Biology subjects

Lin, Z.

Publications and source records attributed to Lin, Z..

At least 55 records · Page 3Linked to original sources

bioRxiv2025

Cryogenic 3D electron microscopy captures whole plant cell native ultrastructure

Volume electron microscopy (vEM) provides nanometer-scale, three-dimensional imaging of cells, but applying it to plant systems remains challenging. Cell walls, large vacuoles, and tissue thickness complicate sample preparation and cryogenic imaging. Here we report a cryogenic vEM (cryo-vEM) workflow for unstained plant protoplasts that achieves volumetric imaging of whole vitrified sorghum stem protoplasts without chemical fixation, dehydration, resin embedding, or heavy-metal staining. The method integrates optimized protoplast isolation, plunge-freezing vitrification for native-state preservation, automated cryogenic focus ion beam scanning electron microscopy (cryo-FIB-SEM) slice-and-view acquisition, contrast enhancement and stack alignment, and AI-assisted human-in-the-loop 3D segmentation. Using sorghum stem protoplasts as a demonstration, the workflow captured large-volume frozen-hydrated protoplast ultrastructure, allowing visualization of major organelles, including the nucleus, mitochondria, vacuoles, ER/Golgi-like membranes, lipid bodies, and subcellular features consistent with nuclear-envelope pores. We further quantified organelle volumes and surface areas from the segmented 3D data, highlighting the potential for quantitative cellular ultrastructure analysis. This cryo-vEM workflow provides a platform for near-native structural studies of isolated plant protoplasts.

plant biology↗

bioRxiv2025

3' to 5' Translation of Circular RNAs?

For decades, the 5[->]3 direction of translation has been a central dogma of molecular biology(1). Here we present evidence that eukaryotic circular RNAs (circRNAs) can serve as templates for 3[->]5 backward translation (BT), yielding polypeptides with distinct sequence and structural features not found in canonical proteomes. Mining of more than 6,000 multi-omics datasets identified [~]1 million candidate BT proteins across eukaryotes, including 59,000 high-confidence human BT proteins supported by mass spectrometry. Genetic combinatorial experiments (KO and KI) and cell-free translation of synthetic circRNAs establish BT as a conserved mechanism. Loss-of-function studies of hs.circCAPN15 in human and dm.circROLS in Drosophila underscore the functional importance of BT-derived proteins. Our work challenges the long-standing unidirectional translation paradigm, expands the functional landscape of the genetic code, and reveals a hitherto hidden layer of proteomic complexity with broad implications for biology and therapeutics.

molecular biology↗

bioRxiv2025

RNA editing fine-tunes transcriptional plasticity and enhances the adaptability of Myzus persicae to diverse host plants

Parthenogenetic organisms can adapt to diverse and rapidly changing environments despite limited standing genetic variation, but the molecular mechanisms enabling such adaptability remain unclear. Here, we investigate this question in the green peach aphid (Myzus persicae, GPA), an asexual generalist herbivore capable of colonizing a wide range of host plants without underlying genetic change. By integrating RNA-seq and DNA-seq from genetically identical clones reared on different hosts, we identified 1,368 high-confidence RNA editing sites (RES), many of which exhibited strong host-dependent differences in editing levels. These RES encompassed all 12 possible nucleotide substitutions, with A-to-I, C-to-U, and U-to-C conversions being particularly prevalent. Editing at several sites shifted dynamically within 48 hours of host transfer and often coincided with changes in transcript abundance. Functional knockdown of the editing genes ADAR1 and ADAR2 reduced editing at specific sites, diminished fecundity, and abolished aphid survival on new hosts. Comparative analyses across aphid species reveal that although the biochemical signatures of A-to-I editing are deeply conserved, the genomic locations and functional roles of edited transcripts are species specific, reflecting lineage-specific adaptive strategies. Together, our findings demonstrate that RNA editing provides a rapid, reversible mechanism for environmental acclimation in GPA, enabling adaptive phenotypic plasticity even in the absence of genetic variation. This work establishes RNA editing as a key molecular basis for the remarkable ecological generalism of clonal organisms. Significance statementA major question in evolutionary biology is how clonal organisms with little genetic variation rapidly adapt to diverse, changing environments. Here we show that RNA editing provides a mechanism enabling the parthenogenetic green peach aphid (Myzus persicae) to colonize a wide range of host plants. RNA editing generates transcriptomic diversity, responds quickly to host shifts, and is essential for adaptation, as silencing ADAR genes disrupts A-to-I editing, reduces fitness, and prevents survival on new hosts. Comparative analyses across aphid species reveal that although the biochemical features of A-to-I editing are conserved, its genomic deployment and functional outcomes differ among species, reflecting lineage-specific strategies. These findings demonstrate that RNA editing creates a dynamic regulatory layer that offsets constraints of asexual reproduction.

molecular biology↗

bioRxiv2025

ClsDiff-AMP30: Generating Antimicrobial Peptides by a Classifier Guidance Noise Predictor

Antimicrobial peptides (AMPs) represent a promising therapeutic strategy to combat the increasing challenge of multidrug-resistant pathogens, a crisis intensified by the overuse of conventional antibiotics. In addition to their broad-spectrum antimicrobial activity, low toxicity, and reduced propensity for resistance development, AMPs offer significant advantages over traditional antibiotic therapies. However, the discovery of novel AMPs through biological experiments remains constrained by high costs, labor-intensive workflows, and time-consuming procedures, underscoring the urgent need for in silico computational methods to design AMP sequences. Notably, shorter AMPs ([≤] 30 residues) demonstrate superior antimicrobial efficacy, improved structural stability, and minimal cytotoxicity toward human cells. To address these challenges, we present a classifier-guided diffusion framework specialized for generating AMPs shorter than 30 residues (ClsDiff-AMP30). The architecture integrates two interdependent submodels, including a noisy AMP classifier that evaluates AMP likelihood at intermediate denoising steps and a noise predictor guided by classifier-derived probability scores, dynamically adjusted via a self-optimized coefficient to modulate guidance strength. ClsDiff-AMP30 achieves a validation accuracy of 66% across 10,000 synthesized sequences by a self-developed AMP classifier. Furthermore, wet lab experiments demonstrated that all 11 selected sequences exhibited high antimicrobial activity against at least one of the three tested bacterial strains and low hemolytic activity.

bioinformatics↗

bioRxiv2025

The functional divergence of two ethylene receptor subfamilies that exhibit Ca2+-permeable channel activity

Ethylene is a gaseous plant hormone crucial for regulating plant growth, development, and stress adaptation, yet the molecular basis underlying ethylene receptor function remains elusive. Here, we show that subfamily I receptors constitute core ethylene-sensing module and function epistatically to subfamily II receptors. Notably, we discover that only subfamily I receptors possess Ca2+-permeable channel activity, which are indispensable for ethylene-induced cytosolic calcium influx. Overall, this work supports a mechanistic framework in which subfamily I receptors integrate ethylene sensing with Ca2+ influx, providing new insight into how plants translate hormonal cues into downstream signaling events.

plant biology↗

bioRxiv2025

RoBep: A Region-Oriented Deep Learning Model for B-Cell Epitope Prediction

MotivationAccurate in silico identification of B-cell epitope residues is crucial for antibody design and structure-guided vaccine development. Although recent protein language models and structure-aware methods can capture spatial information of tertiary structure when generating residue embeddings, most existing epitope predictors use these embeddings to perform classification for individual residues one by one, without enforcing spatial continuity for reported epitope residues. Such methods often result in biologically implausible predictions because B-cell epitope residues always cluster together on the antigen surface. ResultsWe present RoBep, a region-oriented B-cell epitope predictor that explicitly models the spatial clustering of epitope residues. RoBep introduces a novel region constraint mechanism and combines the advanced protein language model ESM-Cambrian with an equivariant graph neural network. Our method outperforms existing structure-based methods on the benchmark dataset, demonstrating improvements of 26%, 45%, 13%, and 43% in F1, MCC, AUPR, and AUROC0.1, respectively. In addition to residue-level predictions, RoBep can also provide antibody-antigen binding regions. Importantly, the predicted epitope residues are ensured to be spatially compact, enhancing biological plausibility and practical relevance for immunotherapeutic design. AvailabilityA user-friendly website for using RoBep is provided at https://huggingface.co/spaces/NielTT/RoBep. All datasets, source code used in this work, and implementation instructions of the website are publicly available at https://github.com/YitaoXU/RoBep.

bioinformatics↗

bioRxiv2025

Agentic Lab: An Agentic-physical AI system for cell and organoid experimentation and manufacturing

Reproducibility in biological research and manufacturing remains constrained by the complexity of multi-step protocols, fragmented data-analysis pipelines, and the intrinsic variability of experimental execution. Here, we present Agentic Lab, an agentic-physical AI platform that unifies large language model and vision language model (LLMs/VLMs)-driven reasoning with real-world laboratory operations. Agentic Lab uses multi-agent orchestration architecture, comprising of specialized subagents for knowledge retrieval, protocol design, multimodal data analysis, and training-free segmentation and representation learning for intrinsically explainable single-cell and organoid phenotyping. These agents operate under the orchestration of a virtual principal investigator MolAgent that is linked to an augmented reality (AR)-based physical AI interface, which can bridge digital reasoning with human physical execution. Agentic Lab perceives real-world experimental activities, provides context-aware instructions, identifying procedural errors in real time for humans to correct, and continuously evolves with its long-term memory database expanding through the accumulation of experimental data logs from human scientists. This interaction allows scientists and AI agents to collaborate and co-evolve dynamically, closing the loop between planning, action, and analysis in the traditional cell and organoid research lifecycle. We demonstrate Agentic Lab in organoid differentiation from human pluripotent stem cells, where it autonomously generates protocols, monitors culture procedures, and identifies subtle morphological heterogeneity linked to growth conditions. The system interprets these phenotypes, grounds them in literature, and proposes targeted instructions for improving differentiation efficiency. By combining multi-agent reasoning with physical laboratory awareness, Agentic Lab transforms experimentation and biomanufacturing from a static workflow into an adaptive, feedback-driven, bidirectional process that integrates agentic AI into the research lifecycle. This framework establishes a foundation for intelligent laboratories that integrate design, execution, and interpretation within a unified agentic-physical system.

bioengineering↗

bioRxiv2025

Mapping unsolved lipidomes accelerates lipid discovery in major bacterial pathogens

Unlike gene-first approaches to understanding bacterial pathogenesis, molecule-forward discovery can uncover unexpected chemical diversity. Here, new lipidomic analytical methods and quality metrics defined the large scope of unknown lipids in the worlds deadliest pathogen, Mycobacterium tuberculosis (Mtb). This map allowed rapid discovery of Mtb lysyldiacylglycerol linked to the biosynthetic gene lysX, which controls in vivo infection outcomes in moth larvae, mice, guinea pigs, and here, zebrafish. A broader search for orthologous lysyltransferase domains identified the Staphylococcus aureus virulence gene mprF, where the same lipoamino acid was shown to be a previously unknown biosynthetic product. Thus, lipidomic mapping showed that the cell envelope composition of well-studied bacterial pathogens remains substantially unsolved and offers a new way to generate lists of discoverable lipids to accelerate molecular discovery.

microbiology↗

bioRxiv2025

Mechanistic Insights into Cyclic Penta-Adenylate-Mediated Activation of Type III CRISPR Ribonuclease Csm6

Type III CRISPR systems generate cyclic oligoadenylate (cOA, 3 to 6 AMPs) messengers upon detecting viral RNA, activating downstream effectors to defend against viral infection. Although cOA-activated effectors have been extensively characterized, the cA5-specific effectors remained unexplored despite cA5 being among the most abundant cOA species produced during phage infection. Here, we report that Actinomyces procaprae Csm6 (ApCsm6) selectively employs cA5 as its activator. Unlike other characterized Csm6 proteins, ApCsm6 self-limits its ribonuclease activity by degrading cOAs via its HEPN domain, rather than relaying on the CARF domain. Cryo-EM structures of ApCsm6 and its complexes with cA5 and cA6 reveal a homotetrameric assembly, where each monomer binds a single cOA within a composite pocket formed by two tandem CARF-HEPN domains. Binding of cA5, but not cA6, enhances tetramerization and induces large conformational shifts in CARF, which in turn allosterically activates ssRNA cleavage in HEPN. These findings advance our understanding of ligand discrimination and signaling regulation in type III CRISPR immunity. HighlightsO_LIApCsm6 preferentially recognizes cA5 as its activator. C_LIO_LICryo-EM structures of ApCsm6 and its complexes with cA5 or cA6 reveal the structural basis for cA5-selective recognition and allosteric activation. C_LIO_LIApCsm6 acts as a self-limiting ribonuclease by degrading cOAs via its HEPN domain rather than relaying on the CARF domain. C_LI

biochemistry↗

bioRxiv2025

DeviceAgent: An autonomous multimodal AI agent for flexible bioelectronics

The development of flexible bioelectronics remains a complex, multidisciplinary process that demands specialized expertise and labor-intensive efforts, limiting scalability, adaptability and accessibility. Here, we introduce DeviceAgent, an autonomous multimodal AI agent that integrates large language models (LLMs), vision-language models (VLMs), and domain-specific computational tools into a unified framework for bioelectronics research. Leveraging the emergent reasoning abilities of LLMs and VLMs, DeviceAgent enables zero and few-shot generalization, contextual learning, and flexible task execution across modalities. A multimodal context memory system orchestrates these capabilities, providing end-to-end support across the experimental pipeline-from high-level design objectives to fabrication protocol generation, visual defect inspection, and electrophysiological signal analysis, while maintaining human oversight at critical decision points. We demonstrate its capabilities through the development of stretchable mesh electronics for interfacing with human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs), a representative application involving complex device architectures, heterogeneous material nanofabrication, and electrophysiology analysis. DeviceAgent autonomously (1) generates customized bioelectronic layouts; (2) creates comprehensive fabrication protocols tailored to specific materials and processes; (3) identifies microscopic defects using visual reasoning; and (4) analyzes cardiac electrophysiological recordings in an interpretable manner. By embedding LLMs and VLMs within a structured, tool-augmented architecture, DeviceAgent establishes a scalable and accessible paradigm for AI-scientist collaboration in nanofabrication and bioelectronics research.

bioengineering↗

bioRxiv2025

Probing the genomic and proteomic basis of encystment in Oxytricha granulifera

Protozoan encystment is a crucial survival strategy for resisting environmental stresses, yet its molecular mechanisms remain poorly understood due to limited genomic resources and lack of integrated multi-omics analyses in model ciliates. In this study, we aimed to elucidate the molecular mechanisms underlying the encystment process in Oxytricha granulifera by sequencing and assembled its genome, and by conducting comprehensive transcriptome, proteome and morphology analyses to reveal gene expression and morphologic changes between vegetative and cyst stages. Morphological observation revealed ciliary dedifferentiation and cyst-wall formation during encystment, events respectively supported by the downregulation of microtubule dynamics related genes and the upregulation of vesicle transport related genes in cyst stage. Expanded gene families for carbohydrate metabolism, autophagy and cellular acidification align with the species mucocyst signature and their observed autophagic clearance, hinting at previously unrecognized mechanisms underlying cyst formation. Elevated expression of the ubiquitin-proteasome system and autophagy pathways facilitates protein turnover, and upregulation of antioxidant enzymes genes helps mitigate oxidative damage, while a rewired post-transcriptional regulation that increases spliceosome activity and alternative splicing frequency--each trend validated at the protein level. Concurrently, the methyltransferases responsible for DNA N6-adenine methylation (6mA), with homologous genes of AMT1 and AMT6/7 significantly downregulated. These results suggesting a multilayered regulation of ciliate encystment and the first integrated evidence that alternative splicing modulates dormancy. Our findings establish the first multi-omics framework for O. granulifera encystment, offering baseline data and tentative clues to the dormancy mechanism that will inform future inquiries into how single-celled eukaryotes endure adverse environments. IMPORTANCEOxytricha species are widely distributed in freshwater and terrestrial ecosystems, playing significant ecological roles in microbial communities. Their ability to undergo encystment provides a powerful model for studying cellular differentiation and stress adaptation in microbial eukaryotes. This study presents the first multi-omics analysis of encystment in Oxytricha granulifera, revealing microbial survival strategies through enhanced protein turnover, autophagy, alternative splicing, and DNA methylation reprogramming. These findings offer fundamental insights into dormancy mechanisms and environmental adaptation in protists, advancing our understanding of microbial resilience, evolutionary innovation, and ecological success in fluctuating environments.

genomics↗

bioRxiv2025

Determination of α-Synuclein Protein Interactions by μMap Photo-proximity Labeling

Fibrillar aggregates of the natively disordered protein -synuclein (S) are hallmarks of Parkinsons disease and related neurodegenerative disorders termed synucleinopathies. Here, we used micromap ({micro}Map) photo-proximity labeling to determine the interactomes of S monomers and fibrils in mouse brain lysate to better understand both the loss of healthy function and gain of toxic function aspects of synucleinopathies. Several S variants were synthesized and characterized, showing that the small size (1 kDa) of the Ir catalyst attached through a Cys-maleimide linkage makes it minimally-perturbing to S, with a narrow labeling radius that allows one to identify interactome differences between different regions of S. Monomer and fibril interactomes were compared to each other and to previous proximity labeling data sets for validation and several examples of further investigations are demonstrated, including Western blotting, super-resolution microscopy, and {micro}Map in primary neurons.

biochemistry↗

bioRxiv2025

An AI Agent for cell-type specific brain computer interfaces

Decoding how specific neuronal subtypes contribute to brain function requires linking extracellular electrophysiological features to underlying molecular identities, yet reliable in vivo electrophysiological signal classification remains a major challenge for neuroscience and clinical brain-computer interfaces (BCI). Here, we show that pretrained, general-purpose vision-language models (VLMs) can be repurposed as few-shot learners to classify neuronal cell types directly from electrophysiological features, without task-specific fine-tuning. Validated against optogenetically tagged datasets, this approach enables robust and generalizable subtype inference with minimal supervision. Building on this capability, we developed the BCI AI Agent (BCI-Agent), an autonomous AI framework that integrates vision-based cell-type inference, stable neuron tracking, and automated molecular atlas validation with real-time literature synthesis. BCI-Agent addresses three critical challenges for in vivo electrophysiology: (1) accurate, training-free cell-type classification; (2) automated cross-validation of predictions using molecular atlas references and peer-reviewed literature; and (3) embedding molecular identities within stable, low-dimensional neural manifolds for dynamic decoding. In rodent motor-learning tasks, BCI-Agent revealed stable, cell-type-specific neural trajectories across time that uncover previously inaccessible dimensions of neural computation. Additionally, when applied to human Neuropixels recordings-where direct ground-truth labeling is inherently unavailable-BCI-Agent inferred neuronal subtypes and validated them through integration with human single-cell atlases and literature. By enabling scalable, cell-type-specific inference of in vivo electrophysiology, BCI-Agent provides a new approach for dissecting the contributions of distinct neuronal populations to brain function and dysfunction.

neuroscience↗

bioRxiv2025

Algebraic and Generative Design of DNA Origami via Origami Monoids

AO_SCPLOWBSTRACTC_SCPLOWWe integrate the elegance of monoid structures with the power of deep learning models, specifically Sequential variational autoencoders (SeqVAEs), to explore the vast and stability-dependent structural space of DNA origami. While existing computational tools provide well-established protocols for diverse applications, they lack large-scale datasets that could fully leverage modern computational power. To address this gap, we introduce an encoding framework based on the algebraic properties of Jones monoids. This approach enables both a top-down construction of DNA origami designs and a systematic understanding of how crossover arrangements influence overall shape. This work thus establishes a new framework for the design and analysis of DNA origami, bridging algebraic formalism with generative modeling.

bioinformatics↗

bioRxiv2025

Developmental Connectomics of the Mouse Cerebellum

To uncover the developmental processes that establish the precise patterns of synaptic connectivity in the CNS, we employed a connectomic approach in the mouse cerebellar cortex between birth and 2 weeks of age. There were dramatic quantitative and qualitative changes in the structure and connectivity of cerebellar cells. Parallel fiber synapses onto Purkinje cells increased [~]500-fold, with the most rapid growth taking place a week after birth. To support this profound synaptogenesis, Purkinje cells generated thousands of transient parallel fiber-oriented filopodia that received nascent synapses from parallel fibers. Importantly, we find that granule cells initiate synaptic output onto Purkinje cells only after receiving mossy fiber input, revealing a sequential, input-dependent logic for circuit assembly. In sharp contrast to the concurrent pruning of climbing fiber inputs, parallel fiber connectivity expanded and became highly individualized during development. Despite anatomical overlap, neighboring Purkinje cells share significantly fewer parallel fiber inputs than expected by chance. Moreover, parallel fibers themselves diverged spatially, further enforcing selective input allocation and resulting in highly specific parallel fiber cohorts for each Purkinje cell. Our findings uncover a mechanistic sequence in which early afferent activity and transient cellular structures guide the selective wiring and expansion of parallel fiber input to Purkinje cells, establishing developmental principles that ensure functional specificity in the mature brain.

neuroscience↗

bioRxiv2025

Artificial Intelligence-driven Whole-brain Cell Mapping with Highly Multiplexed In Situ Hybridization

Recent advances in three-dimensional single-cell-resolution imaging have begun to link organ-wide and cellular level research in development and disease research. Harnessing the power of whole-mount cell staining and tissue-clearing, it became possible to quantify the cell populations throughout an intact organ. While powerful, whole-organ imaging remains limited by the inability to stain a broad range of molecular markers simultaneously and by the lack of an analytical scheme to precisely quantify the cell population. Here, we present a highly multiplexed whole-mount staining technique, utilizing the repeated application of fluorescent in situ hybridization. This technique, termed mFISH3D, was designed by extensively dissecting the chemical basis of hybridization reactions in fixed tissue. mFISH3D enabled the visualization of 10 types of mRNAs in an intact mouse brain and has been demonstrated in various biological specimens including the human brain. To achieve unprecedented levels of accuracy in spatial cell mapping, we developed artificial intelligence (AI)-driven workflow using self-supervised learning, significantly reducing the need for extensive manual annotations. The integration of mFISH3D with our AI solution sets a standard for high-dimensional tissue analysis, provides a new systematic framework for analyzing complex cellular ecosystems and enables comprehensive investigation of selective cellular vulnerabilities in diseases.

neuroscience↗

bioRxiv2025

SynAnno: Interactive Guided Proofreading of Synaptic Annotations

Connectomics, a subfield of neuroscience, aims to map and analyze synapse-level wiring diagrams of the nervous system. While recent advances in deep learning have accelerated automated neuron and synapse segmentation, reconstructing accurate connectomes still demands extensive human proofreading to correct segmentation errors. We present SynAnno, an interactive tool designed to streamline and enhance the proofreading of synaptic annotations in large-scale connectomics datasets. SynAnno integrates into existing neuroscience workflows by enabling guided, neuron-centric proofreading. To address the challenges posed by the complex spatial branching of neurons, it introduces a structured workflow with an optimized traversal path and a 3D mini-map for tracking progress. In addition, SynAnno incorporates fine-tuned machine learning models to assist with error detection and correction, reducing the manual burden and increasing proofreading efficiency. We evaluate SynAnno through a user and case study involving seven neuroscience experts. Results show that SynAnno significantly accelerates synapse proofreading while reducing cognitive load and annotation errors through structured guidance and visualization support. The source code and interactive demo are available at: https://github.com/PytorchConnectomics/SynAnno.

neuroscience↗

bioRxiv2025

Interpretable spatial multi-omics data integration and dimension reduction with SpaMV

Spatial multi-omics technologies have revolutionized our understanding of biological systems by providing spatially resolved molecular profiles from multiple perspectives. Existing spatial multi-omics integration methods often assume that data from different modalities share a common underlying distribution, aiming to project them into a single unified latent space. This assumption, however, can obscure the unique insights offered by each modality, thereby limiting the full potential of multi-omics analyses. To address this limitation, we present the Spatial Multi-View (SpaMV) representation learning algorithm, which captures both the shared information across modalities and the distinct, modality-specific information, enabling a more comprehensive and interpretable representation of spatial multi-omics data. Through extensive evaluation on both simulated and real-world datasets, SpaMV demonstrates superior spatial domain clustering performance and provides users with more interpretable dimension reduction for downstream analysis. Moreover, SpaMV effectively annotates cell types within clusters of a mouse thymus dataset, highlighting its effectiveness in interpretable dimensionality reduction.

bioinformatics↗