Biology subjects

Lu, C.

Publications and source records attributed to Lu, C..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

Hepatocyte FGFR2 Mediates the Antifibrotic Effects of FGF10 in Advanced Liver Fibrosis.

Fibroblast growth factor 10 (FGF10) supports epithelial repair, but its role in liver fibrogenesis remains uncertain. Here we found that hepatic FGF10 and its receptor FGFR2 decline with progression to advanced fibrosis in patient biopsies and in carbon tetrachloride (CCl4)- and diet-induced mouse models. Restoring hepatic FGF10 with recombinant human FGF10 or adeno-associated virus mediated liver-targeted expression attenuated and partially reversed established advanced bridging fibrosis, reduced inflammation, and decreased hepatocyte apoptosis, and these benefits required hepatocyte FGFR2 because hepatocyte-specific Fgfr2 deletion abolished protection. In primary hepatocytes, FGF10 activated FGFR2-FRS2, increased inhibitory GSK3{beta}(Ser9) phosphorylation, and suppressed NF-{kappa}B activation, reducing TGF-{beta}1 and other cytokines and thereby limiting paracrine hepatic stellate cells activation. The molecular and histologic benefits extended to metabolically primed steatohepatitis, supporting translatability beyond toxicant injury. Mechanistic signaling readouts confirm FGFR2 engagement, while the main text centers on the hepatocyte FGF10-FGFR2 axis as an intrinsic brake on fibrogenesis. Together, these data identify a hepatocyte-centric FGF10-FGFR2 axis as a negative regulator and potential regressor of fibrogenesis, highlighting it as a tractable therapeutic target in advanced liver fibrosis.

pharmacology and toxicology↗

bioRxiv2025

Pan-cancer 3D genomic analysis revealed extremely long Polycomb loops as the biomarker for sensitivity to Polycomb inhibition

Polycomb targeted loci form long-range chromatin interactions independent of CTCF-cohesion and are demarcated by low DNA methylation(1-3). Polycomb targets form extremely long-range loops (long Polycomb loops) that can separate anchors up to 60 Mb. These loops predominantly occur in cells of self-renewal status, such as human hematopoietic stem cells (HSC) and mouse embryonic stem cells (ESC), but rarely in cell lines. To identify long Polycomb loops in cancer, we initiated a pan-cancer survey of long Polycomb loops in a collection of 264 tumor samples (33 acute myeloid leukemias, 17 T cell lymphoblastic leukemia, 29 Breast cancer samples, 63 pediatric brain tumors, 70 prostate cancers, and 42 colon cancers). We found most cancers, including all prostate cancers and colon cancers, lack long Polycomb loops except pediatric brain tumors and certain AMLs. In pediatric brain tumors, we found 30% of the ependymoma PFA subtype (which shows globally depleted H3K27me3) notably displayed strong long Polycomb loop interactions. AML displayed more diverse levels of long Polycomb loop interactions. Most AMLs lost both long Polycomb loops with Polycomb binding loss at the loop anchors. Whereas 10% of AMLs retain long Polycomb loops as strong as in HSCs. These AMLs recurrently carry mutations in CEBPA and STAG2, which are not associated with the Polycomb complex or DNA methylation machinery. We found that long Polycomb loop strong AML is sensitive to EZH2 inhibition, which induces cell differentiation. Conversely, PRC1 component dependency can be used to predict long Polycomb loop formation in B-cell lymphoblastic leukemia cell lines. Our analysis suggests that the oncogenesis process antagonizes the long Polycomb loop maintenance in most cancers, yet certain cancers may still preserve strong long Polycomb loops from cell-of-origin. The maintenance of long Polycomb loops sensitizes cells to Polycomb inhibition, indicating that such loops could be an epigenomic biomarker for pharmacological or genetic Polycomb inhibition.

cancer biology↗

bioRxiv2025

A machine learning method for calculating highly localized protein stabilities

The residue-level free energy of opening ({Delta}Gop) is the ultimate thermodynamic descriptor of localized protein stability, providing valuable information about the protein ensemble at physiologically relevant timescales and conditions. PFNet instantly determines {Delta}Gop for arbitrarily large proteins and complexes from conventional peptide-level hydrogen exchange/mass spectrometry (HX/MS) datasets. It unlocks the full potential of HX/MS, democratizing the method and establishing quantitative, scalable and accessible analysis (https://github.com/glasgowlab/PFNet).

biophysics↗

bioRxiv2025

HXMS: a standardized file format for HX/MS data

MotivationHydrogen/deuterium exchange-mass spectrometry (HX-MS) is a rapidly expanding technique used to investigate protein conformational ensembles. The growing popularity and utility of HX-MS has driven the development of diverse instrumentation and software, resulting in inconsistent, non-standardized data analysis and representation. Most HX-MS data formats also employ only mean deuteration representations of the data rather than full isotopic mass spectra, which reduces the information content of the data and limits downstream quantitative analysis. ResultsInspired by reliable protein structure and genomics data formats, we present HXMS, a unified, lightweight, scalable, and human-readable file format for HX-MS data. The HXMS format preserves the isotopic mass envelopes for all peptides, captures the full experimental time-course including fully deuterated control samples, and contains all other key information. It supports multimodal distributions, post-translational modifications (PTMs), and experimental replicates. To promote compatibility with existing HX-MS workflows, we also developed PFLink, a Python package that converts exported data files from commonly used HX-MS software to the HXMS format. PFLink and the HXMS format will enable more quantitative, higher-resolution data processing, improved data sharing and storage among HX-MS practitioners, future machine learning applications, and further developments in HX-MS analysis. Availability and implementationPFLink is publicly available to install locally on HuggingFace, alongside documentation, or use online at HuggingFace (https://huggingface.co/spaces/glasgow-lab/PFlink). The supplemental information includes sample input files, sample HXMS files, and a generic unfilled PFlink custom CSV file that users may populate with key experimental conditions and results, which can then be read and converted into the HXMS format.

biophysics↗

bioRxiv2025

A newly identified Duck orbivirus as the etiological agent of egg-production decline in Chinese breeder ducks

The genus Orbivirus comprises double-stranded RNA viruses, many of which are transmitted by arthropods and cause febrile, hemorrhagic, or reproductive diseases in animals. In recent years, unexplained outbreaks characterized by marked egg-production decline have occurred in breeder duck flocks in China. To identify the causative agent, samples were collected from affected farms in multiple regions across China between 2022 and 2025. Representative samples were analyzed by high-throughput sequencing, followed by phylogenetic analysis, transmission electron microscopy, host cell infection assays, and animal challenge experiments. A previously unreported Orbivirus, designated Duck orbivirus (DORV), was identified. Epidemiological investigation indicated that outbreaks mainly affected breeder ducks during the peak laying period (21-55 weeks of age), with higher incidence in summer and autumn. The DORV genome contains ten double-stranded RNA segments encoding seven structural and five nonstructural proteins. Phylogenetic analysis showed that DORV is closely related to Parrys Lagoon virus (PLV) and Corriparta virus (CORV). Electron microscopy revealed icosahedral particles 30-40 nm in diameter. DORV replicated efficiently in DEF, DEL and C6/36 cells but not in MDCK and LMH cells, suggesting a restricted host range with potential arthropod-borne features. Infection of breeder ducks with DORV-SD01 reduced egg production by 30-40% and caused lesions in the ovary, liver, and spleen. This study describes a novel duck-origin Orbivirus linked to egg-production decline, offering new etiological insights and extending the host spectrum of Orbivirus. The results provide a foundation for future surveillance and prevention of emerging waterfowl diseases.

microbiology↗

bioRxiv2025

Commonality and Variability in Functional Networks In Young Children Under 5 Years Old

Functional brain networks support human cognition, yet how individualized network architecture emerges in early childhood remains poorly understood. Averaging across participants can obscure age-specific organization and person-to-person differences, particularly in slowly developing association cortices. We developed an age-appropriate functional reference that captured common structure across toddlers without averaging away individual variability, enabling estimation of each childs networks from resting-state fMRI. Across cohorts of 8-60-month-old children, we found individualized network organization--including finer-scale subdivisions and emerging language lateralization--well before age five. Network layouts showed longitudinal stability, with greater consistency in sensory than association regions. Within-network connectivity was stronger and explained age-related variance when networks were defined using individualized rather than group-consensus topography. Left-lateralization of language networks tracked age-normalized verbal ability, linking early functional architecture to emerging cognition. These findings show that behaviorally relevant brain networks arise far earlier than previously recognized, providing a foundation for studying typical development and early biomarkers.

neuroscience↗

bioRxiv2025

From Dataset Curation to Unified Evaluation: Revisiting Structure Prediction Benchmarks with PXMeter

Recent advances in deep learning have significantly improved the accuracy of structure prediction for biomolecular complexes; however, robust evaluation of these models remains a major challenge. We introduce PXMeter, an open-source toolkit that support consistent and reproducible evaluation of diverse predictive models across a broad spectrum of biological complex structures. PXMeter provides a unified and reproducible benchmarking framework, offering valuable insights to support the ongoing improvement of structure prediction methods. We also present a high-quality benchmark dataset curated from recently deposited structures in the Protein Data Bank (PDB). These entries are manually reviewed to exclude non-biological interactions, ensuring reliable evaluation. Using these resources, we conducted a comprehensive benchmark of several structure prediction models, namely Chai-1, Boltz-1, and Protenix. Our benchmarking results demonstrate the advancements achieved by deep learning models, while also identifying ongoing challenges--especially in modeling protein-protein and protein-RNA interactions. Project Pagehttps://github.com/bytedance/PXMeter

bioinformatics↗

bioRxiv2025

Scar-associated endothelial-stellate cellular crosstalk drives fibrosis resolution in MASH

Fibrosis, or scarring, can affect many organs including liver, lung, heart, kidney, intestines etc. and is responsible for [~]40% of mortality in the industrialized world. Compared to other organs, fibrosis in the liver typically resolves when the source of injury is extinguished. Elucidating the molecular mechanisms that underlie spontaneous fibrosis resolution in the liver may lead to novel antifibrotic strategies for all organs. In this study we established a robust mouse model of fibrosis regression in MASH (Metabolic dysfunction-Associated Steatohepatitis), a highly prevalent chronic liver diseases worldwide, and performed single cell and in situ molecular profiling of the liver to define novel drivers of fibrosis regression. As fibrosis regressed, we detected a reduction of inflammatory cells and an expansion of endothelial cells. Prediction of cell-cell communication using the Calligraphy pipeline identified a Wnt9b-Sfrp2 crosstalk that emerges as fibrosis resolved in our model. To establish the Wnt9b-Sfrp2 crosstalk as a driver of fibrosis resolution we treated mice with recombinant Sfrp2, which slowed spontaneous fibrosis regression compared to vehicle treated mice. From our single cell datasets we identified a subset of endothelial cells, termed "Endo4", as the source of Wnt9b. Immunostaining of the Endo4 marker VWF using tissue clearing and 3D imaging revealed VWF+ vasculature enveloped by activated hepatic stellate cells (HSCs) that penetrated deep into the fibrotic septa, establishing Endo4 as de facto scar-associated endothelial cells and providing a structural basis of their cellular crosstalk with HSCs. Finally, using a recently developed in situ protease activity screen, prominent serine protease activity co-localized with both scar-associated Endo4 cells and HSCs. In summary, we uncovered an WNT-dependent endo-stellate crosstalk within the fibrotic niche as a novel regulatory node underlying murine MASH fibrosis regression, and a promising therapeutic target.

molecular biology↗

bioRxiv2025

Distinct sex differences in the production of steroids and neuropeptides in the adult zebrafish brain-pituitary gonadal axis

Zebrafish are increasingly used as experimental models in studies of human disease, environmental toxicology, and reproductive biology. However, sex differences in hormone production are rarely examined, despite evidence from gene mutation studies indicating differential effects in females and males. The emerging reproductive peptide secretoneurin (SN) has not been quantitatively compared between sexes in any species. Here, we employed a newly developed extraction and LC-MS method to simultaneously quantify and compare levels of five steroids and thirteen peptides in brain, pituitary, and gonads. As expected, testosterone (T) and 11-ketotestosterone (11-KT) were higher in male tissues, while estrone (E1) and estradiol (E2) were elevated in the female pituitary/ovary and brain, respectively. Estriol (E3) was more abundant in testes. Gonadotropin-releasing hormones Gnrh2 and Gnrh3 were notably higher in testes. Oxytocin (isotocin) and vasopressin (vasotocin) were elevated in the female brain and in testes. Kisspeptins 1 and 2 also showed higher levels in testes. Similarly, SNa and SNb were more abundant in the female brain and pituitary, and markedly higher in testes than ovaries. Several smaller SN fragments were detected at low levels, with patterns suggesting sex-specific enzymatic processing. These findings reveal pronounced sex differences in both classical and emerging reproductive hormones and identify the SN peptide family as a new component of the brain-pituitary-gonadal axis. This dataset provides a foundation for future studies on sexually differentiated neuropeptide production and function across tissues.

physiology↗

bioRxiv2025

Human myelocyte and metamyelocyte-stage neutrophils suppress tumor immunity and promote cancer progression

Tumor-infiltrating neutrophils (TINs) are highly heterogeneous and mostly immunosuppressive in the tumor immune microenvironment (TIME). Current biomarkers of TINs and treatment strategies targeting TINs have not yielded optimal responses in patients across cancer types. Here, we separated human and mouse neutrophils into three developmental stages, including promyelocyte (PM), myelocyte & metamyelocyte (MC & MM), and banded & segmented (BD & SC) neutrophils. Based on this separation, we observed the predominance of human but not mouse MC & MM-stage neutrophils in bone marrow (BM), which exhibit potent immunosuppressive and tumor-promoting properties. MCs & MMs also occupy the majority of TINs among patients with 17 cancer types. Moreover, through the creation of a NCG-Gfi1-/- human immune system (HIS) mouse model, which supports efficient reconstitution of human TIN, we found a significant increase of BM MCs & MMs in tumor-bearing mice. By comparing the scRNA-seq analysis results of human neutrophils from both BM and tumors, we found CD63 and Galectin-3 distinguish MC & MM from neutrophil populations in cancer patients. Furthermore, we proposed a strategy with Flt3L to specifically induce the trans-differentiation of MCs & MMs into monocytic cells, and trigger tumor control in NCG-Gfi1-/- HIS mice. Thus, our findings establish an essential role of human MC & MM-stage neutrophils in promoting cancer progression, and suggest their potential as targets for developing potential biomarkers and immunotherapies for cancer.

immunology↗

bioRxiv2025

Untangling overlapping barcodes in spatial genomics data

Difficulty in resolving spatially overlapping barcodes is a major bottleneck for imaging-based spatial genomics methods. Here, we present an approach for untangling spatially overlapping barcodes by using strong encoding and global optimization to reduce spurious solutions resulting from recombinations of barcode signals. We demonstrate experimentally that cellular regions with average local densities of 127 barcodes per {micro}m2 can be decoded with an estimated FDR of less than 4%, enabling a new type of super-resolution microscopy by coding.

bioengineering↗

bioRxiv2025

Optimization of Genome-wide CRISPR Screens using dual-guide RNA Infection with Cas9 Electroporation (DICE)

Single-guide RNA lentiviral infection with Cas9 protein electroporation (SLICE) enables CRISPR screening in primary cell types that require transient Cas9 expression yet is limited by scalability and robustness. Here, we introduce dual-guide RNA infection with Cas9 electroporation (DICE), which expresses two guides from the same lentiviral construct that target the same gene. In genome-wide screens, DICE outperformed SLICE in defining essential genes and modulators of PD-L1 expression in IFN gamma activated THP1 cells. Collectively, these data demonstrate that DICE can be utilized for reduced-scale CRISPR screens in cell types with transient Cas9 protein expression without sacrificing screening quality.

genomics↗

bioRxiv2025

Development a computational framework for risk evaluation of emerging flu viruses based on genomic sequences and AI

The recent resurgence of highly pathogenic avian influenza H5N1 viruses in North America and Europe has heightened global concerns regarding potential influenza pandemics. Despite significant progress in the surveillance and prevention of emerging influenza viruses, effective tools for rapid and accurate risk assessment remain limited. Here, we present FluRisk, an innovative computational framework that integrates viral genomic data with artificial intelligence (AI) to enable rapid and comprehensive risk evaluation of emerging influenza strains. FluRisk incorporates a curated database of over 1,000 experimentally validated molecular markers linked to key viral phenotypes, including mammalian adaptation, mammalian virulence, mammalian transmission, human receptor-binding preference, and antiviral drug resistance. Leveraging these markers, we developed three state-of-the-art machine learning models to predict human adaptation, mammalian virulence, and human receptor-binding potential, all of which demonstrated superior performance compared to traditional approaches such as BLAST, prior models, and baseline classifiers. In addition, a reference-based method was implemented to provide preliminary estimates of human transmissibility and resistance to six commonly used antiviral drugs. To facilitate broad accessibility and practical application, we developed a user-friendly web server that integrates both the molecular marker atlas and predictive tools for influenza virus phenotyping (available at: http://www.computationalbiology.cn/FluRisk/#/). This computational platform offers a valuable resource for the timely risk assessment of emerging influenza viruses and supports global influenza surveillance efforts.

bioinformatics↗

bioRxiv2025

Validation and pre-analytical considerations for processing cerebrospinal fluid samples on a high-throughput proximity extension assay platform

BackgroundAnalysis of cerebrospinal fluid (CSF) facilitates the understanding of brain-specific molecular changes that may associate with disease progression. Proximity extension assays (PEA) have been deployed in several CSF studies, however the validation of the assay and impact of freeze-thaw cycles on the protein signal has not been documented. We sought to (1) validate the assay on the PEA platform and (2) evaluate the effect of freeze-thaw cycles on the detectability of analytes on the PEA platform. ResultsWe have validated the PEA with Next Generation Sequencing (NGS) readout assay and report on the detectability and coefficient of variation observed in CSF samples. We have also evaluated proteomic signals with a minimum of 3 and a maximum of 9 freeze thaw cycles and detected very minimal change in signal with increasing cycle number. ConclusionOur study is the first to validate PEA using NGS readout platform with CSF samples. We report lower protein detection rates and higher variability in the expansion panels compared to the original 4 panels, with acceptable variation above detectability threshold. In addition, our work demonstrates that the proteomic signal is robust and continues to be stable across multiple freeze thaw cycles. This is highly impactful to the processing and analysis of clinical samples and facilitates the investigation of samples with variable pre-analytical conditions.

neuroscience↗

bioRxiv2025

Conserved energetic changes drive function in an ancient protein fold

While the ongoing revolution in structural biology offers an unprecedented understanding of the relationship between protein structure and function, it also confirms a puzzling, widely applicable principle: protein domains with highly conserved three-dimensional folds can perform radically disparate biochemical functions. To gain insight to this fundamental structural enigma, we mapped the energetic landscapes of a family of bacterial transcription factors and their anciently diverged structural homologs, the periplasmic binding proteins. Using hydrogen exchange/mass spectrometry, bioinformatics, X-ray crystallography, and molecular dynamics, we uncovered an unexpected contrast: despite binding the same sugars, the two families evolved unique "energetic blueprints" to support their distinct functional requirements. To test if differences in energetic ensembles have functional consequences, we rationally redesigned the protein fold for tunable ligand-driven transcriptional responses. Strikingly, energy-driven protein engineering produced synthetic transcription factors with the theoretically anticipated ligand-induced transcriptional outputs. Thus, decoding energetic blueprints among conserved protein folds provides a novel explanation for diverse functional adaptations, paves an alternative roadmap for protein design, and offers a new approach for engineering challenging drug targets.

biophysics↗

bioRxiv2025

REPOP: bacterial population quantification from platecounts

Bacterial counts from native environments, such as soil or the animal gut, often show substantial variability across replicate samples. This heterogeneity is typically attributed to genetic or environmental factors. A common approach to estimating bacterial populations involves successive dilution and plating, followed by multiplying colony counts by dilution factors. This method, however, overestimates the heterogeneity in bacterial population because it conflates the inherent uncertainty in drawing a subsample from the total population with the uncertainty in the sample arising from biological origins. In other words, this approach may obscure features that may otherwise be present in the data hinting at the presence of genuine subpopulations. For example, in plate counting applied to C. elegans gut microbiota, observed multimodality is often interpreted as large host-to-host variance, while the randomness introduced by measurement is frequently ignored. To explicitly account for the uncertainty introduced by dilution and plating randomness, we introduce REPOP, a PyTorch-based library to REconstruct POpulations from Plates within a Bayesian framework. Beyond simple cases, REPOP addresses more complex scenarios, including multimodal populations and correcting the mathematically subtle, but experimentally relevant, bias introduced by excluding plates deemed too crowded to distinguish individual colonies. We demonstrate REPOPs ability to resolve distinct population peaks otherwise obscured by standard multiplication methods. Applications to both simulated and experimental datasets, including bacterial samples of different concentrations and ones from the gut microbiota of C. elegans, show that REPOP accurately recovers the underlying multimodality by properly accounting for error propagation, where naive multiplication fails. REPOP is available on GitHub: https://github.com/LabPresse/REPOP.

microbiology↗

bioRxiv2025

The neurocomputational mechanisms of hierarchical linguistic predictions during narrative comprehension

Language comprehension requires a listener to predict the upcoming inputs of linguistic units with multiple timescales based on previous contexts, but how the prediction process is hierarchically represented and implemented in the human brain remains unclear. Combining the natural language processing (NLP) approach and functional Magnetic Resonance Imaging (fMRI) in a narrative comprehension task, we first applied the group-based general linear model (gGLM) to identify the neural underpinnings associated with the language prediction on word and sentence. Our results revealed a cortical architecture supporting the prediction, extending from the superior temporal cortices to the regions in the default mode network. Then, we investigated how these adjacent levels interact with each other by testing two rival hypotheses: the continuous updating hypothesis posits that the higher level of the representational hierarchy is continuously updated as inputs unfold over time, while the sparse updating hypothesis states that the higher level is only updated at the end of their preferred timescales of linguistic units. By conducting computational modeling and autocorrelation analysis, we found the sparse model outperformed the continuous model and the updating might occur at the sentence boundaries. Together, our results extend the linguistic prediction from the small timescales such as words to large timescales such as sentences, providing novel insights into the neurocomputational mechanisms of information updating within the linguistic prediction hierarchy.

neuroscience↗

bioRxiv2025

Reconstructing signaling histories of single cells via perturbation screens and transfer learning

Manipulating the signaling environment is an effective approach to alter cellular states for broad-ranging applications, from engineering tissues to treating diseases. Such manipulation requires knowing the signaling states and histories of the cells in situ, for which high-throughput discovery methods are lacking. Here, we present an integrated experimental-computational framework that learns signaling response signatures from a high-throughput in vitro perturbation atlas and infers combinatorial signaling activities in in vivo cell types with high accuracy and temporal resolution. Specifically, we generated signaling perturbation atlas across diverse cell types/states through multiplexed sequential combinatorial screens on human pluripotent stem cells. Using the atlas to train IRIS, a neural network-based model, and predicting on mouse embryo scRNAseq atlas, we discovered global features of combinatorial signaling code usage over time, identified biologically meaningful heterogeneity of signaling states within each cell type, and reconstructed signaling histories along diverse cell lineages. We further demonstrated that IRIS greatly accelerates the optimization of stem cell differentiation protocols by drastically reducing the combinatorial space that needs to be tested. This framework leads to the revelation that different cell types share robust signal response signatures, and provides a scalable solution for mapping complex signaling interactions in vivo to guide targeted interventions.

developmental biology↗