Search bioRxiv⌕ Search

Biology subjects

Tian, T.

Publications and source records attributed to Tian, T..

At least 19 recordsLinked to original sources
bioRxiv2026

Ileal Gpbar1 gates depressive phenotypes via gut to brain "vagal-NTS-PVN"GABAergic axis

Depression represents a typical mental condition, yet whether intestinal metabolic signals drive central neural circuits remains unclear. Here we identify ileal enteric neuronal Gpbar1 as a previously unrecognized peripheral node that coordinates gut-to-brain communication to prevent depression-like behavior through a vagus-dependent, nucleus tractus solitaries (NTS)-paraventricular nucleus (PVN) signaling axis. The enteric plexus Gpbar1 is predominantly expressed on a subpopulation of GAD1-positive GABAergic neurons. Chronic restraint stress suppresses Gpbar1 expression, triggers glycolytic reprogramming and mitochondrial distress in both ileum and hypothalamus. Oral Gpbar1 agonism with INT-777 reverses hypothalamic HK2/TOM20 alterations and behavioral abnormalities by enhanced ileal mTOR/HIF-1 alpah signaling and glycolysis-related metabolite shifts, an effect specifically abolished by subdiaphragmatic vagotomy or chemogenetic silencing of PVN GABAergic neurons. Our findings establish ileal Gpbar1 as a gut-derived metabolic sensor that engages a defined vagal-brainstem-hypothalamic inhibitory circuit to regulate depressive phenotypes, shifting the roadmap from brain-centered to gut-initiated antidepressant mechanisms and unveiling a compelling peripheral therapeutic strategy that bypasses intractable central brain-directed intervention.

neuroscience↗
bioRxiv2026

Tirzepatide attenuates atherosclerosis through weight loss-independent anti-inflammatory mechanisms

BackgroundAtherosclerosis is a chronic inflammatory vascular disorder with persistent residual inflammation even after standard lipid-lowering therapy. Mounting evidence from bench to bedside suggests that diabetes and obesity accelerate atherosclerosis development. Tirzepatide (TZP), a dual Glucagon-Like Peptide-1 Receptor/Glucose-Dependent Insulinotropic Polypeptide Receptor (GLP-1R/GIPR) agonist approved for treating diabetes and obesity, has demonstrated proven cardiometabolic efficacy in large cardiovascular outcome trials. However, it remains largely uncertain whether TZP attenuates atherosclerosis independent of its anti-diabetic and anti-obese effects through direct actions on the vasculature. MethodsWe established atherosclerotic mouse models under diabetic, obese, and non-diabetic/non-obese conditions. Analysis of covariance (ANCOVA) and pair-feeding experiments were applied to experimentally decouple weight-dependent metabolic improvement from intrinsic vasculoprotection. Molecular and cell biological assays in human umbilical vein endothelial cells (HUVECs) and human aortic endothelial cells (HAECs) were performed to dissect the underlying signaling mechanisms. ResultsTZP markedly reduced aortic plaque burden and inflammation, restrained necrotic core enlargement, and improved plaque stability across all experimental mouse models. Both ANCOVA and pair-feeding experiments confirmed that these atheroprotective effects were independent of food intake and body weight loss. Furthermore, TZP attenuated systemic and vascular inflammation in Tumor Necrosis Factor- (TNF)-treated C57BL/6J mice, and this protection occurred without changes in body weight or blood glucose levels. Mechanistically, TZP directly targeted endothelial cells and activated the cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA)/endothelial nitric oxide synthase (eNOS) pathway, increased eNOS phosphorylation and nitric oxide bioavailability, consequently downregulating the expression of the pro-inflammatory adhesion molecules Vascular Cell Adhesion Molecule-1 (VCAM-1) and Intercellular Adhesion Molecule-1 (ICAM-1). ConclusionsTZP arrests atherosclerosis progression through weight loss-independent anti-inflammatory mechanisms. These findings implicate TZP as a promising therapeutic drug for mitigating residual vascular inflammation in patients with atherosclerotic cardiovascular disease (ASCVD), irrespective of glycemic status or obesity. Clinical PerspectiveO_ST_ABSWhat Is New?C_ST_ABSO_LITirzepatide exerts direct anti-atherosclerotic effects in preclinical mouse models of atherosclerosis under diabetic, obese, and non-obese conditions. C_LIO_LITirzepatide directly targets endothelial GLP-1R/GIPR and downstream cAMP/PKA/eNOS signaling pathway to suppress NF-{kappa}B-driven vascular inflammation, thereby uncovering a previously unrecognized vasculoprotective mechanism underlying its cardiovascular benefits C_LI What Are the Clinical Implications?O_LITirzepatide exerts direct vascular protective effects independent of body weight reduction, suggesting that its cardiovascular benefits may extend beyond glycemic control and obesity management. C_LIO_LITirzepatide may represent a promising therapeutic drug for addressing residual vascular inflammation in ASCVD patients, including those without overt diabetes or obesity C_LI

physiology↗
bioRxiv2026

Modelling Plant Cortical Mictrotubules: Curvature Sensing from Bending Energy

Within plant cells, the self-organization of cortical microtubules (MTs) into ordered arrays is an important process for directional growth. There is growing evidence that cortical MTs respond to cell shape and/or mechanical stresses in the cell wall, requiring in silico models on complicated surfaces to provide a complete understanding. Most models assume that MTs are directionally persistent, following geodesics of the surface. This ignores the expected tendency of these elastic filaments to minimize curvature. Our recent model incorporated minimization of MT curvature in cylindrical cells and found curvature to be significant in biasing the array organization. Here, we generalize to a larger class of surfaces, studying individual microtubule shapes to provide insights into the role of geometric cues and highlight differences with previous models that use the geodesic assumption. We first show that geodesic models, including current models with finite persistence lengths, exhibit an invariance across certain geometries, leading to biophysically counterintuitive results. Incorporating curvature minimization, we show the difficulties imposed by high-curvature cell edges, elucidating potential new roles of proteins in helping microtubules traverse edges. Lastly, we show that geometries with competing curvature cues result in diverse curves previously not considered. These results provide geometric intuition for how various cell geometries affect individual cortical microtubules, helping us to better understand the processes required for the establishment of microtubule arrays in broad contexts such as: bundles spanning adjacent cell faces in prism-like root and leaf epidermis cells, protruding geometries of trichome cells, and rounded surfaces such as confined protoplasts.

Developmental Biology↗
bioRxiv2026

Mechanochemically-reprogrammed stem cell exosomes reconcile the biogenesis internalization trade-off for pulmonary fibrosis therapy

Matrix stiffness serves as a pivotal biophysical cue that profoundly dictates exosome biogenesis and cellular internalization, yet often creates a functional trade-off that impedes clinical translation. Herein, we developed a mechano-chemo-transductive strategy to engineer mesenchymal stem cell (MSC) exosomes endowed with robust biogenesis and superior delivery potency. Specifically, we revealed that MSCs cultured on soft matrices secreted a significantly elevated exosome yield and demonstrated enhanced competence to drive macrophage towards anti-inflammatory M2 polarization. Conversely, stiff matrices upregulated ATP-binding cassette transporter A1 (ABCA1) expression, enriching exosomal membrane cholesterol and facilitating cellular internalization by recipient cells. By taking advantages of these unique mechano-responses, we engineered MSCs via substrate softening combined with ABCA1 modulation to generate mechanochemically reprogrammed exosomes with concurrently enhanced yield and internalization efficiency. In a murine model of pulmonary fibrosis characterized by restrictive biological barriers, inhaled mechanochemically reprogrammed exosomes treatment demonstrated superior lung retention and deep tissue penetration. Furthermore, they effectively orchestrated immune homeostasis by repolarizing alveolar macrophages to reverse fibrotic remodeling and restore lung function. Collectively, by reconciling the intrinsic trade-off between biogenesis and cellular uptake, this strategy represents a paradigm shift in exosome engineering and paves the way for next-generation therapeutics against refractory fibrotic diseases.

bioengineering↗
bioRxiv2026

The Female Biomarker Challenge: Sex-Specific Network Robustness Constrains Biological Age Estimation and Geroscience Trial Design

Geroscience has more candidate interventions than it can afford to test. Trials powered for mortality or functional decline require large cohorts and long follow-up, and the persistent female survival advantage could further increase sample-size requirements if women and men age through different biology. Existing ageing biomarkers -- including frailty indices, biological-age clocks, allostatic load and homeostatic dysregulation -- compress multidimensional physiology to a scalar, discarding which combination of biomarkers changed. In 29,053 US adults measured on 38 routine laboratory and examination markers, we decomposed displacement from sex-specific young-healthy physiology into total magnitude and alignment with a cross-fitted mortality-associated direction. Women were no closer to young-health than men but had substantially lower disease mortality. Adjustment for magnitude widened the female survival advantage; adjustment for mortality-directed projection attenuated 87% of the sex difference on the log-hazard scale. Sex-specific directions transferred almost without loss, indicating a substantially shared predictive geometry with different positions along it. Across the population, a median 68% of the positive mortality-associated contribution came from measurements lying within conventional clinical limits and 57% from within the young-healthy range; these shares reflect how numerous such measurements are, not greater information per measurement. Ten routine measurements reproduced the 38-marker direction, and seven of eight predictions registered before analysis were confirmed in NHANES III. High-sensitivity C-reactive protein was strongly prognostic alone but showed no incremental information beyond the routine panel, whereas measuring lung-function added more held-out prognostic information than 33 of 34 individual blood and urine markers in the replication cohort. Retrospective sorting on this geometry produced mortality-enriched and physiologically displaced candidate trial groups; pooled thresholds were strongly sex-imbalanced, whereas within-sex thresholds restored balance with only a 1.5-3.8% reduction in expected five-year events. These findings suggest that inexpensive routine biomarkers plus simple functional testing could support smaller, more informative and sex-balanced geroscience trials, but prospective validation of treatment enrichment is required.

systems biology↗
bioRxiv2026

Heterogeneous and specific localization of mRNAs at dendritic spine and its dependence on microtubule entry and kinesin motor

Local mRNA translation near neuronal synapses is crucial for structural plasticity during brain development and memory formation. Whereas mRNAs are transported from soma to dendrites as individual ribonucleoprotein (RNP) granules, how do they subsequently localize at synapses is unclear. In particular, whether different dendritic mRNAs might display differential synaptic localization and interact with specific postsynaptic proteins remains unexplored. Here we found that two well-known dendritic mRNAs, Actb and Camk2a, are targeted to different subdomains of dendritic spines and delivered to different spine populations after synaptic stimulation. Surprisingly, the induced mRNA localization to dendritic spines is associated with microtubule (MT) entry and specifically requires the MT-associated motor KIF5A. Proximity-labelling proteomics further identifies distinct postsynaptic proteins enriched by Actb and Camk2a, from which we uncover the cytoskeletal protein Septin7 to be co-localized with the two mRNAs at different subdomains of dendritic spines. Our study reveals an unexpectedly specific mRNA localization mechanism at synapses and demonstrates proximity-labelling as a useful approach to map their synaptic partners.

neuroscience↗
bioRxiv2026

Sympathetic innervation regulates metabolic flexibility of skeletal muscle

The sympathetic nervous system (SNS) is recognized for its role in the physiological regulation of organs, such as heart, vasculature and lungs, and has emerged as a potential player in skeletal muscle metabolic and neuromuscular junction (NMJ) health. However, the mechanism through which SNS signaling influences skeletal muscle function and adaptation to exercise remains unclear. Using molecular, electrophysiological, immunohistochemical, and high-resolution respirometry techniques, we tested the role of sympathetic innervation to skeletal muscle in response to exercise. Our findings reveal that sympathetic denervation disrupts the NMJ, reducing motor and sympathetic receptor expression, with concomitant deficits in skeletal muscle function. Mechanistically, these deficits are linked to diminished CPT1 enzyme activity, which impairs long-chain fatty acid-mediated oxidation in skeletal muscle mitochondria. These findings reveal a key role for sympathetic innervation in maintaining mitochondrial metabolic function and by extension, skeletal muscle performance, offering novel insight into the interplay between the SNS, exercise, and muscle mitochondria.

cell biology↗
bioRxiv2026

Covalent Leader Peptide Probes Enable Proteome-Level Mapping of RiPP Enzymes beyond Biosynthetic Gene Clusters

Genome-based analyses have enabled widespread discovery and biosynthetic annotation of ribosomally synthesized and post-translationally modified peptides (RiPPs) pathways, yet often fail to capture all participating enzymes, particularly those encoded outside canonical biosynthetic gene clusters. In this study, leveraging the central role of leader peptides (LPs) in RiPP enzyme-substrate recognition, we developed a substrate-guided chemical-proteomic strategy using covalent LP probes to directly identify LP-binding enzymes from native proteomes. Using the LctA-LctM and PatE-LynD systems, we demonstrate that LP probes carrying a photocrosslinker engage enzymes at their authentic LP-recognition sites while preserving full catalytic competence. Application of this strategy to a model strain Streptomyces sparsogenes uncovered lanthipeptide synthetases encoded outside canonical biosynthetic gene clusters and revealed graded cross-cluster activities among multiple RiPP pathways. This work establishes a proof-of-concept model platform for proteome-level discovery of LP-binding RiPP enzymes and highlights cross-cluster enzymatic crosstalk as a latent mechanism for structural diversification.

biochemistry↗
bioRxiv2025

Ethylene signal-driven plant-multitrophic synergy boosts crop performance

Efficient nutrient use in agriculture depends on the dynamic interplay between plant roots, soil, and microbial communities. The root-rhizosphere interface is central to nutrient uptake and serves as a key hub for interactions with beneficial microbes. Arbuscular mycorrhizal (AM) fungi, positioned at the nexus between plant roots and soil microbiota, play a critical role in enhancing crop performance under nutrient-limited conditions. In this study, we dissected the genetic and molecular basis of AM fungi-induced lateral root development in maize (Zea mays), focusing on the role of ethylene-responsive transcription factors (ERFs). We identified ERF genes as essential regulators of pericycle cell division, acting downstream of ethylene biosynthesis genes (ACS6 and ACS7) and AM fungal signaling. Our findings reveal that AM fungi promote lateral root initiation by activating ERF expression and reprogramming flavonoid metabolism, particularly reducing the accumulation of flavonols such as kaempferol and quercetin, which otherwise inhibit root development when over-accumulated. Furthermore, we demonstrated that Massilia, a beneficial rhizobacterium, synergizes with AM fungi to enhance lateral root formation by colonizing fungal hyphae, degrading flavonoids, and contributing to auxin production. Together, our results uncover a tripartite signaling network linking ethylene signaling, flavonoid-mediated microbial recruitment, and AM symbiosis. This study highlights ERFs as central integrators of plant-microbe interactions and provides a molecular framework for engineering root architecture and microbiome assembly to improve nutrient acquisition and support sustainable crop productivity.

plant biology↗
bioRxiv2025

Biofouling on aquaculture mesh: disentangling seasonal and environmental effects with eDNA

Biofouling presents ecological and economic challenges for marine industries, yet patterns of community development in relevant environments remain underexplored. In this study, we examined seasonal (winter vs summer) and spatial (sheltered vs exposed) patterns in eukaryotic biofouling across six coastal sites in New Zealand. Using standardised mesh substrates and environmental DNA (eDNA) metabarcoding with the 18S rRNA gene, we characterised over 10,000 Amplicon Sequence Variants (ASVs) across 216 samples, revealing 541 species from 249 taxonomic classes. Exposed sites exhibited marked seasonal contrasts: winter communities were dominated by hydroids (e.g., Coryne eximia), while summer samples shifted to amphipod-rich assemblages (e.g., Jassa slatteryi). Sheltered sites showed more stable biomass and diversity but greater site-level variability, especially in summer. Temperature and wind fetch emerged as key environmental drivers, with larger seasonal fluctuations observed under high-exposure conditions. Notably, several dominant species were non-native, underscoring the importance of early detection. These findings highlight the value of molecular monitoring for predicting biofouling risks and informing site-specific antifouling strategies in dynamic marine environments.

ecology↗
bioRxiv2025

SpaTranslator: A deep generative framework for universal spatial multi-omics cross-modality translation

Recent advances in spatial omics technologies have enabled the simultaneous analysis of multiple molecular patterns in tissue sections, offering unprecedented insights into cellular microenvironments. However, the high cost of measurements and the sparsity of data restrict the availability of paired spatial multi-omics datasets. Here, we present SpaTranslator, a deep generative framework that integrates graph neural networks with an adversarial variational generative model to fully capture spatial characteristics and enable effective cross-modality translation of spatial omics data, enabling simulation of paired spatial multi-omics data from single-omics measurements. Extensive experiments demonstrate that SpaTranslator consistently outperforms baseline methods in both clustering accuracy and biological coherence across various real-world scenarios, including spatial transcriptomics-epigenomics and spatial transcriptomics-proteomics translation tasks. Furthermore, SpaTranslator provides biologically meaningful insights through marker genes and proteins recovery, motif enrichment analysis, and gene regulation inference. Our work offers an effective and adaptable solution for spatial multi-omics cross-modality translation, supporting a broad range of biological and biomedical research.

bioinformatics↗
bioRxiv2025

Transfer Learning Assessment of Data-driven Crystallisation Processes via Constrained Neural Ordinary Differential Equations

Modelling complex crystallisation processes remains challenging due to limited experimental datasets, high measurement noise, and the need for generalisability across varying operating conditions. Neural Ordinary Differential Equations (NODEs) and transfer learning (TL) offer promising tools to overcome these limitations by providing data-efficient, flexible, and transferable modelling frameworks. This work investigates the use of NODEs to model protein crystallisation dynamics under data-scarce conditions. A NODE trained on a data-rich source system successfully captures solute consumption and particle size dynamics, but when applied to data-sparse target systems, scratch-trained NODEs exhibit limited generalisation and unphysical behaviours. To address this, several TL strategies are evaluated, including layer freezing, parameter deviation penalisation, and system-embedding within the neural architecture. Results show that layer freezing and deviation penalty consistently improve knowledge transfer, while system-embedding offers robustness in noisy or undersampled datasets. In addition, physics-informed NODEs, constrained to enforce monotonic concentration decay and crystal growth, demonstrate greater stability under high noise and sparse measurement regimes, ensuring physically consistent predictions. Overall, the combination of constrained NODEs with appropriate TL strategies provides a robust framework for accurate, transferable modelling of crystallisation systems in low-data regimes.

bioengineering↗
bioRxiv2025

Macrophage-derived reactive oxygen species promote Salmonella aggresome formation contributing to bacterial antibiotic persistence

In this study, we reveal that macrophage-derived reactive oxygen species (ROS) can trigger the rapid formation of Salmonella aggresomes, which substantially contribute to the increased frequency of persisters induced by phagocytosis. Salmonella containing aggresomes exhibited a dormant phenotype characterized by reduced adenosine triphosphate (ATP) levels and decreased metabolic activity. Furthermore, these dormant bacteria showed upregulated expression of Salmonella pathogenicity island 1 (SPI-1)-encoded type III secretion system (T3SS)-related genes, followed by later expression of SPI-2 T3SS-related genes when macrophages ROS production declined. Our results demonstrate that Salmonella containing aggresomes can enter a dormant state to escape antibiotic attack, while crucially maintaining the ability to resuscitate when the stress environment is improved. Research on bacterial aggresomes could potentially provide therapeutic strategies to combat bacterial antibiotic persistence. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=122 SRC="FIGDIR/small/657582v1_ufig1.gif" ALT="Figure 1"> View larger version (55K): org.highwire.dtl.DTLVardef@7e5f1forg.highwire.dtl.DTLVardef@970a39org.highwire.dtl.DTLVardef@204449org.highwire.dtl.DTLVardef@632c47_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗
bioRxiv2025

Single-cell Characterization of DNA Hydroxymethylation of the Mouse Brain During Aging

DNA methylation dynamics, including 5-hydroxymethylcytosine (5hmC) and 5-methylcytosine (5mC), critically regulate brain function, yet conventional methods cannot distinguish these modifications, obscuring their distinct roles in gene regulation and aging. We present Joint- Cabernet, a bisulfite-free single-cell platform enabling simultaneous profiling of 5hmC, 5mC, and transcriptomes. Applying Joint-Cabernet to 84,071 nuclei from adult and aged mouse brains, we resolved cell-type-specific DNA hydroxymethylation landscapes, revealing elevated 5hmCG and 5hmCH levels in transcriptionally active genes across neuronal subtypes and spatial gradients in cortical layers. During aging, 5hmCG accumulates globally but is selectively enriched at open chromatin loci, aligning with the upregulation of cell-type-specific genes in distinct brain cell types. This single-cell DNA methylation brain cell atlas provides a framework for studying methylation- driven mechanisms in brain aging and neurodegenerative diseases.

genetics↗
bioRxiv2025

Virus-host interactome reveals host cellular pathways perturbed by tick-borne encephalitis virus infection

Tick-borne encephalitis virus (TBEV) poses an increasingly significant threat to public health. Here, we delved into the virus-host interactome and discovered that TBEV has a profound impact on multiple host cellular pathways. Viral pre-membrane protein (prM), non-structural proteins NS1, NS2B3, NS4A, and NS5 remodel the hosts innate immune responses. PrM and NS2B3 are involved in autophagy, NS4A is associated with neurodegenerative diseases, and NS5 participates in spliceosome dynamics, ribosomal biogenesis, cell cycle regulation, and DNA damage response. Notably, TBEV infection causes G0/G1 cell cycle arrest in host cells. NS5 interacts with histone acetyltransferase P300 to upregulate P16 expression, suppressing CDK4 and resulting in cell cycle arrest. Elevated P16 and reduced CDK4 levels were observed in TBEV-infected brain organoids. The P300 inhibitors and CDK4 agonist can reverse virus-induced cell cycle arrest and inhibit viral replication. Further analysis uncovered potential antiviral targets like KAT6A, XIAP, and RIOK1/BRD9. These findings provide valuable insights into TBEV pathogenicity and hold promise for antiviral drug development.

microbiology↗
bioRxiv2025

Regenerative failure of sympathetic axons contributes to deficits in functional recovery after nerve injury

Renewed scientific interest in sympathetic modulation of muscle and neuromuscular junctions has spurred a flurry of new discoveries with major implications for motor diseases. However, the role sympathetic axons play in the persistent dysfunction that occurs after nerve injuries remains to be explored. Peripheral nerve injuries are common and lead to motor, sensory, and autonomic deficits that result in lifelong disabilities. Given the importance of sympathetic signaling in muscle metabolic health and maintaining bodily homeostasis, it is imperative to understand the regenerative capacity of sympathetic axons after injury. Therefore, we tested sympathetic axon regeneration and functional reinnervation of skin and muscle, both acute and long-term, using a battery of anatomical, pharmacological, chemogenetic, cell culture, analytical chemistry, and electrophysiological techniques. We employed several established growth-enhancing interventions, including electrical stimulation and conditioning lesion, as well as an innovative tool called bioluminescent optogenetics. Our results indicate that sympathetic regeneration is not enhanced by any of these treatments and may even be detrimental to sympathetic regeneration. Despite the complete return of motor reinnervation after sciatic nerve injury, gastrocnemius muscle atrophy and deficits in muscle cellular energy charge, as measured by relative ATP, ADP, and AMP concentrations, persisted long after injury, even with electrical stimulation. We suggest that these long-term deficits in muscle energy charge and atrophy are related to the deficiency in sympathetic axon regeneration. New studies are needed to better understand the mechanisms underlying sympathetic regeneration to develop therapeutics that can enhance the regeneration of all axon types.

neuroscience↗
bioRxiv2024

Deep generative design of neutralizing nanobodies againstSARS-CoV-2 variants

In recent years, single-domain heavy chain antibodies (nanobodies), with only one-tenth the molecular weight of conventional antibodies, have emerged as important therapeutic proteins in the fight against SARS-CoV-2. However, the rapid mutation of the virus often renders existing nanobodies ineffective, underscoring the need to develop nanobodies that specifically target new variants. Traditional methods for discovering nanobodies are time-consuming and complex, making it difficult to efficiently identify nanobodies that bind to specific epitopes. To address this, we propose a de novo nanobody design method based on a Generative Adversarial Network (GAN). We developed a deep generative model, AiCDR, which consists of three discriminators and one generator. By determining the contribution of these three discriminators to the entire network, we can enhance the discrimination of generated sequences and reduce the similarity of the generated CDR3 sequences with natural peptides and random sequences, thereby ensuring their nature-like properties. These generated CDR3 sequences were then grafted onto humanized nanobody scaffolds, resulting in a structural library of approximately 104 nanobodies with natural-like properties. Using computational methods, we screened this library against the Spike (S) protein of the SARS-CoV-2 Omicron variant and identified 10 candidate nanobodies. Functional assays confirmed that two of these nanobodies exhibited neutralizing activity against the S protein. Our study demonstrates the potential of deep learning in nanobody design and offers a novel approach for developing nanobodies that target rapidly evolving viral variants.

bioinformatics↗