Biology subjects

Han, W.

Publications and source records attributed to Han, W..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

AlphaGEM Enables Precise Genome-Scale Metabolic Modelling by Integrating Protein Structure Alignment with deep-learning-based Dark Metabolism Mining

Constructing high-quality genome-scale metabolic models (GEMs) for less-studied species remains challenging. To address this, we developed AlphaGEM, a versatile toolbox leveraging proteome-scale structural alignment and deep-learning-based predictions for efficient genomic mining to generate GEMs ready for applications. Our findings show that the structural alignment or protein-language-model-based prediction (i.e., PLMSearch), could identify more homologous protein relationships than sequence-blast-based alignment, contributing to the accurate profiling of metabolism from target organisms. Additionally, AlphaGEM encompasses an ensemble procedure empowered by multiple deep learning toolboxes to effectively mine the dark metabolic functions encoded by nonhomologous proteins, significantly expanding species-specific metabolic networks. We demonstrate AlphaGEMs accuracy by building GEMs for eukaryotes (e.g., S. pombe, C. albicans) and prokaryotes (e.g., K. pneumoniae, B. subtilis), achieving predictions comparable to manually curated models while outperforming existing tools. AlphaGEM also successfully reconstructs GEMs for M. musculus and C. griseus, showcasing its great potential for uncovering dark metabolism in complex mammals. Lastly, we demonstrate that AlphaGEM could facilitate the automatic GEMs reconstruction for 332 distinct yeast species with high prediction fidelity. In conclusion, AlphaGEM provides unprecedented opportunities for the precise, rapid construction of GEMs across diverse domains, which sets a solid foundation for universal functional analysis of non-model organisms having genome sequences available.

bioinformatics↗

bioRxiv2025

Functional Comparison to Ezh2 Reveals PRC2-Independent Functions of Jarid2 in Hematopoietic Stem Cell Lineage Commitment

Previous studies showed the Polycomb Repressive Complex 2 (PRC2) co-factor Jarid2 represses self-renewal transcriptional networks in mouse multipotent progenitor cells (MPPs). But only a fraction of de-repressed HSC-specific genes were associated with loss of H3K27me3, implying Jarid2 may have non-canonical (PRC2-indpendent) in hematopoiesis. Here we sought to delineate any PRC2-independnent functions by comparing stem and progenitor cells genetically deficient for either Jarid2 or Ezh2 (enzymatic component of PRC2). Loss of Ezh2 increased myeloid differentiation in transplantation assays. In contrast, loss of Jarid2 enhanced T-cell output. Single cell transcriptomics showed while loss of Jarid2 had minimal impact across progenitor populations, loss of Ezh2 led to accumulation of lymphoid-biased MPP4 cells and B-cell progenitors in the bone marrow. Functional assays confirmed a differentiation block at the pre-pro B-cell stage. The maturational arrest of Ezh2-deficient B-cell progenitors contrasts with increased T-cell output from loss of Jarid2, suggesting Jarid2 has non-canonical functions in hematopoiesis.

genetics↗

bioRxiv2025

Estimating Actual Striking Forces Using Attenuation Properties of Taekwondo Protectors

BackgroundProtectors attenuate impact forces significantly, however, their quantitative attenuation characteristics under varying conditions remain poorly understood. Accurate measurement of striking forces is essential for evaluation. However, direct measurement of athletes striking forces presents significant practical and safety challenges, necessitating the development of indirect estimation approaches. MethodA controllable pendulum-based impact testing apparatus was developed to evaluate force attenuation characteristics of Taekwondo body protectors. The system delivered repeatable impacts across varying magnitudes and contact durations to a mannequin equipped with a protector. Simultaneous measurements of input and transmitted forces through the protector enabled direct quantification of attenuation ratios under controlled conditions. To demonstrate practical application, five Taekwondo athletes performed standardized kicks on the instrumented protector setup, allowing estimation of their actual striking forces using the derived attenuation factor. ResultsForce attenuation ratios demonstrated high consistency across impact magnitudes ranging from 300 to 5800 Newtons, with impact duration variations showing minimal influence on attenuation performance. Linear attenuation relationships were established between input and transmitted forces (R2[≥] 0.987), enabling derivation of a predictive attenuation factor. Validation testing with the pendulum system showed that the estimated input force trajectories achieved an nRMSE of 3.5%. Averaged across the entire force range, the MAPE was 4.7% for maximum force, 1.4% for impact duration, and 3.5% for impulse. Applying the attenuation factor to standardized kicks performed by five Taekwondo athletes yielded estimated maximum striking forces ranging from approximately 1300 to 1800 Newtons, demonstrating the methods capability to quantify actual forces that would otherwise be difficult to measure directly. DiscussionThis study establishes the first quantitative characterization of impact attenuation in Taekwondo body protectors, providing a validated attenuation factor for estimating striking forces from transmitted forces. These findings demonstrate a practical methodology for quantifying striking forces that are otherwise difficult to measure directly. Author summary

bioengineering↗

bioRxiv2025

Treating Myotonic Dystrophy with artificial RNA endonucleases to specifically degrade toxic RNA expansions

Myotonic dystrophy type 1 (DM1), the most common autosomal dominant muscular disorder, is driven by expanded CUG repeats in the 3' UTR of DMPK gene, which sequester RNA-binding proteins and cause aberrant alternative splicing. Despite extensive study, no effective treatment exists, and current care remains limited to symptomatic management. Using a customized dual-color reporter carrying RNA repeats in different lengths, we engineered a new class of programmable artificial enzymes (artificial RNA cleavers, or ARCs) that preferentially degrade the pathogenic repeats both in cells and in a mouse model. We further examined the in vivo efficacy of ARCs by systematically delivering ARCs into the DM1 mouse model with muscle-specific adeno-associated viruses, and validated that ARCs can rescue the muscular pathology and movement ability. The treatment by ARCs showed sustained efficacy, low immunogenicity, and minimal off-target effects in animal model, which are key concerns in clinical translation of gene therapies. Collectively, these findings establish ARCs as a potent and precise RNA-targeting platform with strong clinical potential for the treatment of DM1 and other repeat expansion disorders.

synthetic biology↗

bioRxiv2025

Genome-wide CRISPR screens identify PTGES3 as a novel AR modulator

The androgen receptor (AR) is a critical driver of prostate cancer (PCa). To study regulators of AR protein levels and oncogenic activity, we created the first live cell quantitative endogenous AR fluorescent reporters. Leveraging this novel AR reporter, we performed genome-scale CRISPRi flow cytometry sorting screens to systematically identify genes that modulate AR protein levels. We identified and validated known AR protein regulators including HOXB13 and GATA2 and also unexpected top hits including PTGES3, a poorly characterized gene in PCa. PTGES3 repression resulted in loss of AR protein, cell cycle arrest, and cell death in AR-driven PCa models. PTGES3 is not a commonly essential gene, and our data nominate it as a prime PCa therapeutic target. Clinically, analysis of PCa data demonstrate that PTGES3 expression is associated with AR-directed therapy resistance. Mechanistically, we show PTGES3 binds directly to AR, forms a protein complex with AR in the nucleus, regulates AR protein stability in vitro and in vivo and modulates AR function in the nucleus at AR target genes. PTGES3 represents a novel therapeutic target for overcoming known mechanisms of resistance to existing AR-directed therapies in PCa.

cancer biology↗

bioRxiv2025

Adaptive Evolution of Gene Regulatory Networks in Mammalian Neocortical Neurons

Mammals have evolved a plethora of adaptations that have enabled them to thrive in diverse environments. Among the most significant is the emergence of a more complex brain, exemplified by the dramatic transformation of the dorsal cortex from a single layer of excitatory projection neurons (ExNs) in ancestors to a multilayered cerebral neocortex enriched with diverse intratelencephalic (IT) and extratelencephalic (ET) ExN subtypes. These ExNs established specialized projection systems, such as the corticospinal tract and corpus callosum, enhancing brain connectivity and functionality. However, the evolutionary mechanisms underlying these mammalian-specific adaptations remain elusive. By comparing the landscape of gene expression and cis-regulatory elements (CREs) in mouse ExN subtypes and by cross-species examination of mammalian and non-mammalian CREs, we identified mammalian-specific CREs and expression patterns. The mammalian-specific CREs include a subset bound by ZBTB18 that are associated with genes defining IT and ET subtypes and connectivity. Both ZBTB18 and these target genes have previously been implicated in intellectual disability and autism. Deletion of Zbtb18 in mouse ExNs dysregulated target gene expression, reduced molecular diversity, diminished corticospinal and callosal projections, and increased intrahemispheric cortico-cortical association projections to the prefrontal cortex, resembling features of non-mammalian dorsal pallium. Interestingly, ZBTB18 binding motifs are highly enriched in callosally projecting IT-biased CREs, where they show higher conservation specifically in mammals. This study uncovers critical components and mammalian-specific evolutionary adaptations within a regulatory node essential for neocortical ExN identity and connectivity, with implications for neurodevelopmental and neuropsychiatric disorders.

neuroscience↗

bioRxiv2025

ReCellTy: Domain-specific knowledge graph retrieval-augmented LLMs workflow for single-cell annotation

To enable precise and fully automated cell type annotation with large language models (LLMs), we developed a graph-structured feature-marker database to retrieve entities linked to differential genes for cell reconstruction. We further designed a multi-task workflow to optimize the annotation process. Compared to general-purpose LLMs, our method improves human evaluation scores by up to 0.21 and semantic similarity by 6.1% across 11 tissue types, while more closely aligning with the cognitive logic of manual annotation.

bioinformatics↗

bioRxiv2025

Novel insights into the genetic architecture and mechanisms of host/microbiome interactions from a multi-cohort analysis of outbred laboratory rats

The intestinal microbiome influences health and disease. Its composition is affected by host genetics and environmental exposures. Understanding host genetic effects is critical but challenging in humans, due to the difficulty of detecting, mapping and interpreting them. To address this, we analysed host genetic effects in four cohorts of outbred laboratory rats exposed to distinct but controlled environments. We found that polygenic host genetic effects were consistent across environments. We identified three replicated microbiome-associated loci, one of which involved a sialyltransferase gene and Paraprevotella. We found a similar association in a human cohort, between ST6GAL1 and Paraprevotella, both of which have been linked with immune and infectious diseases. Moreover, we found evidence of indirect genetic effects on microbiome phenotypes, which substantially increased their total genetic variance. Finally, we identified a novel mechanism whereby indirect genetic effects can contribute to "missing heritability".

genetics↗

bioRxiv2025

Deep contrastive feature compression with classical machine learning enables ligand discovery through efficient triage of large chemical libraries

Improving in silico compound-protein interaction (CPI) predictability is critical for productive drug discovery. Current deep learning approaches largely rely on end-to-end models trained on limited labeled CPI datasets, overlooking the representational power of large-scale biochemical foundation models. We present COMRADE (Contrastive Multirepresentation Accelerated Docking Engine), a hybrid virtual screening framework that accelerates docking by triaging compounds using CE-Screen (Contrastive Embedding-Screen). CE-Screen leverages seven high-dimensional pretrained representations - including those from protein language models and molecular transformers, along with an original physics-based interaction potential encoding - for rapid first-pass screening [~]100x faster than docking. Its contrastive compression neural network maps these inputs onto a single compact, discriminative representation optimized for CPI prediction via a lightweight ensemble classifier. CE-Screen outperforms state-of-the-art end-to-end models by up to 111.11% on retrospective benchmarks and is successfully used to triage [~]10.8 million compounds against five targets, yielding novel hits for each one - including a new scaffold for the branched-chain ketoacid dehydrogenase kinase (BCKDK), an understudied yet high-value target in metabolic disease and oncology.

bioinformatics↗

bioRxiv2025

Sensitive Transcriptomics and Genotyping reveals function of genetic variants in immunity

Advancements in genomics have revolutionized human genetics by defining the genetic architecture of disease. However, identifying causal variants and their mechanisms of action remains a challenge in translating genetics into therapeutic interventions. Here, we developed Sensitive Transcriptomics And Genotyping by sequencing (STAG-seq), a high-throughput platform designed to define mechanistic genotype-phenotype relationships through simultaneous single-cell measurements of genomic DNA and RNA transcripts. Combined with base-editing, STAG-seq enables functionalization of variants in relevant cellular contexts. We demonstrate the applicability of this approach in several settings. First, we screened genetic perturbations to identify monoallelic and biallelic variant effects in primary human macrophages treated with innate immune stimuli. Next, we phenotyped clinically relevant missense variants associated with immunodeficiency and autoimmunity. Finally, we defined a noncoding variant in a pleiotropic autoimmunity locus that governs TNRC18 expression in primary T cells. STAG-seq thus enables variant phenotyping at scale to advance functional genomics and disease biology.

genomics↗

bioRxiv2025

Viral Infection Detector: Ensemble Learning for Predicting Viral Infection in Single-cell Transcriptomics of Virus-Induced Cancers

Oncoviruses cause around 15-20% of all human cancers. As 3 end-based single-cell RNA-sequencing (scRNA-seq) methods become more popular, the relationship between viral infection and tumor progression at the transcriptomic level has been increasingly studied at single-cell resolution. However, the identification of infected cells is challenging as viral reads from infected cells may not be captured, leading to high false negative rates. To recover these infected cells that have been missed, we developed VID, a stacked ensemble machine-learning model that predicts virus-infected cells in scRNA-seq datasets. Using VID, we uncovered biologically meaningful differences between infected and uninfected cells across different oncovirus-mediated cancers and different cell types. Our VID tool provides a user-friendly way to study changes induced by viral infections in scRNA-seq datasets.

bioinformatics↗

bioRxiv2025

A two-step clockwork mechanism opens a proteo-lipidic pore in PIEZO2 channels

Mechanosensitive PIEZO channels are thought to open their pore through tension-induced flattening of large transmembrane arm domains. Yet, the structural basis of this activation remains unclear. Here, we uncover the conformational coupling between arm flattening and pore opening in PIEZO2 by capturing protein motions across length scales using hybrid-resolution molecular dynamics simulations. Sampling multiple microsecond-long trajectories under physiological activation tension show that arm flattening correlates with anticlockwise rotation of the pore domain and with clockwise twisting of inner pore helices, enabling dilation and hydration of a transmembrane pore gate. These clockwork motions enable PIEZO2 to populate two open states with distinct conductance depending on applied membrane tension, in agreement with single-channel electrophysiology. Pore opening is accompanied by the separation of pore helices, creating interhelical gaps which become filled with lipids, resulting in the fully conducting pore being walled by both lipids and protein. The fully open PIEZO2 state recapitulates minimal pore size, conductance, ion selectivity, and outward rectification of chloride currents observed electrophysiologically. This work reveals how tension-induced large-scale rearrangements of the PIEZO2 arms funnel into subtle and dynamic gating motions, providing invaluable structural insights for future structure-function and drug discovery studies.

biophysics↗

bioRxiv2025

Effective in vivo RNA base editing via engineered cytidine deaminase APOBECs fused with PUF proteins

Base editing stands at the forefront of genetic engineering, heralding precise genetic modifications with broad implications. While CRISPR-based DNA and RNA base editing systems capitalize on sgRNA-guided specificity and diverse deaminase functionalities, the pursuit of efficient C-to-U RNA editing has been hampered by the inherent constraints of cytidine deaminases. Here, we report an RNA base editing platform by refining cytidine deaminases, termed professional APOBECs (ProAPOBECs), through systematic enhancements and AI-driven protein engineering. ProAPOBECs demonstrate unprecedented catalytic versatility, particularly fused with RNA-recognizing Pumilio and FBF (PUF) proteins. We present the first effective use of RNA base editing in the brain with ProAPOBECs in Mef2c mutant mice, a model for autism. The AAV-mediated RNA base editing via ProAPOBEC not only corrects genetic mutation in mRNAs but significantly alleviates autistic-like behaviors in the mice. This work introduces a pioneering collection of RNA base editing instruments, emphasizing their therapeutic potential in combatting genetic disorders.

molecular biology↗

bioRxiv2025

SRS microscopy identifies inhibition of vitellogenesis as a mediator of lifespan extension by caloric restriction in C. elegans

The molecular mechanisms of aging are not fully understood. Here, we used label-free Stimulated Raman scattering (SRS) microscopy to investigate changes in proteins and lipids throughout the lifespan of C. elegans. We observed a dramatic buildup of proteins within the body cavity or pseudocoelom of aged adults that was blunted by interventions that extend lifespan: caloric restriction (CR) and the reduced insulin/insulin-like growth factor signaling (IIS) pathway. Using a combination of microscopy, proteomic analysis, and validation with mutant strains, we identified vitellogenins as the key molecular components of the protein buildup in the pseudocoelom. Vitellogenins shuttle nutrients from intestine to embryos and are homologous to human apolipoprotein B, the causal driver of cardiovascular disease. We then showed that CR and knockdown of vitellogenins both extend lifespan by >60%, but their combination has no additional effect on lifespan, suggesting that CR extends the lifespan of C. elegans in part by inhibiting vitellogenesis. The extensive dataset of more than 12,000 images stitched into over 350 whole-animal SRS images of C. elegans at different ages and subjected to different longevity intervention will be a valuable resource for researchers interested in aging.

cell biology↗

bioRxiv2025

Optogenetic control of Protein Kinase C-epsilon activity reveals its intrinsic signaling properties with spatiotemporal resolution

The regulation of PKC epsilon (PKC{varepsilon}) and its downstream effects is still not fully understood, making it challenging to develop targeted therapies or interventions. A more precise tool that enables spatiotemporal control of PKC{varepsilon} activity is thus required. Here, we describe a photo-activatable optogenetic PKC{varepsilon} probe (Opto-PKC{varepsilon}) consisting of an engineered PKC{varepsilon} catalytic domain and a blue-light inducible dimerization domain. Molecular dynamics and AlphaFold simulations enable rationalization of the dark-light activity of the optogenetic probe. We first characterize the binding partners of Opto-PKC{varepsilon}, which are similar to those of PKC{varepsilon}. Subsequent validation of the Opto-PKC{varepsilon} tool is performed with phosphoproteome analysis, which reveals that only PKC{varepsilon} substrates are phosphorylated upon light activation. Opto-PKC{varepsilon} could be engineered for recruitment to specific subcellular locations. Activation of Opto-PKC{varepsilon} in isolated hepatocytes reveals its sustained activation at the plasma membrane is required for its phosphorylation of the insulin receptor at Thr1160. In addition, Opto-PKC{varepsilon} recruitment to the mitochondria results in its lowering of the spare respiratory capacity through phosphorylation of complex I NDUFS4. These results demonstrate that Opto-PKC{varepsilon} may have broad applications for the studies of PKC{varepsilon} signaling with high specificity and spatiotemporal resolution. SummaryWe have developed a photo-activatable optogenetic PKC{varepsilon} probe, which demonstrates differential activity in light versus dark. The tool is subsequently validated with protein association studies and phosphoproteome analysis. It enables dissection of signaling events arising from its activation at defined subcellular locations. For instance, sustained activation of PKC{varepsilon} at the plasma membrane is required for its phosphorylation of the insulin receptor at Thr1160.

molecular biology↗

bioRxiv2024

Base-modified nucleotides mediate immune signaling in bacteria

Signaling from pathogen sensing to effector activation is a fundamental principle of cellular immunity. While cyclic (oligo)nucleotides have emerged as key signaling molecules, the existence of other messengers remains largely unexplored. Here, we reveal a bacterial anti-phage system that mediates immune signaling through nucleobase modification. Immunity is triggered by phage nucleotide kinases, which, combined with the system-encoded adenosine deaminase, produce deoxyinosine 5-triphosphate (dITP) as immune messengers. The dITP signal activates downstream effector to mediate cellular NAD+ depletion, resulting in population-level defense through the death of infected cells. To counteract immune signaling, phages deploy specialized enzymes that deplete cellular dAMP, the precursor of dITP messengers. Our findings uncover a nucleobase modification-based anti-phage signaling pathway, establishing noncanonical nucleotides as a new type of immune messengers in bacteria.

microbiology↗

bioRxiv2024

Integrative analysis of transcriptome and metabolism reveals functional roles of redox homeostasis in low light and salt combined stress in Leymus chinensis

Soil salinity (Salt) is one of the major limiting factors of Leymus chinensis (named sheepgrass) performance and alters their primary metabolic process and gene expression. Salt stress could accelerate inhibitive effects concomitant with low light (LL-Salt). However, little is known about physiological and molecular mechanisms under such LL-Salt in sheepgrass. This study aims to uncover the key reprogrammed metabolic pathways induced by LL-Salt through an integrated analysis of transcriptome and metabolism. Plant seedlings were exposed to six combinations of light intensity (moderate light, ML, and LL) and NaCl concentrations (0, 50, and 200mM) for 20 days. Results suggest that the growth of sheepgrass seedlings was dramatically inhibited with severe lodging phenotype, especially when exposed to LL-Salt combined conditions. Activities of an antioxidant enzyme, catalase, were significantly increased in LL but significantly decreased in salt stress, leading to a reverse pattern for H2O2 in LL and salt stress. Transcriptome analysis reveals 4921 downregulated differentially expressed genes (DEGs) with carbon metabolism pathways were significantly enriched compared to light response without salt (ML_0 vs LL_0). There are 194 overlapped downregulated DEGs induced by two salt treatments (ML_50 vs ML_0 and ML_200 vs ML_0), where the carbon metabolism pathway was also significantly enriched. In terms of interactive effects of LL-Salt treatments, we found that there are 16 overlapped DEGs, including a phytochrome-interacting transcription factor 4 (PIF4) with upregulation in LL treatment while downregulation in salt treatment, which was validated by qPCR. Metabolism analysis confirmed that in carbon metabolism, the contents of seven differentially abundant metabolites (DAMs), including sucrose, were all downregulated in both LL and salt treatment. In contrast, serine and glycolate in the photorespiration pathway were downregulated in LL while upregulated in salt treatment. A photorespiration regulatory gene, glycolate oxidase expression, was confirmed by qPCR. Collectively, we found that serval antioxidant redox pathways, including photorespiration, GSG/GSSH redox, and ABA signaling, participated in response to LL and salt combined events and highlighted the roles of cellular redox homeostasis in LL-Salt response in sheepgrass.

plant biology↗

bioRxiv2024

Stimulus history, not expectation, drives sensory prediction errors in mammalian cortex

Hierarchical predictive coding (HPC) models have recently flourished in neuroscience1-9. Feedforward and feedback processing are at the heart of HPC models. Previous experimental studies using fMRI, EEG/MEG, and LFP9-11 do not reliably resolve feedback modulation from local computations and feedforward outputs. Here, using open-science8, multi-species, multi-area, high-density12, laminar neurophysiology13, we empirically test whether hierarchical predictive coding is a key component shaping cortical processing of visual stimuli. To isolate visual information processing and eliminate motor/reward confounders9-11, we use a no-report task. Our task leveraged so-called global oddballs (GO) as unpredictable, deviant stimuli that circumvent low-level adaptation. We examined their responses relative to local oddballs (LO) that we habituated into highly predictable priors. Four surprising findings in this dataset challenge many existing hierarchical predictive coding models. First, GO responses were exclusive to higher-order, more cognitive areas rather than early-to-mid visual cortex. Second, inhibitory interneuron-targeted optogenetics in primates and mice and waveform shape analysis in primates revealed no evidence that predictive suppression was implemented via these interneurons. Third, highly predictable LO responses dominated in over 50% of all neurons, including in higher-order cortex, which should have anticipated them, indicating limited evidence for predictive suppression. Lastly, prediction error responses evoked by GOs did not evoke feedforward processing. These results reveal circuit dynamics that govern how prediction shapes visual processing, motivating more neurally constrained predictive processing models.

neuroscience↗