly elevated in tumor tissues and HNSCC cell lines and was associated with poor overall survival. Moreover, high-risk HPV, a major etiological factor in HNSCC, upregulated NAT10 protein expression through the viral oncoproteins E6 and E7. Functional inhibition of NAT10, either by genetic depletion or the small-molecule inhibitor Remodelin, activated tumor-intrinsic innate immune responses, as evidenced by increased IRF3 phosphorylation and induction of type I/II interferons and interferon-stimulated genes. Depletion of NAT10 was able to suppress tumorigenic phenotypes, including cell proliferation, migration, and colony formation in HNSCC cells. Importantly, activation of the STING signaling pathway using agonist cyclic di-GMP further amplified immune activation in NAT10-inhibited cancer cells. Together, our findings establish NAT10 as a previously unrecognized negative regulator of tumor-intrinsic immunity in HNSCC and support NAT10 targeting, particularly in combination with STING agonists, as a promising immunotherapeutic strategy.

Cancer Biology↗

bioRxiv2026

Identifying Ecological Restoration Priority Areas through a Function-Structure-Dynamic Framework: Integrating Ecological Security Patterns and Future Land-Use Simulation in the Anhui Section of the Yangtze River Basin

Rapid urbanization has intensified ecological degradation and threatened ecosystem sustainability in the Yangtze River Basin. This study proposes a Function-Structure-Dynamic (FSD) framework integrating ecological protection importance assessment, ecological security pattern (ESP) construction, and future land-use simulation to identify ecological restoration priorities in the Anhui section of the Yangtze River Basin. Ecologically important areas were identified by coupling ecosystem services and ecological vulnerability. MSPA, landscape connectivity analysis, resistance surface modeling, and circuit theory were applied to construct the ESP, while the PLUS model simulated land-use change and ecological risks in 2032. Results identified 37 ecological sources covering 6,901.39 km{superscript 2} and 84 ecological corridors with a total length of 916.80 km. Ecological pinch points, barrier areas, and 17 ecological warning areas were further delineated. Based on these findings, four ecological restoration zones were proposed with differentiated management strategies. The FSD framework effectively integrates ecological functions, spatial structure, and future dynamics, providing scientific support for ecological restoration and sustainable territorial spatial planning in rapidly urbanizing regions.

ecology↗

bioRxiv2026

Scalable in vivo cardiac functional genomics with compressed AAV-Perturb-seq reveals a common mitochondrial response to perturbation

Efficient identification of new targets to treat human disease requires a scalable way to link genotype to phenotype directly in the target organ. Pooled CRISPR screening with single-cell RNA sequencing as a readout (Perturb-seq) has emerged as a method for functional genomics but is typically applied in vitro and is limited in scale. Here, we combine in vivo Perturb-seq via adeno-associated virus (AAV)-mediated delivery with a statistical framework allowing for signal deconvolution after multiple random perturbations per cell (compressed Perturb-seq), to develop in vivo compressed AAV-Perturb-seq, a scalable way to perform high-throughput, cell-autonomous functional genomics in a desired target organ. We apply this approach to study the effect of 585 gene knockouts on the cardiomyocyte transcriptome. We identify that alterations in the mitochondrial transcriptome are a common response to genetic perturbation, a finding independently validated across species and perturbation modalities. We identify very few synergistic perturbations, despite observing frequent combinatorial effects. Broadly, our work establishes a platform to facilitate in vivo functional genomics in a target organ with direct applicability to identifying therapeutic targets for the treatment of human disease.

genomics↗

bioRxiv2026

Deep learning-guided design of hydrolases for crystalline PET depolymerization

Poly(ethylene terephthalate) (PET), a ubiquitous polyester used in packaging and textiles, persists in the environment due to its high crystallinity and stability, contributing substantially to global plastic pollution. Enzymatic depolymerization by PET hydrolase (PETase) offers a chemically precise and environmentally sustainable route to convert PET into its monomeric building blocks, enabling recycling. However, practical implementation remains hindered by the rigid, crystalline architecture of PET, which restricts enzyme access and necessitates energy-intensive pretreatment to enable efficient depolymerization. In addition, most PETases achieve only partial depolymerization with low terephthalic acid (TPA) yields and accumulate inhibitory intermediates, while limited thermostability and slow surface kinetics further restrict efficiency. To overcome these barriers, we report VenusPETase, an engineered variant of KbPETase designed using PET-Flow, a state-of-the-art computational framework for PETase engineering. Compared with KbPETase, VenusPETase exhibits a 3.4-fold increase in hydrolytic activity, up to a 27-fold improvement after heat treatment of protein, and a 12 {degrees}C enhancement in thermostability. VenusPETase exhibits rapid degradation across a wide crystallinity range (8%-50%) at 50 {degrees}C, effectively spanning the entire spectrum of commercial PET products. Moreover, VenusPETase outperformed nine high-performance PETases under their respective optimal conditions and degraded untreated PET substrates across 8%-50% crystallinity, producing TPA as over 95% of the released products with minimal intermediate accumulation. X-ray crystallography and molecular dynamics simulations suggest that dynamic modulation, elevated surface electrostatic potential enhance the interaction of VenusPETase with crystalline PET, thereby lowering the hydrolytic energy barrier and improving catalytic performance. We also demonstrate that untreated, postconsumer-PET from nine different products can all be degraded by VenusPETase. The recovered monomers can be directly repolymerized into virgin-quality PET, demonstrating a closed-loop enzymatic recycling process. In a 100 L bioreactor, VenusPETase completely depolymerizes post-consumer crystalline PET (28% crystallinity) within 24 h under 50 {degrees}C. These results establish VenusPETase as a robust biocatalyst that enables efficient, closed-loop recycling of crystalline PET under mild conditions.

biochemistry↗

bioRxiv2026

BRCA1 promotes homologous recombination through separable DNA and RAD51 binding activities within its disordered region.

The breast and ovarian tumor suppressor BRCA1 heterodimerizes with BARD1 to promote DNA double-strand break repair by homologous recombination (HR) and to protect stressed DNA replication forks against nuclease attack. The large, intrinsically disordered central region of BRCA1 harbors binding sites for DNA and multiple repair factors, but its lack of stable structure has hindered mechanistic dissection of these activities. Using biochemical mapping and NMR spectroscopy, we delineate the DNA binding and RAD51 interaction interfaces within this region and construct separation-of-function mutants that selectively ablate each activity. Both DNA binding and RAD51 interaction are required for BRCA1-BARD1 to promote RAD51-mediated DNA strand invasion, and DNA binding also contributes to BLM-DNA2 end resection. These findings provide mechanistic insights into how individual ligand binding activities within BRCA1 contribute to genome maintenance.

biochemistry↗

bioRxiv2026

Hierarchical Interplay between H3K27ac and H3K4me3 in Transcriptional Regulation

H3K27ac and H3K4me3 are enriched at transcriptional start sites and have been implicated in transcription. However, how these marks concertedly regulate transcription is not fully understood. Here, we developed a dual chemically inducible CRISPR/dCas9-based epigenome editing system that enables independent, temporal and transcription stage-specific modulation of H3K27ac and H3K4me3 at a specific gene locus. Stage-specific removal of H3K4me3 impaired RNA polymerase II recruitment, increased promoter-proximal pausing, reduced productive elongation, and accelerates mRNA decay via increased m6A deposition. Losing both H3K27ac and H3K4me3 rapidly abolished transcriptional activity, while preserving H3K4me3 without H3K27ac can partially sustain transcription. These findings revealed a functional hierarchy and interdependence between H3K27ac and H3K4me3 in different transcription stages and the established versatile tool will contribute to the functional dissection of the temporal dynamics of chromatin modifications in gene regulation.

genomics↗

bioRxiv2026

Guidance for high-quality functional gene embeddings from large language models

Large language models (LLMs) are increasingly used to generate gene embeddings, yet systematic benchmarks of prompting strategies and practical guidance for obtaining biologically meaningful representations remain limited. Here we present GEbench, an evaluation framework for assessing LLM-derived gene embeddings across different tasks, prompting strategies, and LLM architectures. GEbench revealed that embedding quality depends primarily on whether the input text contains explicit functional information, rather than on sparse gene identifiers or model size. Identifier-based embeddings showed weak biological organization, whereas embeddings derived from functional descriptions consistently achieved stronger functional separation and predictive performance. Notably, Self-Des, which extracts embeddings from model-generated gene function descriptions, enabled locally deployable LLMs to generate high-fidelity representations that approach the quality of expert-curated databases. Genome-scale analyses further supported these findings, indicating that explicit functional descriptions are an effective design principle for generating high-quality gene embeddings from LLMs.

bioinformatics↗

bioRxiv2026

Inhibition of oogenic JNK preserves fertility and ovarian hormones during DNA-damaging cancer therapy

Primary ovarian insufficiency (POI) and related infertility, early menopause, and endocrine disorders due to hormonal deficiency are major side effects in young female cancer patients undergoing cancer therapy. Current strategies preserving the fertility and hormonal functions of the ovary remain imperfect due to concerns of feasibility, efficacy, or safety. Herein, we identified c-Jun N-terminal kinase (JNK) as a pivotal regulator of the DNA damage response (DDR) signaling in oocytes of primordial follicles in response to DNA-damaging cancer therapy. Using pharmacological JNK inhibition and a genetically modified mouse model with oocyte-specific JNK deletion, together with histological, bioinformatic, and molecular approaches, we demonstrated that JNK inhibition prevented chemotherapy-induced oocyte apoptosis and POI, and preserved long-term reproductive cycles and fertility. Mechanistically, JNK was activated in response to chemotherapy-induced DNA damage in oocytes of primordial follicles, causing activation of transcription factor TAp63 and subsequent oocyte apoptosis, ultimately resulting in diminished ovarian reserve and POI. A more clinically relevant breast cancer-bearing mouse model revealed that JNK inhibition preserved the ovarian reserve without compromising anti-cancer efficacy of chemotherapy. Together, our study identifies oocyte-intrinsic JNK as a promising target for developing ovarian protectants and safeguarding reproductive health and fertility in young female cancer survivors.

pharmacology and toxicology↗

bioRxiv2026

CHORD: a framework for cross-species single-cell integration across gene, cell and cell-type levels

Quantifying cross-species relationships among cell types from single-cell transcriptomic data can reveal both conserved and divergent patterns of cell-type hierarchies. However, existing cross-species integration methods can be limited in modeling genes beyond orthologs by leveraging cell-type-resolved transcriptional context, or in learning explicit type-level representations. Here we present CHORD, a cross-species integration framework that jointly learns representations of genes, cells and cell types. We demonstrate that CHORD can integrate cross-species single-cell atlases and support cell-type annotation with unknown cell-type detection. In the frog-zebrafish embryogenesis and mammalian motor cortex atlases, CHORD infers cell-type trees that place conserved cell types from different species in relative proximity and summarize hierarchical relationships among cell types. CHORD also supports cross-species comparison of continuous phenotypic variation by placing embryonic cells along an aligned developmental timeline. CHORD further yields gene embeddings that capture orthologous and functional relationships, and gene importance scores linking genes to cell types.

bioinformatics↗

bioRxiv2026

TopicVI: A Knowledge-guided deep interpretable model for resolving context-specific gene programs

Mechanistic insights from single-cell and spatial transcriptomics largely rely on cell clustering, differential expression analysis, and interpretation through prior biological knowledge. However, this approach is often limited by the reliance on curated biological priors that fail to capture context-specific gene programs, particularly in complex disease states. To address this gap, we introduce TopicVI, a deep interpretable model that integrates established biological knowledge with data-driven refinement to discover context-dependent gene programs in single-cell and spatial transcriptomic data. TopicVI jointly infers cell clusters and gene topics using optimal transport to flexibly align prior gene programs with observed data while permitting context-specific refinements. Comprehensive benchmarking demonstrates that TopicVI outperforms existing methods in biological conservation, batch correction, topic coherence, and rare cell identification. TopicVI effectively disentangles multiple sources of biological variation, such as separating anatomy-specific expression patterns from disease-associated signatures in spatial transcriptomics. Applying TopicVI to glioblastoma datasets, we identify gene topics related to cell cycle regulation and EGFR signaling that reveal convergent tumor states across distinct drug perturbations. By integrating prior knowledge with data-driven discovery, TopicVI enables identification of interpretable gene programs that illuminate biological processes and therapeutic mechanisms in complex transcriptomics data.

bioinformatics↗

bioRxiv2026

Target RNA abundance controls the collateral activity of RfxCas13d in human cells and zebrafish embryos

Collateral RNA cleavage by CRISPR-Cas13 effectors presents a critical obstacle to their application in biological research and therapeutics, yet the molecular determinants of this activity remain poorly understood. Here, we systematically investigate the collateral activity of various Cas13 variants in human cells in vitro and zebrafish embryos in vivo. Among these nucleases, RfxCas13d displays robust collateral activity that is highly dependent on target abundance. Targeting moderately expressed RNAs activates only a limited subset of RfxCas13d molecules, resulting in selective degradation of ectopically expressed transcripts, while endogenous RNAs remain largely protected. This selectivity indicates higher accessibility to exogenous RNAs through spatial proximity and temporal colocalization with activated RfxCas13d nuclease domains. In contrast, the recognition of highly abundant RNA targets drives simultaneous activation of a large fraction of cellular RfxCas13d, leading to widespread collateral cleavage of cellular RNAs, disruption of proteome homeostasis, and consequent cell toxicity and developmental defects in zebrafish embryos. Notably, transgenic zebrafish with target RNA expression restricted to endothelial or neuronal cell lineages exhibit localized collateral activity, leading to tissue-specific developmental abnormalities and motility deficits. These findings reveal that RfxCas13ds collateral activity is threshold-dependent, with abundant target RNA acting as a molecular switch for widespread collateral RNA degradation. This work underscores the need for careful consideration of target abundance when deploying RfxCas13d, and highlights PspCas13b as an alternative for RNA silencing free of collateral activity.

molecular biology↗

bioRxiv2026

A self-complementary recombinant adeno-associated virus vector coding for an anchorless prion protein carrying the G127V mutation extends survival in a rodent prion disease model

The replacement of a single codon in the human prion gene, causing the substitution of glycine with valine at position 127 (G127V) of the prion protein (PrP), prevents development of prion disease. We set out to explore if prion disease survival extension manifests in mice if the V127 mutant is delivered through a recombinant adeno-associated virus (rAAV) packaged as a self-complementary DNA. The notorious delivery limitations of rAAVs were overcome using a cross-correction approach that relied on the expression of the mutation in the context of glycosylphosphatidylinositoI-anchorless ({Delta}GPI) PrP. In this proof-of-concept study, we inoculated Rocky Mountain Laboratory (RML) prions into knock-in mice, in which the endogenous murine prion protein gene (Prnp) was replaced with the bank vole prion protein gene (BvPrnp). Prion-inoculated mice that were retro-orbitally transduced with a protective rAAV vector encoding BvPrnpV127{Delta}GPI survived [~]50 days longer than control mice that were unprotected. A deep proteomic analysis revealed that BvPrnpV127{Delta}GPI was protective by slowing perturbations to the proteome observed in late-stage RML prion disease. In addition to capturing details of synaptic decay and depletion of proteins in proximity to PrP, the proteomic dataset revealed the identity of proteins of potential diagnostic value that may be central to the brains attempt to fight prion disease by contributing to astrocytosis or microgliosis, by coping with calcium influx, or by enhancing the endoplasmic reticulum processing of essential proteins. Taken together, our results demonstrate that a gene therapy based on a GPI-anchorless PrP containing the G127V mutation can delay the onset of prion disease in mice, providing a framework for development of a corresponding therapy in humans. AUTHOR SUMMARYA rare change in the human prion protein, involving a single building block, has been linked to strong protection against prion diseases--fatal neurodegenerative disorders. This study tested whether that protective effect could be reproduced using gene therapy in mice. To this end, we exposed the animals to infectious prions and then delivered the protective version of the protein into mice using a viral carrier. Treated mice survived about seven weeks longer than untreated animals, showing that the approach can meaningfully slow disease progression. To understand why, we examined changes in brain proteins during disease and found that treatment helped preserve the normal protein levels of cellular proteins, particularly those involved in communication between nerve cells. The analysis also identified proteins altered in the disease that are linked to the brains defense responses, including inflammation, stress handling, and protein processing, some of which may serve as future disease markers. Importantly, the limited protection observed was not due to poor delivery of the therapy but likely reflects biological limits of the model used. Overall, the findings support the idea that gene therapies based on naturally protective human variants could help slow prion diseases and improve understanding of how the brain responds to them.

neuroscience↗

bioRxiv2026

Dingent: An Easily Deployable Database Retrieval and Integration Agent framework

AI-driven data search and integration represent an emerging research direction. Although several LLM-based backend frameworks and agentic frameworks have emerged, significant gap remains in developing a one-stop, configurable agent framework that supports various data sources and provides a web interface for efficient data retrieval using natural language. To address this, we present Dingent, a novel and configurable agent framework that facilitates data access from various resources and enables the flexible constructions of agent applications. We demonstrate its capabilities across three distinct application scenarios, achieving promising results. The Dingent framework can be readily applied to other fields, such as earth sciences and ecology, to facilitate data discovery.

bioinformatics↗

bioRxiv2026

Learning Continuous Morphological Trajectories via Latent Principal Curves

Inferring continuous morphological transformations from collections of static biological snapshots is an important, yet challenging problem. In the context of cellular biology, prevailing approaches reduce 3D shape collections to static reconstructions or hand-crafted descriptors, which fail to capture smooth, multi-dimensional transitions. We present MorphCurveVAE, a two-stage pipeline for constructing continuous morphological trajectories from sets of static, segmented 3D microscopy images. Stage 1 learns a smooth, compact latent manifold of volumetric morphologies using a multi-branch convolutional variational auto-encoder (VAE) that can encode multiple correlated substructures into disentangled subspaces. Stage 2 extracts a constrained, topologically-aware principal curve through the augmented latent space to produce directional and correlated trajectories of structural dynamics. To demonstrate our framework, we apply MO_SCPLOWORPHC_SCPLOWCO_SCPLOWURVEC_SCPLOWVAE to a large public dataset (Allen Institute WTC-11) of segmented volumetric cell and nucleus images spanning the mitosis cycle. Our results indicate high-quality reconstructions, low projection errors to the fitted principal curve, and biologically and visually plausible continuous animations. These results suggest MorphCurveVAE as a practical tool for modeling biological morphological trajectories, while remaining broadly applicable to other biological imaging domains where time-resolved observations are unavailable.

biophysics↗

bioRxiv2026

Cell type-specific gene regulatory network inference from single cell transcriptomics with ctOTVelo

Inferring gene regulatory networks (GRNs) from gene expression is a crucial task for understanding functional relationships. Gene expression data (transcriptomics) provide a snapshot of gene activity, encoding information about gene regulatory relationships. However, gene regulation is a dynamic process, modulating across time and with different cell types. Temporal GRN inference methods aim to capture these dynamics by utilizing time-stamped transcriptomics, gene expression data of similar samples captured across discrete timepoints, or pseudotime transcriptomics, computationally ordering cells based on an inferred trajectory. These methods can estimate constant or temporal gene regulatory relationships, but may not capture finer, cell type specific relationships. We propose ctOTVelo, an extension to our previous work to account for cell type specificity during GRN inference. ctOTVelo incorporates cell type labels or proportions when inferring the GRN from single cell transcriptomics data. Our methods achieve state-of-the-art performance in GRN prediction in time-stamped and pseudotime-stamped transcriptomics. Furthermore, ctOTVelo is able to generate cell type specific GRNs, allowing cell type resolution analysis of gene regulatory relationships.

genomics↗

bioRxiv2026

IFI207 promotes antiviral responses by modulating STING ubiquitination and degradation

Aim2-like receptors (ALRs) play crucial roles in innate immune signaling pathways and demonstrate strong positive selection likely driven by pathogens. IFI207, an ALR found in all Mus species, enhances interaction with and stabilization of STING, contributing to the control of Murine Leukemia Virus (MLV) infection. We show here that IFI207 enhances the type 1 interferon response by inhibiting activation-induced K63-linked ubiquitination of STING, thereby preventing its recognition by hepatocyte growth factor-regulated tyrosine kinase substrate (HRS), a key component of the ESCRT complex, and its subsequent degradation in lysosomes. IFI207 promotes downstream signaling in the STING pathway in multiple cell types and moreover enhances the STING-dependent response to herpesvirus simplex 1 infection ex vivo and in vivo. We also show that IFI207 likely functions in dendritic cells to suppress MLV infection. Our study reveals that IFI207 acts as a modulator in the STING pathway, strengthening the hosts defense against viral infections and suggests that the expansion of the Alr locus in mice may have occurred in response to endemic viruses.

immunology↗

bioRxiv2026

Connectome lateralization in autism across the first 14 years: heterogeneity related to developmental stage, hemisphere, and pathophysiology

Hemispheric lateralization is fundamental feature to neurodevelopment, yet its dynamic evolution in autism across childhood and underlying molecular mechanisms remain unknown. Thus, using a large-scale resting-state fMRI dataset (n = 1553, ages 1-14), we investigated developmental heterogeneity in connectome lateralization. At the group level, abnormalities in dynamic connectome lateralization strength (DCLS) progressed from focal patterns in early childhood ([≤] 6 years) to widespread distributions by late childhood ([≤] 14 years), independent of symptom severity. The late-stage patterns were validated using the ABIDE dataset (n=153). Late childhood also showed marked increases in inter-individual heterogeneity. These alterations were significantly associated with clinical phenotypes, transcriptomic profiles, and multiple neurotransmitter systems, with distinct patterns across childhood groups. Our findings reveal a developmentally dynamic trajectory of connectome lateralization in autism and highlight late childhood as a critical window for stage- and individual-specific interventions tailored to substantial individual heterogeneity.

neuroscience↗