Biology subjects

Feng, W.

Publications and source records attributed to Feng, W..

At least 19 recordsLinked to original sources

bioRxiv2026

Epstein-Barr Virus in HIV-Associated Conjunctival Squamous Cell Carcinoma and Transcriptional Reprogramming of Conjunctival Epithelial Cells

The involvement of Epstein-Barr virus (EBV) in conjunctival squamous cell carcinoma (cSCC) remains incompletely defined. This study integrated analyses of primary tumors with an experimental model of the conjunctival epithelium. Five Nigerian tumors from patients with HIV and three Japanese tumors from patients without HIV were examined by EBV-encoded RNA in situ hybridization (EBER-ISH), EBV DNA PCR, and virome analysis of RNA sequencing data. EBER localized to epithelial tumor cells in four Nigerian specimens and none of the Japanese specimens. EBV DNA and RNA were each detected in three Nigerian tumors, with every EBER-positive tumor supported by at least one additional assay. Human telomerase reverse transcriptase-immortalized conjunctival epithelial cells infected with Akata-GFP-EBV expressed latency-associated and lytic viral genes. Comparison of three infected and three control cultures identified 96 differentially expressed host genes and 23 enriched Hallmark gene sets at a false discovery rate below 0.05. RT-qPCR confirmed increased HES1, CXCL8, CXCL2, and LOX and decreased FGFBP1, CSTA, SAPCD2, and TMEM40 expression. Prominent responses involved inflammatory signaling, epithelial remodeling, metabolism, and cell-cycle regulation. These findings identify tumor-cell-associated EBV in a subset of HIV-associated cSCC and establish an epithelial model for investigating infection-associated host responses.

microbiology↗

bioRxiv2026

Cardiomyocyte Perm1 sustains cardiac bioenergetic reserve and exercise capacity

Objective: Mitochondrial oxidative capacity in striated muscle is governed by the PGC-1/ERR transcriptional axis, which controls not only steady-state respiration but the capacity to expand energy production when demand rises. Whether this reserve function requires dedicated downstream effectors, and whether its loss is detectable at rest, remains unclear. We asked whether Perm1, a striated muscle-specific PGC-1/ERR-induced regulator, is required cell-autonomously for cardiac bioenergetic reserve. Methods: We generated cardiomyocyte-specific Perm1 knockout mice using alphaMHC-nCre, complemented by an adult tamoxifen-inducible alphaMHC-MerCreMer model, and assessed cardiac development, structure, and function longitudinally through 12 months of age in both sexes. We evaluated mitochondrial function by high-resolution respirometry, transmission electron microscopy, and molecular profiling. We tested energetic capacity under load by treadmill exercise with tissue ATP quantification and by beta-adrenergic stimulation of cultured neonatal cardiomyocytes. Results: Perm1 deletion did not perturb embryonic cardiac development, and neither constitutive nor adult-onset deletion impaired cardiac structure, systolic function, or diastolic function through 12 months. Despite this preserved resting phenotype, mitochondrial respiration was reduced by 20 - 25% across complex I- and complex I+II-supported conditions, without changes in citrate synthase activity, mtDNA copy number, or OXPHOS subunit abundance. Instead, the MICOS components IMMT (MIC60) and CHCHD3 (MIC19) were selectively reduced, and mitochondrial subpopulations shifted toward fewer subsarcolemmal and more interfibrillar mitochondria, with unchanged total mitochondrial density. Functionally, Perm1-deficient mice of both sexes showed reduced treadmill capacity at 6 and 12 months and failed to maintain cardiac ATP during exercise at a submaximal workload, while skeletal muscle ATP and plasma lactate were unaffected. This deficit was cardiomyocyte-autonomous: beta-adrenergic stimulation increased ATP content in control cardiomyocytes, but Perm1 knockdown reduced this response. Conclusions. Perm1 is a cardiomyocyte-intrinsic determinant of bioenergetic reserve, dispensable for resting cardiac function but required to expand ATP production under increased demand. These findings identify a mode of metabolic regulation that is invisible to resting functional phenotyping.

physiology↗

bioRxiv2026

NSD2 deficiency disrupts the epigenetic landscape of cochlear hair cells to enhance ECM expression and to perturb cochlear hair cell function

Mammalian cochlear hair cells (HC) are crucial for hearing, and histone modifications may play an important role in HC maturation and functions. However, whether and how NSD2, a histone H3 lysine 36 (H3K36) dimethyltransferase, affects cochlear HC maturation remains unclear. Here, we established an HC-specific Nsd2 knockout mouse model and discovered that the loss of NSD2 results in severe structural defects in stereociliary bundles and profound hearing impairment. Integrated multi-omics profiling revealed that NSD2 deficiency leads to a decreased expression of H3K36me2, which in turn alters genome-wide chromatin accessibility and gene expression. Mechanistically, NSD2 deficiency triggers the aberrant upregulation of extracellular matrix (ECM) genes at the transcriptional level, causing excessive collagen accumulation and compromising tissue structure. Crucially, ECM intervention can mitigate NSD2-deficiency mediated hearing damage. Our study reveals an epigenetic mechanism by which NSD2 regulates cochlear HC maturation through inhibition of chromatin remodeling-induced ECM accumulation, and provides a potential new therapeutic direction for hearing impairment.

cell biology↗

bioRxiv2026

A closed-loop reinforcement learning framework for rapid compound directed optimization

Generative artificial intelligence (AI) holds transformative potential for drug discovery, yet existing architectures typically operate in open loops without experimental feedback. Here we introduce rapid compound directed optimization (RCDO), a closed-loop reinforcement learning framework that accelerates the optimization process by bridging dry-lab computation with wet-lab feedback. RCDO couples a three-dimensional structure-guided generative model with a multi-level reward system updated after each design cycle using experimental measurements from all synthesized compounds, including inactive or developability-failed compounds. By continuously aligning the generative model with accumulated wet-lab measurements, RCDO substantially compresses optimization timelines. We evaluated RCDO through retrospective benchmarking against historical optimization trajectories and prospective wet-lab campaigns targeting ROR1, NLRP3, and NSD3. Across prospective evaluations, RCDO rapidly resolved key optimization bottlenecks within two to three design cycles: improving the oral exposure of an ROR1 inhibitor by 40-fold while maintaining antitumor efficacy, reducing CYP2C19 inhibition of an NLRP3 antagonist by 20-fold while preserving inflammasome activity, and boosting the binding affinity of an NSD3 hit by 18-fold. By directly coupling wet-lab feedback to generative learning, RCDO establishes an efficient platform for compound directed optimization, transforming AI-driven drug discovery from static generation into continuous experimental adaptation.

bioinformatics↗

bioRxiv2026

TRAPPC8 Is an Endogenous Brake on UFMylation That Suppresses Tauopathy

UFMylation, a ubiquitin-like protein modification, drives tau spread through the brain, but what keeps this process in check has remained unclear. We show that TRAPPC8 acts as a natural brake on UFMylation, binding directly to the E1 enzyme UBA5 to dampen pathway activity. In Alzheimer's disease brain tissue, TRAPPC8 is reduced while UFMylation is elevated, suggesting this brake fails as disease progresses. Sustaining UFMylation in human iPSC-derived neurons increased tau aggregation and spread while disrupting lysosomal function and lipid balance. Restoring TRAPPC8-UBA5 binding reversed these defects through the lysosomal protein CLN8, which restored lysosomal function and reduced tau pathology. Notably, expressing just the UBA5-binding region of TRAPPC8 was enough to suppress tau pathology in vivo, marking it as a promising therapeutic target.

neuroscience↗

bioRxiv2026

CryoForge: A Self-Correcting Agent for Cryo-EM Model Building That Learns When to Act and When to Stop

Automated atomic model building has accelerated cryo-EM structure determination, but different builders leave distinct residual error profiles requiring expert inspection. The post-building challenge is to decide which local interpretations are sufficiently supported by experimental evidence to be retained, corrected or rejected. Here we introduce CryoForge, an evidence-gated post-builder agent that separates repair proposal from repair acceptance. Rule and learning-based components identify candidate regions and prioritize legal actions, whereas an independent evidence gate evaluates each edit using map and half-map support, stereochemistry, connectivity and local structural context. Supported edits are retained; unsupported or conflicting modifications are rejected, rolled back, stopped or escalated for expert review. Across a resolution-stratified benchmark, 84.9% of 26,153 released trajectories yielded standard validated improvements and 3.7% yielded low-confidence partial improvements, with no quality-degrading edit retained in the final promoted models. Relative to rule-only control, learned prioritization reduced non-improving candidates and harmful actions while preserving global structural stability. External evaluations using an alternative initializer, same-team automated/manual-assisted challenge submissions and three recently released complex assemblies showed that CryoForge adapts to distinct residual error phenotypes and performs bounded, evidence-supported correction without uncontrolled remodeling. CryoForge provides a builder-independent, scalable and auditable correction layer between automated model generation and expert structural interpretation.

bioinformatics↗

bioRxiv2026

The HES1-SOD1 Antagonism Shapes Senescence Heterogeneity and Impacts Metastatic Relapse of Circulating Tumor Cells

Circulating tumor cells (CTCs) encounter multiple challenges within the blood microenvironment, including oxidative stress, flow shear forces, and immune surveillance, often leading to anoikis. Recently, a transitional state of CTC senescence has been identified, contributing to metastatic inefficiency. However, the molecular mechanisms linking senescent CTCs to disease relapse remain to be defined. By integrating a genetic model of cortactin knockdown-induced CTC senescence with single-cell multi-omic analyses, we revealed two distinct senescent CTC subpopulations marked by HES1 expression levels. These HES1low and HES1high subpopulations exhibited differential evolutionary trajectory dynamics and unique molecular and metabolic signatures, which were significantly correlated with adverse clinical outcome across several patient cohorts. HES1low senescent CTCs displayed enhanced mitochondrial fitness, oxidative phosphorylation, and ROS-detoxifying capabilities, resulting in more efficient tumor regrowth with a pro-inflammatory and thrombotic phenotype when compared to the HES1high group. Mechanistically, HES1 directly bound to the Sod1 promoter and repressed its expression, leading to redox imbalance and mitochondrial dysfunction that were linked to weakened tumor regrowth capacity. Both senescent CTC subpopulations were broadly resistant to cytotoxic and targeted therapies, yet they showed elevated dependency on anti-apoptosis programs that make them susceptible to dual blockade by SOD1 inhibitor and the anti-senolytic drug ABT737 in vivo. Finally, in a prospective cohort of on-treatment melanoma patients, HES1 senescent CTCs were highly enriched in patients with progressive disease. Thus, the HES1-SOD1 antagonism shapes CTC senescence heterogeneity and contributes to differential tumor relapse, which can be therapeutically explored for eliminating residual metastatic disease.

cancer biology↗

bioRxiv2026

Vancomycin promotes key microbiota-pathogen interactions and removes protective bottlenecks to enteric infection

Antibiotic exposure disrupts enteric pathogen colonization resistance, yet how antibiotics reshape pathogen population dynamics, infection bottlenecks, and strain-level heterogeneity in the gut remains poorly understood. Here, we combine high-resolution pathogen barcoding, transcriptomic, and metabolomic analyses to quantify how short-term vancomycin perturbation alters infection ecology in vivo. We use Citrobacter rodentium as a model for human infection by pathogenic Escherichia coli-an antimicrobial resistance priority group-to demonstrate that just two days of vancomycin pre-treatment profoundly reshapes infection trajectories, driving rapid, global gut colonization, a dramatic increase in pathogen founding population size, and preservation of strain diversity across intestinal sites. Notably, vancomycin eliminated the hallmark heterogeneity of C. rodentium infection, resulting in fully reproducible colonization across hosts. Population-level analysis revealed that antibiotic treatment relaxes competitive constraints both with the resident microbiota and among clonal pathogen lineages, allowing early-established founders to persist and expand. Despite accelerated pathogen engraftment and tissue pathology, transcriptomic analysis revealed reduced virulence gene expression. Instead, antibiotic-induced metabolic restructuring of the gut created permissive conditions for pathogen expansion. Interactions with a vancomycin-altered microbiota, dominated by Akkermansia and Bacteroides, further promoted nutrient cross-feeding and influenced epithelial attachment. Together, we illustrate how short-term antibiotic exposure reshapes enteric infection by removing ecological bottlenecks that normally constrain strain diversity and infection outcomes. These findings have implications for antibiotic use, antimicrobial resistance transmission, and therapeutic strategies that rely on competition-driven dynamics, such as strain replacement.

microbiology↗

bioRxiv2026

Dependencies in heterogeneous, lineage plastic patient-derived prostate cancer organoids revealed through integrated single-cell multiomics and CRISPR screening

Lineage plasticity and tumor heterogeneity limit the effectiveness of targeted therapies, yet the functional dependencies used to nominate therapeutic targets are often derived from homogeneous systems that fail to capture this complexity. Here, we establish a framework to resolve state-specific genetic vulnerabilities by integrating single-cell multiomics (RNA and ATAC) with pooled CRISPR-Cas9 screening across a large panel of patient-derived organoids (PDOs) from castrate-resistant prostate cancer (CRPC) and neuroendocrine prostate cancer (NEPC). We generate a single-cell multiome atlas spanning >190,000 cells across 22 PDOs, defining seven lineage states--including intermediate and plastic populations not resolved by bulk profiling--and demonstrate that these lineage programs robustly classify independent transcriptomic datasets from prostate cancer patient tumors. By systematically coupling this atlas to subtype-resolved CRISPR screens, we construct a functional dependency map linking cell state in heterogeneous 3D human tumor models. We show that intratumoral heterogeneity fundamentally reshapes the interpretation of gene essentiality, whereby gene-level depletion reflects the composite behavior of co-existing subpopulations, and identify a general principle in which resistant "limiting" populations disproportionately determine aggregate fitness effects. This framework reveals both canonical and previously unrecognized lineage-restricted dependencies within highly plastic tumor and NEPC states, including a therapeutically targetable dependency on the aryl hydrocarbon receptor (AHR) in a novel hybrid stem-like/ASCL1 population. Together, these data establish an extensive multi-dimensional prostate cancer resource, identify novel lineage-resolved biology, and provide a generalizable strategy for interpreting functional genomics in heterogeneous human tumors.

cancer biology↗

bioRxiv2026

Generative machine learning unlocks the first proteome-wide image of human cells

The spatial organization of proteins within cells governs virtually all cellular functions, yet current imaging can simultaneously visualize only tens of proteins, orders of magnitude below the thousands populating a single human cell. Here we present ProtiCelli, a deep generative model that simulates microscopy images for 12,800 human proteins from just three cellular landmark stains. Trained on 1.23 million Human Protein Atlas images, ProtiCelli outperforms existing methods in reconstruction accuracy and textural fidelity, and generalizes to unseen cell types and drug perturbations. Simulated images preserve hierarchical subcellular organization, recapitulate known protein protein interaction landscapes, and resolve compartment-specific functions of moonlighting proteins at single cell resolution. Remarkably, the model infers drug-induced changes in protein expression and localization from cell morphology alone, predicts cell cycle stage without dedicated markers, and enables unsupervised segmentation of subcellular compartments and spatial decomposition of gene sets into functional regions. We leverage ProtiCelli to generate Proteome2Cell, a dataset of 30.7 million simulated images spanning 2,400 virtual cells across 12 human cell lines, enabling hierarchical single-cell models that distinguish conserved from dynamic protein architectures. Integrated into the Human Protein Atlas, Proteome2Cell democratizes exploration of these virtual cells. By computationally bridging the experimental scalability gap, ProtiCelli establishes a foundation for spatial virtual cell modeling.

cell biology↗

bioRxiv2026

IgG Propels Atherosclerosis by Noncanonically Activating Macrophages

Despite being a central component of adaptive immunity and a highly abundant serum protein, the contribution of IgG to the milieu of atherosclerosis remains unappreciated. Here, we identify a pro-atherogenic role for IgG as it activates an innate immune cascade, independent of its classical antigen-neutralizing function. Analyses of human coronary artery plaques reveal a positive correlation between IgG and cardiovascular and cerebrovascular disease severity. Integrated single-cell plaque analyses localize IgG, coinciding with its recycling receptor FcRn, to pro-inflammatory and foamy macrophages. Genetic ablation of FcRn in myeloid cells prevents IgG from accumulating in mouse atherosclerotic lesions, diminishing plaque size and inflammation. Mechanistically, IgG acts as an endogenous ligand for TLR4, triggering NF-{kappa}B-NLRP3 inflammasome signaling without requiring its antigen-binding domain. Additionally, IgG accelerates macrophage foam cell formation through upregulation of downstream effector LCN2. Our work uncovers a role for previously overlooked adaptive immune molecules in the pathogenesis of atherosclerosis through a noncanonical mechanism linked with innate immunity.

immunology↗

bioRxiv2026

Interactome mapping in human excitatory neurons reveals novel risk genes and pathways in Alzheimer's disease

Alzheimers disease (AD) is an irreversible neurodegenerative disease defined by its molecular hallmarks - amyloid beta peptide plaques and neurofibrillary Tau tangles. Despite significant progress that has been made in uncovering a large number of genetic risk factors through extensive genomic sequencing and genetic studies, the molecular mechanisms driving AD-associated pathology and cognitive decline remain poorly understood. Therefore, alongside the identification of more risk genes, it is also paramount to study how these genes function and influence each other within the cellular pathways and overall molecular networks in AD-relevant brain cell types. However, current human protein-protein interactome datasets were all generated in either yeast or generic human cell lines. Consequently, many important neuronal interactions, especially neuron-specific ones, have yet been discovered. To address this critical gap, we developed a highly scalable, high-quality interactome mapping pipeline in human excitatory neurons derived from induced pluripotent stem cells (iPSC), and generated a comprehensive, neuron-specific interactome map, named ADNeuronNet, for key AD risk genes. ADNeuronNet consists of 1,767 high-confidence interactions among 1,189 proteins and is the only dataset enriched with neuron-specific genes when compared to known protein interactions, including previous large-scale interactome maps, for the same baits in the literature. Within ADNeuronNet, we identified 1,375 novel interactions, many of which are likely neuron specific. For example, we identified a neuron-specific interactor, RIN2, for major AD risk factor BIN1 and confirmed RIN2s function in recruiting BIN1 to RAB5 positive early endosomes, a process that has been well-associated with AD etiology. Additionally, we performed quantitative interaction perturbation analyses on AD risk genes with AD-associated mutations or isoforms and identified significant changes in 99 protein interactions among 11 different protein variants. Finally, we found that subunits from the anaphase-promoting complex/cyclosome (APC/C), another novel BIN1 interactors identified by ADNeuronNet, mediated modulation of Tau-aggregation in neurons via regulation of APOE expression, uncovering a previously unrecognized BIN1-APC/C-APOE regulatory axis in AD pathobiology. In summary, these findings illustrate how our neuron-specific ADNeuronNet can be leveraged to uncover new risk gene candidates and cellular pathways that help advance our understanding of molecular mechanisms underlying AD etiology.

systems biology↗

bioRxiv2026

cGAS inhibition delays TDP-43-driven ALS Pathogenesis

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder marked by motor neuron loss and cytoplasmic mislocalization of TAR DNA-binding protein 43 (TDP-43), a key regulator of RNA splicing. However, the upstream modulators of this process remain poorly defined. Here we identify cyclic GMP-AMP synthase (cGAS) as a central mediator of TDP-43 pathology and associated mis-splicing. cGAS expression was elevated in ALS patient brains and enriched across activated microglia. In human iPSC-derived microglia-motor neuron co-cultures, neuronal TDP-43 pathology triggered microglial cGAS activation, whereas pharmacological inhibition with a potent human cGAS inhibitor reduced phosphorylated TDP-43, restored lysosomal and phagocytic programs, normalized microglial reactivity, and reversed TDP-43-associated RNA splicing defects. In vivo, cGAS inhibition in TDP-43 Q331K mice reversed widespread RNA splicing abnormalities across neurons and oligodendrocyte lineage cells, attenuated neurodegenerative pathology, and preserved motor function. Together, these findings identify cGAS as a druggable upstream regulator linking innate immune signaling to TDP-43-dependent RNA mis-splicing and neurodegeneration, and establish cGAS inhibition as a promising therapeutic strategy for ALS.

neuroscience↗

bioRxiv2025

PBX-dependent and independent Hox programs establish and maintain motor neuron terminal Identity

Motor neuron (MN) diversity is essential for producing the broad repertoire of animal movements, yet the molecular mechanisms that specify MN subtypes remain incompletely defined. Here, we investigate how Hox genes and their PBX cofactors shape cholinergic MN subtype identity along the anterior-posterior (A-P) axis of the C. elegans ventral nerve cord (VNC). In anterior MNs, we show that the anterior Hox genes ceh-13 (Lab/Hox1) and lin-39 (Scr/Dfd/Hox4-5) collaborate with the Hox cofactor ceh-20 (Exd/Pbx1-4) and the terminal selector unc-3 (Collier/Ebf1-4) to activate terminal identity genes. In posterior nerve cord MNs, the mid-body Hox gene mab-5 (Antp/Hox6-8) represses terminal identity gene expression by antagonizing unc-3 in a ceh-20-dependent manner. Notably, mab-5 and ceh-20 are required not only during early development but also in later life stages to maintain posterior MN identity. In lumbar MNs, the posterior Hox gene egl-5 (Abd-A/Abd-B/Hox9-13) collaborates with unc-3 to activate lumbar-specific MN terminal identity genes in a ceh-20-independent manner. We further find that ceh-20 is necessary for Hox gene expression (ceh-13, lin-39, mab-5) in VNC MNs, supporting a model where Hox positive autoregulation requires PBX activity. Together, these findings reveal PBX-dependent and independent roles for Hox genes in establishing and maintaining MN identity, illustrating how combinatorial interactions between Hox factors and terminal selectors generate neuronal subtype diversity. AUTHOR SUMMARYAnimals rely on many different types of motor neurons to generate precise and flexible movements, but how these neuron subtypes are specified remains an open question. In this study, we examine how a family of developmental genes called Hox genes, together with their cofactors, help define distinct motor neuron identities in the nervous system of the nematode Caenorhabditis elegans. We find that different Hox genes act in specific regions of the ventral nerve cord to either turn motor neuron identity genes on or keep them off. In anterior motor neurons, certain Hox genes work together with a cofactor called PBX and a neuron-specific regulator (UNC-3) to activate genes required for proper motor neuron function. In contrast, a mid-body Hox gene suppresses these genes in posterior neurons, while a more posterior Hox gene activates a unique set of genes in lumbar motor neurons through a different mechanism. Importantly, we show that some Hox genes and PBX are needed not only during early development but also later in life to maintain motor neuron identity. Together, our findings reveal how combinations of Hox genes and cofactors generate and preserve motor neuron diversity, providing insight into general principles of nervous system development.

developmental biology↗

bioRxiv2025

tBid mediated genetic ablation of connective tissue cells reveal their key regulatory function during limb regeneration in axolotls

Limb regeneration in adulthood is a fascinating phenomenon unique to salamanders, requiring precise coordination and interaction of various cell types. Connective tissue (CT) constitutes a major component of the limb and serves as the primary reservoir of positional information during regeneration. However, the direct evidence and the extent of CT in positional contribution remain limited. Here we develop an inducible Cre-LoxP-mediated tBid cell ablation system in axolotls and demonstrate efficient elimination of targeted muscle tissue and muscle stem cells through transient electroporation or genetic approach. Ablation of CT cells at early stages of regeneration results in delayed regeneration and loss of proximal limb segments (e.g., upper and lower limb), with minimal impact on hand differentiation. CT ablation during development yields similar defects as observed during regeneration, indicating an essential role for CT cells in determining the positional identity of developing and regenerating limb structures. Further scRNA-seq reveal a progressive proximal-to-distal transition among CT cells and identify distinct CT subtypes that contribute to proximal and distal segments. CT ablation results in reduction of the differentiated CT1--a subpopulation that maintains proximal identity throughout regeneration, and a delay of distal transformation in the proliferative CT1--a major CT subtypes showing progenitor property. These cellular shifts likely lead to the observed morphological deficits in regenerating limbs after CT ablation. Moreover, differentiated CT1 cells enhanced their interactions with surrounding cell types after CT ablation, to modulate adaptive proliferation in other populations, e.g., muscle stem cells. Our work establishes a tBid-based ablation strategy for functional studies of CT and other cell types in axolotls, and provides direct evidence demonstrating pivotal roles of CT cells in guiding positional memory and segmental patterning during organ regeneration.

molecular biology↗

bioRxiv2025

Benchmarking of duplex sequencing approaches to reveal somatic mutation landscapes

Detecting somatic mutations in normal tissues is challenging due to sequencing errors and the low allele fractions of post-zygotic variants. Duplex sequencing greatly reduces errors and can detect mutations at any allele fraction, but systematic, cross-platform comparisons are lacking. We present a comprehensive benchmarking of six duplex sequencing technologies used by the SMaHT Network: CODEC, CompDuplex-seq, HiDEF-seq, NanoSeq, ppmSeq, and VISTA-seq. We evaluated their performance using cord blood DNA, a tumor-normal cell line mixture, and homogenates from six human tissues. Each method shows distinct profiles in genomic footprint, sensitivity, and cost. Despite differences in library construction and sequencing platforms, estimates of mutation rates and mutational signatures are highly concordant. Integration with ultra-deep whole-genome sequencing shows that duplex approaches sensitively capture mutations and signatures beyond embryonic or clonally expanded variants. These results provide a foundation for selecting duplex methods and interpreting their data, enabling scalable single-molecule analyses of somatic mutation landscapes. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=101 SRC="FIGDIR/small/692823v1_ufig1.gif" ALT="Figure 1"> View larger version (46K): org.highwire.dtl.DTLVardef@123b221org.highwire.dtl.DTLVardef@83ba0aorg.highwire.dtl.DTLVardef@2b0ab0org.highwire.dtl.DTLVardef@1cae35e_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

bioRxiv2025

A unified language model bridging de novo and fragment-based 3D molecule design delivers potent CBL-B inhibitors for cancer treatment

The rational design of small molecules is central to drug discovery, yet current artificial intelligence (AI) methodologies for generating three-dimensional (3D) molecules are often siloed, focusing on either de novo design or fragment-based design. The lack of a holistic framework limits AIs application across the complex and multi-step pipeline spanning from novel scaffold identification to lead compound optimization, and prevents AI from effectively learning from the entire process. Here, we introduce UniLingo3DMol, a language model for 3D molecular generation, empowered by fragment permutation-capable molecular representation alongside multi-stage and multi-task training strategy. This integrated design enables UniLingo3DMol to seamlessly span both de novo and fragment-retained molecular design, demonstrating superior performance over existing generation models in in silico evaluations across more than 100 diverse biological targets. We further leveraged UniLingo3DMol in the design of inhibitors targeting CBL-B, a crucial immune E3 ubiquitin ligase and attractive immunotherapy target. This strategy led to a lead compound demonstrating excellent in vitro activity and robust in vivo anti-tumor efficacy. Our findings establish UniLingo3DMol as a generalized and powerful platform, showing the strong potential to advance AI-driven drug discovery.

bioinformatics↗

bioRxiv2025

Functional Role of Small Extrachromosomal Circular DNA in Colorectal Cancer

Extrachromosomal circular DNA (eccDNA) are circular DNA molecules that originate from chromosomal DNA but exist independently. While large eccDNA (ecDNA) contributes to tumorigenesis, the role of smaller eccDNA (<100,000 base pairs) in cancer remains unclear. Our analysis of 25 colorectal cancer (CRC) tumors and adjacent non-tumorous tissues revealed that eccDNA is significantly more abundant in tumor tissues, correlating strongly with chromosomal amplifications. The presence of whole intact genes on 1.29% of eccDNA was non-random. We identified 84 genes that recurred across tumors of multiple patients when present on eccDNA with 19% of genes being cancer-associated. eccDNA-borne genes were often accompanied by increased expression, and their contribution to expression was much larger than that from linear amplifications and the larger ecDNA. The cytokine gene CXCL5 exemplified this phenomenon, showing substantial copy-number increase and upregulation when present on eccDNA. Functional validation in cell lines showed that CXCL5 eccDNA enhanced transcriptional output and immune cell recruitment function. The recurrence and overexpression of CRC-related genes on eccDNA indicate their selection in tumors, suggest eccDNA can serve as a novel mechanism for dynamically influencing gene expression and is capable of conferring cancer phenotypes to cells. Analysis of chromatin landscapes revealed that eccDNA preferentially forms at sites of open chromatin and active transcription, with architectural boundaries marked by CTCF protein. Clinically, patients with higher eccDNA levels showed poorer relapse-free survival. These findings suggest that circular DNA elements across the entire size spectrum participate in cancer evolution, positioning eccDNA as a potential therapeutic target and prognostic biomarker.

cancer biology↗