Biology subjects

Wu, Y.

Publications and source records attributed to Wu, Y..

At least 325 records · Page 18Linked to original sources

bioRxiv2025

Behavioral Time Scale Synaptic Plasticity (BTSP) endows Hyperdimensional Computing with attractor features

Human reasoning depends on reusing pieces of information by binding them together in new ways, thereby "making infinite uses of finite means" (Alexander von Humboldt). Needed for that is a binding mechanism that enables fast composition and decomposition of tokens of information. Binding can easily be implemented in symbolic computations through parentheses and ordering of symbols. But it is a highly nontrivial operation for distributed representations, where the tokens are encoded by activity patterns in large neural networks or large language models, or more abstractly, by a high dimensional vector. Vector Symbolic Architectures (VSAs) provide partial solutions, but are lacking the flexibility of the brain in information retrieval, e.g. retrieving the tokens from a composed representation or retrieval of a composed representation by just providing some tokens as a cue. We show that a mechanism which the brain employs for binding distributed representations, Behavioral Time Scale Synaptic Plasticity (BTSP), overcomes these deficiencies. In particular, it combines binding with attractor features that make information retrieval substantially more flexible and robust. We evaluate its performance on various applications, including encoding and decoding complex visual scenes and hierarchical binding. We also show that it enhances models for natural language processing and abstract brain computation. BTSP-based binding only requires binary synaptic weights and simple local synaptic plasticity, and can therefore easily be implemented through in-memory computing or other innovative designs for energy-efficient AI implementations.

animal behavior and cognition↗

bioRxiv2025

The decameric repeat (DR) of PSGL-1 functions as a basic antiviral unit in restricting HIV-1 infectivity

PSGL-1 (P-selectin glycoprotein ligand-1) is a dimeric, mucin-like surface glycoprotein that interacts with selectins on the endothelium, facilitating leukocyte tethering and rolling during transmigration. Previous studies have demonstrated that PSGL-1 acts as an HIV restriction factor, blocking HIV infectivity by hindering particle attachment to target cells. A large part of PSGL-1s extracellular region consists of a mucin-like domain, characterized by decameric repeats (DR), which comprise 14 to 16 DR tandems. Each DR is composed of 10 consensus amino acids (-A-T/M-E-A-Q-T-T-X-P/L-A/T-) and is notably rich in O-glycosylated threonines (30%) and prolines (10%). A proposed function of the DR is to elongate the protein backbone needed for selectin binding. However, the precise role of DR in PSGL-1s antiviral mechanisms has yet to be fully elucidated. In this study, we performed DR deletion mutagenesis and molecular modeling, systematically deleting from one DR to all DRs, and quantified their effects on PSGL-1s antiviral functions. Here, we demonstrate that DR is necessary for PSGL-1s antiviral activity. Deleting DR did not affect virion incorporation of PSGL-1, but diminished PSGL-1s ability to hinder virion attachment to target cells. We also discovered that individual single DR mutants retained 3.5 to 18% of the antiviral activity exhibited by full-length PSGL-1, and the basic antiviral activity of DR is cumulative; increasing the number of DRs correlates with heightened antiviral activity. Additionally, the intrinsic anti-HIV capability of DR was found to be transferrable; inserting the DR domain into the extracellular Ig-like domain of CD2 permitted the hybrid CD2-DR molecules to partially acquire the anti-HIV properties of PSGL-1. Further molecular modeling utilizing all-atom replica exchange molecular dynamics simulations highlighted a structure-function relationship between the anti-HIV potency of DR and the elongation of the DR backbone, as measured by DR backbone dihedral angles ({varphi}) and the radial distributions of peptide atom number density (rmax). These results collectively suggest that DR serves as an essential antiviral unit within the framework of PSGL-1s restriction mechanism against HIV.

immunology↗

bioRxiv2025

Enhancing Biomedical Optical Volumetric Imaging via Self-Supervised Orthogonal Learning

Optical volumetric imaging grapples with inherent noise problems arising from photon budget constraints, light scattering, and space-bandwidth product bottlenecks, all of which degrade structural fidelity and become even more pronounced than in planar imaging. While deep learning presents potential for denoising, supervised methods are hindered by the impracticality of paired dataset, and existing self-supervised approaches fail to fully exploit the intrinsic volumetric structural redundancy inherent to optical imaging. Here, we present a self-supervised Volumetric biomedicAL Imaging Denoiser (VALID) that leverages intrinsic three-dimensional spatial coherence for highly efficient volumetric denoising through a self-supervised orthogonal learning framework. VALID demonstrates robust denoising performance across diverse imaging modalities, including two- and three-photon microscopy, light-field microscopy, and optical coherence tomography, substantially enhancing structural fidelity in deep-tissue, multimodal, and dynamic imaging scenarios. By combining computational efficiency with zero-shot adaptability, VALID establishes a transformative approach to volumetric image enhancement with structure-aware precision.

neuroscience↗

bioRxiv2025

The NHEJ machinery is translocated off DNA ends toenable resection

Error-prone non-homologous end joining (NHEJ) and homologous recombination (HR) compete to repair DNA double strand breaks [1, 2, 3]. Given that NHEJ factors such as Ku and DNA-PKcs are rapidly recruited to DSBs and sterically occlude ends, they must be removed to enable 5' DNA end resection, the commitment step of HR. While resection is well characterized, how the NHEJ machinery is displaced from DNA ends has remained poorly understood. Here, we performed single-molecule imaging in Xenopus laevis egg extracts to monitor the removal of NHEJ factors from DNA ends while simultaneously tracking the progress of DNA end resection. We find that MRN and CtIP are essential to evict the core NHEJ factor Ku from DNA ends. Surprisingly, while Ku removal and resection of the 5' end are normally coordinated, these processes arise from distinct activities of the MRN complex and are differentially regulated by the action of cyclin dependent kinases and ATM. By directly imaging the position of Ku on DNA using smFRET, we found that Ku is translocated independently of DNA end resection in an ATP-dependent manner that requires both MRN and CtIP. In summary, our results provide a new mechanistic framework to understand how eviction of the NHEJ machinery is coordinated with DNA resection to initiate HR.

biophysics↗

bioRxiv2025

Sequential Amino Acid Mutagenesis-Driven De Novo Evolution of Adenine Deaminases Enables Efficient In Vivo Base Editing in Primate

Base editing allows for the precise modification of genetic information, providing new avenues for treating diseases1. The adenine base editor ABE8e is currently the most efficient and widely used tool for adenine base editing. ABE8e was developed through multiple rounds of directed evolution of Escherichia coli tRNA adenine deaminase, including phage-assisted continuous evolution (PACE)2,3. While PACE is highly effective, it is a complex system that poses challenges for implementation4. Despite its high efficiency, ABE8e is associated with limitations such as relatively higher bystander editing effects and elevated off-target activity5, which need to be addressed to further enhance its precision and safety. Here, we developed a novel method for the de novo discovery of evolved ABE components, particularly adenine deaminases. This process involves identifying candidate proteins through AI-based structural prediction and clustering, followed by the enhancement of deaminase editing activity through screening libraries created by sequential amino acid saturation mutagenesis. This evolutionary strategy simplifies the approach by employing saturation mutagenesis libraries tailored to specific segments, thereby enabling exploration of an expanded sequence space and increasing the likelihood of discovering adenine deaminases with superior capabilities. The newly developed hpABE5.20 here demonstrates a more refined editing window, reduced DNA off-target effects that are both sgRNA-dependent and -independent, and minimized RNA off-target activity, while maintaining robust editing efficiency relative to ABE8e. Furthermore, hpABE5.20 has been successfully applied for precise and effective therapeutic adenine base editing in cellular disease models, humanized mice, and non-human primates.

bioengineering↗

bioRxiv2025

Spatial geometry-aware deep learning for deciphering tissue structure from spatially resolved transcriptomics

Recent advances in spatially resolved transcriptomics have enabled high-throughput gene expression profiling while preserving spatial context, thereby facilitating the investigation of spatial heterogeneity within tissues. Here, we present SpatialGEO, a spatial geometry-aware deep learning framework designed to decipher tissue organizational structures through dual-encoder feature extraction and geometric graph learning. Experimental results demonstrate the superior accuracy of SpatialGEO in identifying spatial regions across datasets encompassing diverse biological contexts and spatial resolutions. Moreover, SpatialGEO uncovers key mechanisms of immune evasion and regulation within the tumor microenvironment and provides novel biological insights into mouse embryonic development.

bioinformatics↗

bioRxiv2025

A Multimodal Graph Learning Framework for Versatile Spatial Transcriptomics Analysis with SpatialModal

The development of spatial transcriptomics(ST) technologies enables the exploration of tissue structure and cellular function within a spatial context. However, mainstream analytical methods primarily focus on the gene expression modality itself and struggle to fully leverage auxiliary information from histological images, which limits precise dissection of complex biological structures. To address this, we propose SpatialModal, a multimodal graph learning framework that effectively integrates gene expression data and histological images from spatial transcriptomics to construct unified spatial representations. SpatialModal effectively captures synergistic interactions between modalities by jointly modeling intra- and inter-modal complementary features while incorporating spatial adjacency information. Furthermore, it employs a dual contrastive learning strategy to enhance the discriminative power of representations, thereby enabling efficient and robust analysis of tissue structures. We validate the effectiveness of this approach on multiple public datasets covering diverse tissue types, species, and resolutions. Experiments demonstrate that SpatialModal exhibits significant advantages in downstream tasks, including spatial domain identification, gene expression reconstruction, and pseudotime inference, and accurately elucidates the hierarchical structure of the mouse cortex as well as the metabolic-immune dynamic boundaries in the breast cancer microenvironment. Additionally, SpatialModal shows exceptional robustness in single-cell resolution and multi-slice integration tasks, providing an efficient and broadly applicable analytical tool for spatial transcriptomics research.

bioinformatics↗

bioRxiv2025

Advanced eTAM-seq enables high-fidelity, low-input N6-methyladenosine profiling in human cells and embryonic mouse tissues

Functional dissection of N6-methyladenosine (m6A), the most abundant internal messenger RNA modification in mammals, demands quantitative, scalable detection technology. We previously reported eTAM-seq, which supports transcriptome-wide quantification of m6A by enzyme-assisted adenosine deamination. While effective, TadA8.20--the enzyme used in our first-generation technology--is sensitive not only to m6A but also to RNA structure, making accurate detection dependent on the inclusion of control transcriptomes. Here, we introduce eTAM-seq-v2, in which we replace TadA8.20 with TadA8r, a further evolved adenosine deaminase with superior catalytic efficiency. eTAM-seq-v2 supports control-free m6A calling with high fidelity. Because enzyme treatment preserves RNA integrity, eTAM-seq surveys >51% of A sites in all expressed genes with moderate sequencing depth (60 million uniquely mapped reads) and delivers robust performance with as little as 10 ng of total RNA ([~]500 cells). With eTAM-seq-v2, we delineate the m6A landscape across six human cell lines and seven embryonic mouse tissues. While uncovering broadly conserved m6A patterns, we reveal that most neighboring m6A sites are independently deposited at the single-molecule level. Moving forward, we envision that eTAM-seq-v2 will enable researchers to survey m6A in diverse biological contexts and uncover new insights into its regulatory roles.

molecular biology↗

bioRxiv2025

A synergistic generative-ranking framework for tailored design of therapeutic single-domain antibodies

Single-domain antibodies (sdAbs) have emerged as powerful therapeutic agents due to their small size, high stability, and superior tissue penetration. However, unlike conventional monoclonal antibodies (mAbs), sdAbs lack an Fc domain, limiting their functional versatility and manufacturability. To address this challenge, we developed TFDesign-sdAb, a deep learning-based generative-ranking framework that enables rational engineering of sdAbs with tailored functionalities. Our framework integrates a structure-aware diffusion model (IgGM) for large-scale candidate generation and a fine-tuned sorter (A2binder) that evaluates and prioritizes them based on predicted functionality. Unlike traditional CDR-focused approaches, TFDesign-sdAb optimizes both complementarity-determining regions (CDRs) and framework regions (FRs), allowing sdAbs to acquire new functional properties while maintaining antigen specificity. We validated our approach by conferring Protein A binding to human VHs and nanobodies that originally lacked this feature, achieving high expression rates, strong binding affinities, and successful purification via industry-standard Protein A affinity chromatography. High-resolution structural characterization (1.49 [A] and 2.0 [A]) of the redesigned sdAb-Protein A complexes revealed conserved FR-mediated binding motifs that recapitulate natural Fc-Protein A interactions, validating the accuracy of our model. Furthermore, our pipeline streamlined the antibody engineering process, achieving a 100% success rate in generating Protein A-binding sdAbs while maintaining their original antigen-binding affinity. This work demonstrates the power of AI-driven design in overcoming long-standing limitations in antibody engineering and presents a scalable, generalizable solution for enhancing sdAb functionality.

bioinformatics↗

bioRxiv2025

Lactate metabolism antagonizes antiviral innate immunity bypromoting VDAC1-lactylation-mediated mtDNA release

Foot-and-mouth disease (FMD) is an acute, highly contagious disease caused by the foot-and-mouth disease virus (FMDV). Due to its zoonotic nature, this disease not only significantly impacts livestock productivity but also poses a serious threat to human health. FMDV exhibits distinct host specificity, mainly infecting cloven-hoofed animals such as cattle, sheep, pigs, and deer, while perissodactyl animals like horses and donkeys show natural resistance. Elucidating the interspecies differences in response to FMDV infection is essential for understanding the pathogenesis of the virus. In this study, we report that FMDV infection in horse epithelial cells induces the production of lactate, which modulates the function of VDAC1 protein through lactylation, facilitating the release of mtDNA from mitochondria into the cytoplasm. This process activates the cGAS-STING signaling pathway, which stimulates interferon-{beta} expression and suppresses FMDV replication. Our findings provide new insights into the mechanism behind horse resistance to FMDV and lay a theoretical foundation for further research concerning the differential pathogenesis of FMDV infection across species.

cell biology↗

bioRxiv2025

Targeting USP2 induces degradation of PML-RARα with or without drug-resistant mutations in acute promyelocytic leukemia

Despite the high efficacy of all-trans retinoic acid (ATRA) and arsenic trioxide (ATO) in treating acute promyelocytic leukemia (APL), approximately 10-20% of patients develop drug resistance due to mutations in PML-RAR and other factors. Here, it is demonstrated that inhibiting USP2 with ML364 or silencing USP2 reduces PML-RAR protein levels in both ATRA-sensitive and -resistant APL cells, an effect reversed by proteasome inhibition. Conversely, USP2 overexpression enhances PML-RAR stability. Mechanistically, USP2 interacts with and deubiquitinates PML-RAR, including drug-resistant mutants. Consistent with PML-RAR degradation, ML364 treatment significantly induces apoptosis in APL cell lines and primary leukemia cells. In conclusion, this study identifies USP2 as a novel deubiquitinating enzyme for PML-RAR and highlights USP2 inhibition as a potential therapeutic strategy for APL with PML-RAR mutations.

cell biology↗

bioRxiv2025

Deep learning guided discovery of antibacterial polymeric nanoparticles

Drug-resistant bacterial infections pose a serious threat to global health, driving the development of novel antibacterial strategies beyond classic antibiotics. Self-assembled polymeric nanoparticles (SANPs) have emerged as promising candidates due to their unique structural properties, which enhance membrane interaction, drug loading, and therapeutic accumulation. The structural diversity and tunable properties of SANPs provide a flexible design platform for antibacterial applications. However, navigating the vast chemical space of SANPs remains a significant challenge due to the complex structure-activity relationships. To address this, we developed PolyCLOVER, a deep learning-guided framework for the efficient discovery of antibacterial SANPs. The framework integrates multi-stage self-supervised learning, active learning, and high-throughput experimentation to iteratively guide the discovery of SANPs with potent antibacterial activity and low toxicity. Through this framework, we identified three potent SANPs within a combinatorial library of [~]100,000 poly({beta}-amino ester)s, with minimum inhibitory concentrations of 4 g/mL and 8 g/mL against clinically multidrug-resistant S. aureus and A. baumannii, respectively. Moreover, they can also act as antibiotic carriers, restoring the sensitivity of pathogens to penicillin G. In vivo studies demonstrated their therapeutic efficacy both as monotherapies and in combination therapies with antibiotics. These results demonstrate PolyCLOVER as a powerful framework for discovering multi-functional SANPs to combat antibiotic resistance.

microbiology↗

bioRxiv2025

Phosphorylation of the α subunit inhibits proteasome assembly and regulates cell division in archaea

Archaea of the order Sulfolobales possess a eukaryotic-like cell division machinery and display a eukaryotic-like cell cycle, however, the cell division and cell cycle control mechanisms remain enigmatic. Here, we demonstrate that phosphorylation of the subunit by a eukaryotic-like protein kinase ePK2 affects 26S proteasome assembly and controls cell division in Saccharolobus islandicus. ePK2 exhibits cell cycle-dependent expression at both transcriptional and translational levels. Deletion or overexpression of ePK2 results in impaired cytokinesis, with the deletion cells being unable to generate single chromosome cells after synchronization and the overexpression cells exhibiting growth retardation and cell enlargement. Interestingly, overexpression of ePK2 leads to a coherent reduction in cellular proteasome activity and degradation of cell division proteins. We identify S200 and T213 of the proteasome subunit as specific target sites for ePK2 phosphorylation. Functional and structural analyses of site-directed mutants at S200 and T213 suggest that S200 phosphorylation disrupts the assembly of 20S into 26S proteasome whereas T213 phosphorylation interferes the de novo ring assembly. Collectively, our study uncovers an ingenious and efficient mechanism of proteasome phosphorylation-mediated cell division regulation, a prototype of the eukaryotic cell cycle regulation system, in Sulfolobales archaea.

microbiology↗

bioRxiv2025

Vessels hiding in plain sight: quantifying brain vascular morphology in anatomical MR images using deep learning

Non-invasive assessment of brain blood vessels with magnetic resonance (MR) imaging provides important information about brain health and aging. Time-of-flight MR angiography (TOF-MRA) in particular is commonly used to assess the morphology of blood vessels, but acquisition of MRA is time-consuming and is not as commonly employed in research studies or in the clinic as the more standard T1- or T2-weighted MR contrasts (T1w/T2w). To enable quantification of brain blood vessel morphology in T1w/T2w images, we trained a neural network model, anat2vessels, on a dataset with paired MR/MRA. The model provides accurate segmentations as assessed in cross-validation on ground truth images, particularly in cases where T2w images are used. In addition, correlation between features that are extracted from model-based vessel segmentations and from ground truth account for as much as 78% of the variance in these features. We further evaluated the model in another dataset that does not include MRA and found that anat2vessels-based vessel morphology features contain information about aging that is not captured by cortical thickness features that are routinely extracted from T1w/T2w images. Moreover, we found that vessel morphology features are associated with individual variability in blood pressure and cognitive abilities. Taken together these results suggest that anat2vessels could be fruitfully applied to a range of existing and new datasets to assess the role of brain blood vessels in aging and brain health. The methods are provided as open-source software in https://github.com/nrdg/anat2vessels/.

neuroscience↗

bioRxiv2025

GWAs reveals SUBER GENE1-mediated suberization via Type One Phosphatases

Suberin deposition in the root endodermis is critical for plant nutrient acquisition and environmental adaptation. Here, we used an unbiased forward genetic approach based on natural variation across 284 Arabidopsis thaliana accessions to identify novel regulators of endodermal suberization. This screen revealed striking diversity in suberin levels and patterns, uncovering broader roles for suberin beyond those observed in the reference accession Col-0. A genome-wide association study pinpointed SUBER GENE1 (SBG1), a previously uncharacterized gene encoding a 129-amino acid protein, as a key regulator of suberin deposition. SBG1 acts through physical interaction with type one protein phosphatases (TOPPs) via conserved SILK and RVxF motifs. Disrupting this interaction abolishes SBG1 function, while topp mutants exhibit enhanced endodermal suberization, mirroring SBG1 overexpression. Our findings uncover a previously unknown regulatory module linking suberin formation to TOPP activity and ABA signaling and provide a framework for improving plant stress resilience through targeted manipulation of root barrier properties.

plant biology↗

bioRxiv2025

Computational Assessment of Circularization Impact on Saxitoxin G-quadruplex Aptamer Performance

Although high-affinity and high-specificity aptamers can be obtained by SELEX (Systematic Evolution of Ligands by Exponential Enrichment), and have good activity in vitro, their performance in vivo is poor. This is mainly because of the degradation of linear nucleic acids by nucleases in body fluids and the difficulty of maintaining stable 3D structures in complex body fluid environments, which seriously hinders the clinical application of aptamers. This makes circular aptamers, which can resist degradation and are more stable in complex body fluid environments, an attractive option. However, the cyclization process may have a considerable impact on the structure of the original linear aptamer, and blind cyclization may lead to a decrease in its affinity. Here, we developed a new method to guide the cyclization of saxitoxin (STX) aptamers based on molecular computation, and verified the activity of circular aptamers through bioactivity experiments to further verify the authenticity and reliability of our methods. Consistent with the computational analysis, although circularization disrupts the G4 structure, circular aptamer not only has a higher affinity than the linear one but also provides certain protection to mice and greatly prolongs their survival, which proves the accuracy and reliability of the circular structure designed by our computational method. Our findings provide important ideas for better guidance in the design of circular aptamers and demonstrate their great potential in therapeutic applications.

biochemistry↗

bioRxiv2025

Locus Coeruleus to Medial Prefrontal Cortex Noradrenergic Neural Circuit Modulates States of Consciousness during Sevoflurane Anesthesia in Mice

How exactly does general anesthesia achieve its effects? It is so widely used in surgeries and medical treatments to enhance comfort for nearly one hundred years, yet its precise mechanisms underlying the loss and recovery of consciousness still remain unclear. Utilizing a variety of research methods-pharmacological, optogenetic, chemogenetic, fiber photometry, and gene knockdown approaches-our study has shed light on the significant modulatory function of locus coeruleus noradrenergic neurons during the transition from sevoflurane anesthesia to awakening. Furthermore, the activation of the LC-mPFC circuit has been found to have a substantial arousal effect, with 1-AR playing a crucial role in this process. Additionally, GABAA-R has been identified as the key binding site for sevoflurane within the locus coeruleus. These findings collectively offer novel insights into the neural network mechanisms underlying general anesthesia, advancing our understanding of this complex medical phenomenon.

neuroscience↗

bioRxiv2025

Biophysical and molecular mechanisms responsible for phytoplankton sinking in response to starvation

Marine phytoplankton face eco-evolutionary pressure to regulate their vertical position in the ocean to access light, which is abundant towards the surface, and nutrients, which are found deeper down the water column. All phytoplankton experience gravitational sinking, which can contribute to their vertical migration. However, the biophysical and molecular mechanisms that impact gravitational sinking have not been systematically characterized across taxa and environmental conditions. Here, we combine simulations with measurements of cell mass, volume, and composition to investigate the effects of nutrient availability on gravitational sinking in 9 representative unicellular pico- and nanoplankton species. We find that gravitational sinking becomes faster in most species when starved, but the biophysical changes responsible for this vary across species and starvation conditions. For example, the faster sinking of Chaetoceros calcitrans is nearly exclusively driven by cell density whereas that of Emiliania huxleyi is due to cell volume. On the molecular level, the altered sinking is predominantly attributed to changes in cellular dry contents, rather than water. For example, starch accumulation increases sinking in 3 green algae species, and lipid accumulation decreases sinking in Phaeodactylum tricornutum. Overall, our work reveals that phytoplankton physiology has evolved multiple mechanisms that impact gravitational sinking in response to starvation, possibly to support the vertical migration of the cell.

cell biology↗