Biology subjects

Zhang, H.

Publications and source records attributed to Zhang, H..

At least 397 records · Page 22Linked to original sources

bioRxiv2025

Engineering a mouse-adapted SADS-CoV and establishing a neonatal mouse model to study its infection

Swine acute diarrhea syndrome coronavirus (SADS-CoV) is an emerging bat-origin alphacoronavirus causing severe disease in neonatal piglets, with significant economic losses to the swine industry as a consequence. The virus exhibits a broad species tropism, infecting cells derived from pigs, humans and mice, highlighting its potential for cross-species transmission. Due to drawbacks associated with the use of young piglets, there is a need for an appropriate small animal model to study SADS-CoV biology. Here we established a mouse infection model based on a murinized mutant of the virus, mSADS-CoV, in which the ectodomain of the viral spike protein was replaced by that of the murine coronavirus mouse hepatitis virus (MHV). This chimeric virus, generated through targeted RNA recombination, replicated efficiently in murine cell cultures and exhibited an age-dependent infection in neonatal mice that was lethal in 2-day-old BALB/c mice affecting various organs, notably the intestine. We validated our infection model by successfully verifying the efficacy of the RNA-dependent RNA polymerase inhibitor remdesivir (RDV). The model will serve as a valuable tool for studying SADS-CoV pathogenesis and for elucidating the roles of host factors in viral replication as well as for preclinical evaluation of vaccine candidates and antiviral compounds targeting the viral replication machinery. Author SummarySADS-CoV poses a dual threat to the swine industry and public health because of its broad species tropism and potential for cross-species transmission. The emergence of other bat-derived coronaviruses, including severe acute respiratory syndrome coronavirus (SARS-CoV), SARS-CoV-2, and Middle East respiratory syndrome coronavirus (MERS-CoV), underscores the need for robust models to study these pathogens. The successful rescue of mSADS-CoV and the development of a mouse infection model represent significant advancements in SADS-CoV research. This model not only enables the evaluation of antiviral therapeutics such as RDV but also provides a powerful platform for investigating viral replication mechanisms and host-pathogen interactions, offering critical insights for pandemic preparedness.

microbiology↗

bioRxiv2025

Ninjurin2 Regulates Vascular Smooth Muscle Cell Phenotypic Switching and Vascular Remodeling through Interacting with PDGF Receptor-β

Vascular smooth muscle cell (VSMC) phenotypic switching from a contractile to synthetic state plays a pivotal role in vascular remodeling. Although genome-wide association studies have linked NINJ2 (encode Ninjurin2) polymorphisms are associated with risk of atherosclerosis and related conditions like ischemic stroke and coronary artery disease, its functional role in regulating VSMC plasticity remains unexplored. Here, using Ninj2 knockout (Ninj2-/-) and smooth muscle-specific overexpressing (Ninj2smcTG) mice subjected to carotid artery ligation, we demonstrate that Ninjurin2 deficiency exacerbates injury-induced intimal hyperplasia, whereas its overexpression attenuates this pathological remodeling. Mechanistically, Ninjurin2 knockdown promoted PDGF-BB-mediated VSMC phenotypic switching, proliferation, and migration, while biochemical studies revealed that Ninjurin2 interacts with PDGFR{beta} to suppress FAK signaling. These findings identify Ninjurin2 as a novel regulator of VSMC phenotypic switching via the PDGFR{beta}/FAK axis, offering new mechanistic insights into vascular remodeling and potential therapeutic avenues for atherosclerosis.

pathology↗

bioRxiv2025

Improving Cellular Senescence Detection with Weak Gene Markers in Single-Cell RNA-seq vi Iterative Imputation

BackgroundCellular senescence represents a stable cell cycle arrest state that plays critical roles in tissue aging and age-related pathologies. Although single-cell RNA sequencing (scRNA-seq) enables comprehensive profiling of millions of cells in aged tissues, reliably identifying senescent cells remains challenging due to the weak and non-specific expression of their marker genes. Existing computational methods can detect active cell states through gene set scoring, but their performance for senescence genes has not been rigorously evaluated. ResultsIn this study, we show that senescence genes display weak, non-specific expression patterns across different tissues and are susceptible to dropout events in scRNA-seq. Current scoring approaches suffer intrinsic limitations in single-cell detection using weakly expressed markers. Simply expanding the gene set of weak markers failed to improve detection accuracy. To overcome these limitations, we developed ICE (Imputation-based Cell Enrichment), a computational framework that combines expression imputation with iterative marker refinement. ICE substantially improved detection precision in pancreatic cells with weakly expressed markers, increasing it from 56% to 98%. Using consensus markers, ICE identified senescent fibro-adipogenic progenitors (FAPs) in muscle tissue, characterized by elevated expression of senescence-associated secretory phenotype (SASP) genes. When applied to individual senescence markers (CDKN1A/p21, CDKN2A/p16, ATF3, and MX1), ICE successfully identified marker-specific cell populations. These included stressed {beta} cells in aging and type-I interferon-responsive microglia in Alzheimers disease (AD). ConclusionsOur study introduces ICE, a rigorous framework for detecting cell states defined by weakly expressed markers at the single-cell level. This tool enables the reliable identification of senescence-associated populations, facilitating a deeper characterization of their heterogeneity and temporal dynamics across diverse human tissues and disease contexts.

genomics↗

bioRxiv2025

DeepAden: an explainable machine learning for substrate specificity prediction in nonribosomal peptide synthetases

Microbial non-ribosomal peptides (NRPs) exhibit remarkable structural diversity and serve as valuable sources of lead compounds for clinical drug development. The biosynthesis of NRPs relies on non-ribosomal peptide synthetases (NRPSs), in which adenylation (A) domains play a pivotal role in defining the core structure by selectively recognizing and activating amino acid substrates. Accurately predicting the substrate specificities of A-domains is thus essential for understanding the core structural and biosynthetic logic of NRPs. Here, we present DeepAden, a two-stage deep learning framework. In the first stage, a graph attention network (GAT)-based model localizes 27-residue binding pockets within 6 [A] of bound substrates and convert these into pocket representations. In the second stage, pocket representations are then encoded alongside substrate information using pretrained language models, and aligned using contrastive learning. In addition, we introduce a SHapley Additive exPlanations (SHAP)-guided data augmentation strategy to mitigate class imbalance and improve robustness, particularly for nonproteinogenic substrates. DeepAden achieves competitive performance compared with state-of-the-art tools on a benchmark dataset, and enabled the identification of two Streptomyces NRPS gene clusters through accurate A-domain substrates specificity predictions. DeepAden offers a powerful tool for precise pocket localization and robust substrate prediction, accelerating the discovery and characterization of novel NRP natural products for future work. The DeepAden web server is available at https://deepnp.site/.

bioinformatics↗

bioRxiv2025

Diffusion Neuroimaging of Speech Acquisition in Infants

How multisystem white-matter maturation supports early speech acquisition in infancy remains unclear. Using longitudinal diffusion MRI from the Baby Con-nectome Project, we quantified neurite density maturation from birth to 24 months across 33 white-matter tracts spanning motor, auditory, visual, ventral-language, and dorsal-language systems. Maturation was coordinated across systems rather than confined to canonical language pathways: dorsal-language association tracts began relatively immature near birth but showed steep postnatal growth, pro-gressively rebalancing with motor and sensory systems. Tract-behavior coupling was likewise distributed across developmental domains, with broadly positive age-adjusted associations for motor outcomes, a sparser, partly motor-like profile for gesture-related phrase comprehension, and predominantly negative associa-tions for lexical speech-related outcomes. Exploratory follow-up analyses further suggested motor and social precursor capacities that may help mediate these in-direct developmental routes from structural maturation to later speech-related outcomes. These findings support a multisystem developmental account in which early speech acquisition emerges not from an isolated language pathway, but from coordinated maturation across multiple systems and the precursor skills that link them.

neuroscience↗

bioRxiv2025

Ketogenic diet prevents obesity-associated pancreatic cancer independent of weight loss and induces pancreatic metabolic reprogramming

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive cancer with poor outcomes. Obesity is a risk factor for several cancers including PDAC due to metabolic dysregulation and inflammation. The ketogenic diet (KD) can alter metabolism and has been evaluated for its effects on tumor progression in non-obese but not obese PDAC using genetically engineered mouse models (GEMMs). We hypothesized that ketone bodies and a KD alter cell and tumor metabolism. We show that ketone treatments altered pyrimidine metabolism in PDAC cells. Moreover, in an obese PDAC GEMM, KD prevented tumor progression independent of weight loss but promoted PDAC in a non-obese PDAC GEMM. The KD-specific delay of obesity-associated PDAC was associated with pancreatic metabolic shifts in pyrimidine, cysteine and methionine, and arginine and proline pathways. These findings suggest potential benefits of a KD in preventing obesity-associated PDAC, but highlights some risks in non-obese settings.

cancer biology↗

bioRxiv2025

DNA2 and FANCM function in two distinctive pathways in disrupting TERRA R-loops and suppressing replication stress at ALT telomeres

Cancers maintain their telomeres through two main telomere maintenance mechanisms (TMMs): 85-90% of cancers rely on telomerase, while 10-15% of cancers adopt the Alternative Lengthening of Telomere (ALT) pathway. Previously, we and others reported that FANCM, one of the Fanconi Anemia proteins, plays a critical role in suppressing replication stress and DNA damage at ALT telomeres by actively disrupting TERRA R-loops [1-4]. Here, we showed that inactivation of DNA2 in ALT-positive (ALT+) cells, but not in telomerase-positive (TEL+) cells, induces a robust increase of replication stress and DNA damage at telomeres, which leads to a pronounced increase of many ALT properties, including telomere dysfunction-induced foci (TIFs), ALT-associated PML bodies (APBs), and C-circles. We further demonstrated that depletion of DNA2 induces a pronounced increase of TERRA R-loops and a decrease in replication efficiency at ALT telomeres. Most importantly, we uncovered a strong additive genetic interaction between DNA2 and FANCM in the ALT pathway. Furthermore, co-depletion of DNA2 and FANCM causes synthetic lethality in ALT+ cells, but not in TEL+ cells, suggesting that targeting DNA2 and FANCM could be a viable strategy to treat ALT+ cancers. Finally, utilizing the single-molecule telomere assay via optical mapping (SMTA-OM) technology, we thoroughly characterized genome-wide changes in DNA2 deficient cells and FANCM deficient cells and found that most chromosome arms manifested increased telomere length. Unexpectedly, we uncovered many chromosome arm-specific telomere changes in those cells, suggesting that telomeres at different chromosome arms may regulate and respond to replication stress differently. Collectively, our study not only shed new light on the molecular mechanisms of the ALT pathway, but also discovered a new strategy for targeting ALT+ cancer.

cancer biology↗

bioRxiv2025

Deep Blood Proteomics Identifies over 12,000 Proteins Providing Valuable Information about the State of the Human Body

Blood is a valuable resource for clinical research, offering insight into physiological and pathological states. However, the specific proteins detectable in blood and the optimal proteomic methods for their detection have not been rigorously investigated and documented. To address this, we conducted various blood proteomic strategies, including directly blood proteomic analysis, high-abundance protein depletion, low-abundance protein enrichment, and extracellular vesicle enrichment using data-independent acquisition or targeted proteomics. These approaches identified 11,679 protein groups in plasma from healthy individuals. In 136 pancreatic ductal adenocarcinoma whole blood samples, 6,956 protein groups were found, including 678 not seen in healthy samples, expanding the total to 12,357 blood proteins. This represents the most comprehensive blood proteome to date. To support broader access and analysis, we developed the Human Blood Proteome (HuBP) database, detailing protein detectability, abundance, and reproducibility across workflows, sample types, and disease contexts.

systems biology↗

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

scDenorm: a denormalisation tool for integrating single-cell transcriptomics data

Integrating single-cell omics data at an atlas scale enhances our understanding of cell types and disease mechanisms. However, the integration of data processed by different normalisation methods can lead to biases, such as unexpected batch effects and gene expression distortion, leading to misinterpretations in downstream analysis. To address these challenges, we present scDenorm, an algorithm that reverts normalised single-cell omics data to raw counts, preserving the integrity of the original measurements and ensuring consistent data processing during integration. We evaluated scDenorms performance on large-scale datasets and benchmarked its impact on data integration and downstream analysis across three datasets.

bioinformatics↗

bioRxiv2025

Aberrant Fibro-Adipogenic Progenitor Subpopulations Drive Volumetric Muscle Loss-Induced Fibrosis

Volumetric muscle loss (VML) injuries result in chronic fibrosis, inflammation, and persistent functional deficits. Fibro-adipogenic progenitor (FAP) cells are a heterogeneous, muscle-resident stromal cell population that play a crucial role in muscle regeneration, but also contribute to fibrosis in muscle disease. The role of FAPs in VML is not well established and may be critical target to ensure functional muscle regeneration after VML. We utilized a VML model in the mouse quadriceps to study the location, secretome, surface marker distribution, gene expression, and single-cell transcriptional profile of FAPs after VML. After VML, a subpopulation of FAPs highly expressed {beta}1-integrin and were elevated in the post-VML muscle tissue; these FAPs had increased fibrotic gene expression and increased myofibroblast differentiation potential. Transforming growth factor-{beta}1 (TGF-{beta}1) and tissue inhibitor of matrix metalloproteinase 1 (TIMP1) were identified as secreted proteins from VML derived FAPs that produced both pro-fibrotic and anti-myogenic signaling. These data establish an aberrant FAP sub-population that are elevated in VML injury and provides novel targets for future scarless muscle regeneration in VML.

pathology↗

bioRxiv2025

Automated behavior classification of julius seizure mutants in Drosophila reveals stereotyped seizure stages with genotype specificity

Bang-sensitive (BS) Drosophila mutants exhibit a stereotyped pattern of seizure behavior after mechanical disturbances. We previously identified mutations in the julius seizure (jus) gene, formerly CG14509, can induce BS seizures. However, the behavioral manifestations of the seizure phenotype of the various jus mutants have not been fully characterized. Here, we developed a machine learning pipeline featuring LASC (Long short-term memory and Attention mechanism for Sequence Classification) for automatic phenotyping of jus mutant videos. LASC achieves 90% classification accuracy in distinguishing five phases: paralysis (P), tonic seizure (T), spasm (S), clonic seizure (C), and recovery (R). Applying the trained LASC model to multiple jus lines showed they use a common repertoire of seizure stages and followed the general P[->]T[->]S[->]C[->]R progression, but each genotype exhibited unique patterns of stage duration and transition probabilities. Remarkably, stage usage patterns are distinct among the mutant genotypes. These findings establish that while all jus mutants adhere to stereotyped behavioral rules, each allele generates a distinct signature in stage usage. This work demonstrates how advanced behavioral quantification can reveal previously hidden relationships between gene mutation and complex motor outputs. More broadly, the complete pipeline presented here can pave the way for high-throughput, automated drug screening for epilepsy.

neuroscience↗

bioRxiv2025

WHEP Domain of Glycyl-tRNA Synthetase Regulates Neuropilin 1 Binding and Vascular Permeability

AbstractAminoacyl-tRNA synthetases (aaRSs) charge tRNAs with their cognate amino acids, ensuring accurate translation of the genetic code from mRNA to protein. During eukaryotic evolution, aaRSs acquired additional domains with unclear functions, including the WHEP domain, a two-helix bundle found in several eukaryotic aaRSs such as glycyl-tRNA synthetase (GlyRS, encoded by GARS1). We generated Gars1{Delta}WHEP mutant mice lacking exon 2, which disrupts most of the WHEP domain. Homozygous Gars1{Delta}WHEP/{Delta}WHEP mice exhibited late embryonic or neonatal lethality, with delayed lung development, characterized by reduced airway dilation (inflatability), and increased vascular leakage. Disruption of the WHEP domain did not impair tRNA aminoacylation but inhibited the free release of GlyRS from cells. Instead, GlyRS{Delta}WHEP was found in the membrane fractions and showed a stronger interaction with the extracellular region of neuropilin 1 (Nrp1) receptor compared to full-length GlyRS. This aberrant interaction enhanced Nrp1s endocytic activity and significantly reduced the localization of the Nrp1 interactor VE-cadherin at the adherens junctions of endothelial cells. A heterozygous knockout of Nrp1 in the Gars1 {Delta}WHEP/{Delta}WHEP mice partially rescued body weight and vascular permeability defects. This study establishes a physiological role for the GlyRS WHEP domain in lung development and its regulation of GlyRS-Nrp1 interaction and vascular permeability.

molecular biology↗

bioRxiv2025

Predicting Neural Activity from Connectome Embedding Spaces

Understanding how structured patterns of neural activity emerge from the underlying connectivity is fundamental to elucidating brain function. While cortical connectomes are intrinsically high-dimensional, population activity typically resides in a much lower-dimensional subspace. Consequently, only a small fraction of the information encoded in the connectome appears relevant for shaping activity. Can we identify low-dimensional features of the connectome features be that reliably predict neural activity? Leveraging the MICrONS dataset, which combines millimeter-scale, nanometer-resolution connectivity with simultaneously recorded in-vivo activity, we demonstrate a statistically significant alignment between morphological and functional similarity, quantified by subspace angles and centered kernel alignment. Topological analyses further reveal that the representation spaces of both the connectome and neural activity share a low-dimensional hyperbolic geometry with exponential scaling. These parallels motivated the hypothesis that embedding anatomical affinities into an appropriate geometric space can isolate the functionally relevant features of the connectome. We therefore applied multidimensional scaling to generate such embeddings and trained a simple linear model to reconstruct neuronal activity. Remarkably, the embedded connectome explained 68% in activity similarity, surpassing models that had direct access to activity similarity itself and outperforming similarly simple models that used the full high-dimensional connectome (56%). Our findings uncover a robust structure-function coupling: geometry-aware dimensionality reduction discards much of the connectomes microscopic detail yet yields superior predictions of neural activity. This suggests that synaptic wiring implicitly encodes an abstract, low-dimensional organization that underlies the observed low-dimensional cortical activity.

neuroscience↗

bioRxiv2025

Unraveling the Molecular Mechanisms of ABHD5 Membrane Targeting

ABHD5 is a master regulator of PNPLA family lipases, particularly PNPLA2 (ATGL), the rate-limiting triglyceride (TAG) hydrolase in metabolic tissues. Despite its central role in lipid metabolism, the molecular basis by which ABHD5 recognizes membranes and regulates enzymatic activity remains poorly understood. Here, we report an integrated computational-experimental study revealing how the /{beta}-hydrolase domain-containing protein 5 (ABHD5) dynamically engages lipid droplet (LD) and endoplasmic reticulum membranes to control lipolytic activation. Using multiscale molecular dynamics simulations, hydrogen-deuterium exchange mass spectrometry, and site-directed mutagenesis, we uncover a sequential dual-site membrane recognition mechanism, where the N-terminus provides initial anchoring and a lid helix within the insertion segment forms a crucial secondary contact. Membrane binding triggers a dramatic conformational switch in this lid, expanding the pseudosubstrate pocket and transforming ABHD5 into an active and membrane-localized regulator. This structural transition is coupled to membrane remodeling, inducing localized curvature and forming a triacylglycerol-enriched nanodomain beneath the ABHD5 pseudosubstrate pocket. This bidirectional interaction between ABHD5 and the membrane provides a persuasive mechanism for interfacial activation. Our findings establish new principles for how LD-binding proteins achieve functional specificity through membrane-dependent regulation, offering novel molecular targets for interventions in metabolic diseases. TeaserA key protein controls fat breakdown by changing shape when it binds membranes and reshapes local fat droplets.

biophysics↗

bioRxiv2025

UNCOVERseq Enables Sensitive and Controlled Gene Editing Off-Target Nomination Across CRISPR-Cas Modalities and Systems

The rapid development of CRISPR-Cas gene editing technologies has revolutionized genetic medicine, offering unprecedented precision and potential for treating a wide array of genetic disorders. However, assessing the risks of unintended gene editing effects remains critical, and is complicated by new editing modalities and unclear analytical guidelines. We present UNCOVERseq (Unbiased Nomination of CRISPR Off-target Variants using Enhanced RhPCR), an improved in cellulo off-target nomination workflow designed to sensitively nominate off-target sites (<0.01% editing) with defined input requirements and analytical process controls to provide empirical performance evidence across diverse circumstances. Using this workflow, we nominated off-targets across 192 guide RNAs (gRNAs) and demonstrated superior performance compared to existing methodologies. We identified a subset of six gRNAs with a dynamic range of specificity and confirmed the relevance and high true positive rate of our nomination method, providing relative risk assessments for multiple modalities (S.p. Cas9 and derived high-fidelity variants / base editors) in a translational system involving hematopoietic stem and progenitor cells (HSPCs). Additionally, we established that double-strand break (DSB) editing retains a strong, positive rank correlation to single-strand break (SSB)-mediated base editing, highlighting the importance of DSB nomination sites as candidate loci for base editing. Overall, UNCOVERseq improves informed risk assessment of gene editing in translational systems by enhancing the quality of off-target nomination.

molecular biology↗

bioRxiv2025

Unveiling the Developmental Dynamics and Functional Role of Odorant Receptor Co-receptor (Orco) in Aedes albopictus: A Novel Mechanism for Regulating Tuning Odorant Receptor Expression

As one of the most aggressive disease vectors, the Asian tiger mosquito Aedes albopictus relies heavily on its olfactory system to search for food in the larval stage, locate hosts after eclosion, and identify suitable oviposition sites after blood feeding. In mosquitoes and other insects, the olfactory system detects environmental odors primarily through a diverse repertoire of tuning odorant receptors (ORs), which require the highly conserved odorant receptor co-receptor (Orco) to function. While Orcos role in enabling tuning receptor function is well established, its cellular localization patterns, developmental expression dynamics, and system-wide impact on olfactory physiology and behavior remain understudied in Ae. albopictus. To address this knowledge gap, we leveraged the Q-system to systematically characterize Orco-expressing neurons (ORNs) across embryonic, larval, pupal and adult stages of Ae. albopictus. We uncovered a dramatic reorganization of the olfactory system during metamorphosis: ORNs were observed as early as the embryonic stage and expanded during larval development before rapid degeneration and re-generation in the pupal stage resulting in the final population of adult ORNs. Orco expression in adults spanned the olfactory neurons of the antennae, labella, and maxillary palps in both male and female mosquitoes, consistent with its conserved peripheral distribution. To further investigate the functional implications of Orco, we generated Orco knockout mutants and strikingly discovered that Orco knockout mutants displayed significant widespread downregulation of tuning Ors, suggesting that Orco may influence OR expression or stability. Electrophysiological recordings confirmed significantly attenuated responses to human volatiles in Orco mutants, and behavioral assays demonstrated a marked decline in blood-feeding efficiency and elimination of host preference in females. Together, these findings reveal dynamic reorganization of ORNs during mosquito development and uncover the critical role of Orco in maintaining the integrity and function of the olfactory system, providing insights that may inform novel, next-generation vector control strategies.

neuroscience↗

bioRxiv2025

MicroProphet: A Digital Twin Framework for Predicting Microbial Community Dynamics with Personalized Precision

The ability to accurately predict the dynamic evolution of microbial communities is critical for advancing personalized medicine, precision intervention, and ecological system management. However, the irregular sampling, high missingness, and complex temporal behaviors that characterize longitudinal microbiome datasets present substantial challenges to reliable forecasting. Here we propose MicroProphet, a personalized digital twin framework capable of accurately forecasting microbial abundance trajectories from incomplete longitudinal observations without the need for data interpolation. By leveraging a time-aware Transformer architecture, MicroProphet reconstructs individualized microbial trajectories using as little as the initial 30% of time points, capturing critical transitional states through its attention mechanism. We demonstrate its robust cross-ecosystem generalizability across synthetic communities, human gut microbiomes, infant gut development, and corpse decomposition. In clinical contexts, MicroProphet enables early identification of disease-related microbial shifts and supports intervention timing optimization, exemplified in inflammatory bowel disease and antibiotic perturbation responses. By transforming incomplete and sparse data into actionable forecasts, MicroProphet establishes a foundation for real-time microbial monitoring, therapeutic decision support, and precision ecological management, paving the way for broader applications of digital twin systems in biology and personalized healthcare. HighlightsO_LIMicroProphet establishes a personalized digital twin framework for forecasting biological dynamics from incomplete longitudinal observations, enabling precision health monitoring and intervention planning. C_LIO_LIBy leveraging a transformer-based architecture, MicroProphet accurately reconstructs microbial community trajectories using as little as 30% of initial time points, without the need for data interpolation. C_LIO_LIThe framework demonstrates robust cross-ecosystem generalizability across clinical, early-life development, and forensic contexts. C_LIO_LIMicroProphet empowers real-time tracking of microbial shifts, early detection of disease-associated transitions, and timing optimization for microbiome-targeted therapeutic strategies. C_LI

microbiology↗