Biology subjects

Fu, C.

Publications and source records attributed to Fu, C..

At least 19 recordsLinked to original sources

bioRxiv2026

Genome evolution and glycosyltransferase diversification shape asperosaponin biosynthesis in Dipsacus asperoides

Deciphering the genomic and enzymatic features associated with high-abundance specialized metabolism is important for understanding and engineering bioactive natural product biosynthesis. Asperosaponin , a hederagenin-based saponin, accumulates at exceptionally high levels in Dipsacus asperoides roots and is by far the predominant triterpenoid saponin. Here, by integrating a chromosome-level genome assembly with comparative genomics, transcriptomics, metabolite profiling, and biochemical characterization, we reconstructed a major biosynthetic route to asperosaponin and identified diversification of triterpenoid-associated gene families. Functional analyses revealed an ordered glycosylation sequence in which DaUGT73GZ2 preferentially accepts C-3-glycosylated substrates retaining a free C-28 carboxyl group, thereby establishing a C-3-before-C-28 glycosylation order during asperosaponin biosynthesis. Phylogenetic and biochemical analyses further uncovered functional differentiation among UGT85CA enzymes involved in C-3 arabinosylation, while genomic and expression analyses revealed lineage-specific patterns of pathway gene duplication and transcriptional regulation. Together, these findings identify genomic, enzymatic, and transcriptional features associated with the distinctive, high-abundance triterpenoid saponin profile of D. asperoides roots and provide a framework for understanding and engineering triterpenoid saponin biosynthesis.

genomics↗

bioRxiv2026

UnibindMD: An Integrated AI and Molecular Dynamics Platform for De Novo Peptide Binder Design

UnibindMD is an integrated computational platform that turns a protein target into validated de novo peptide-binder leads by chaining five orthogonal stages: deep-learning binder design (BindCraft), independent structure prediction and confidence assessment (AlphaFold3-class / Boltz-2), binding-affinity prediction (PRODIGY), developability profiling (ADMET-AI), and all-atom molecular-dynamics validation (GROMACS/CHARMM36m). The platform targets a persistent weakness of single-method design, namely that confidence metrics such as ipTM and pLDDT are self-consistent but not independent, so a design predicted to be confident may still be physically unstable. We demonstrate UnibindMD on three Andes virus (ANDV) intervention points, a setting of urgent need given a case fatality rate approaching 40%, a 2026 MV Hondius cruise-ship outbreak (13 cases, 3 deaths), and no approved therapeutic: the Gn-Gc glycoprotein complex (viral entry), the L-protein endonuclease domain (cap-snatching; PDB 6Q99), and the host VEGFR2 kinase (vascular permeability; PDB 2XIR). Across nine BindCraft campaigns (4,475 trajectories) the platform produced 464 accepted binders (acceptance 2.3-67.6%), with top candidates reaching ipTM = 0.94 and pLDDT = 0.94. The three selected leads were predicted by PRODIGY to bind with G = -9.1 to -11.1 kcal mol-1 (Kd 8 nM-0.2 M), and all three remained stably bound over 100 ns of unrestrained molecular dynamics (receptor CRMSD [≤] 2.6 [A], retained peptide fold, [≥] 99.5% of frames preserving more than half of the initial interface contacts, and 75-82% helical content). ADMET-AI flagged the expected poor oral bioavailability of large polar peptides and no PAINS alerts. By converting single-prediction design into a cross-validated, dynamics-tested workflow, UnibindMD provides a rapid, scalable, and generalizable path from any protein target to validated peptide-binder candidates.

bioinformatics↗

bioRxiv2026

Agentic transcranial functional ultrasound imaging: @fUS

Functional ultrasound (fUS) imaging provides deep, wide-field access to brain function, but instrument cost and complex acquisition and analysis workflows restrict its wider adoption. Here we introduce @fUS, an agentic transcranial imaging platform with an open architecture that integrates purpose-built hardware with executable experimental skills and persistent structured memory. The system combines 128-channel, 16-bit acquisition at 125 MHz with GPU-based processing and enables functional imaging through the intact scalp and skull of mice, with approximately 100-m spatial resolution and 10-Hz temporal sampling, providing a basis for longitudinal studies without cranial-window surgery. Its agent connects scientific objectives to experimental design, direct instrument control and data analysis, retaining context across interactions and supporting inspectable workflows through text and hands-free voice control. Neuroscience trainees with limited engineering experience independently configured the system within 30 min. Agent-assisted exploration of whisker-stimulation datasets revealed low-frequency vascular changes beyond conventional response mapping, supported by complementary two-photon measurements of single-vessel diameter. By combining transcranial imaging with accessible instrument control and analysis, @fUS provides a foundation for wider adoption and larger, more diverse neuroscience datasets. More broadly, it offers a framework for scientific instruments in which measurement, computation and experimental reasoning are developed as parts of the same system.

bioengineering↗

bioRxiv2026

Systemic iron sequestration restricts the intratumoral CD8+ T cell landscape in pancreas cancer

Systemic iron sequestration occurs frequently in cancer due to inflammation-driven expression of the iron-regulatory hormone hepcidin. The impact of systemic iron availability on tumoral immunity is unclear. Here, we show that elevated serum hepcidin is associated with reduced survival and decreased intratumoral CD8+ T cells in patients with pancreas cancer. While hepcidin is not induced in murine tumor models, administration of a hepcidin mimetic phenocopies the T-cell-depleted tumor microenvironment seen in patients. Mechanistically, chronic antigen-driven mitochondrial dysfunction disrupts iron metabolism and selectively depletes high avidity CD8+ T cells during iron restriction. These findings establish a direct link between hepcidin-mediated iron sequestration and tumoral immunity and nominate systemic iron dysregulation as a therapeutic target to enhance anti-tumoral CD8+ T cell responses.

immunology↗

bioRxiv2026

TREM2 Orchestrates Myeloid Cell Programming and Immune Dysregulation in Pulmonary Hypertension

Pulmonary hypertension (PH) is a progressive and fatal disease characterized by pulmonary vascular remodeling, inflammation, and immune dysregulation. Myeloid-derived suppressor cells (MDSCs) and macrophages contribute to PAH pathobiology; however, the molecular regulators of their pathological activation remain poorly defined. The triggering receptor expressed on myeloid cells 2 (TREM2) is an immunomodulatory receptor that shapes myeloid cell metabolism, survival, and immunosuppressive function, yet its role in PH has not been investigated. Methods: Single-cell RNA sequencing (scRNA-seq) data from human pulmonary artery tissue (GSE210248; n=3 PAH, n=3 donors) were analyzed to characterize TREM2 expression across cell populations. Wild-type (WT) and global TREM2 knockout (TREM2 KO) mice were exposed to chronic hypoxia (10% FiO2, 28 days). Hemodynamic, histological, flow cytometric, and ex vivo functional assessments were performed. TREM2 expression was measured by flow cytometry in circulating MDSCs from PAH patients (n=22) and healthy controls (n=13). Results: TREM2 was markedly enriched in monocyte/macrophage populations in PH pulmonary arteries and TREM2-high immune cells exhibited transcriptional downregulation of chemotaxis programs and upregulation of antigen processing and MHC II presentation pathways. TREM2 deficiency significantly attenuated hypoxia-induced RVSP increase, right ventricular dysfunction, pulmonary inflammation, and vascular remodeling. MDSCs from TREM2 KO mice exerted significantly less suppression of CD4+ and CD8+ T cell proliferation. TREM2 was significantly elevated in circulating MDSCs from PAH patients and showed a directional association with hemodynamic severity. Conclusions: TREM2 is a novel regulator of myeloid-driven immunosuppression and vascular remodeling in PAH and warrants investigation as a therapeutic target and biomarker.

immunology↗

bioRxiv2026

Phenotypic screens identify biologic regulators of nanoparticle uptake in diffuse midline glioma

Nanoparticle drug delivery systems hold considerable promise for locoregional administration to central nervous system tumors, yet the biological determinants of nanoparticle-cancer cell interactions remain poorly understood. Using patient-derived histone-mutant diffuse midline glioma (DMG) models, we performed a pooled CRISPR-Cas9 perturbation screen to systematically identify regulators of liposomal nanoparticle delivery. The screen identified candidate genes spanning endocytosis, vesicle transport, and metabolic signaling, revealing that nanoparticle delivery is governed by a broader landscape than previously appreciated. Among these, CTNNB1, or {beta}-catenin, emerged as a common negative regulator across two independent DMG models and two distinct nanoparticle surface chemistries. Transcriptomic profiling of CTNNB1-depleted DMG cells revealed upregulation of membrane remodeling and extracellular matrix gene programs, accompanied by reduced cell stiffness measured by a microfluidic acoustic scattering assay. This resulted in a shift in endocytic activity characterized by decreased bulk-phase macropinocytosis and increased receptor-mediated endocytosis. We further identified MAPK and mTOR pathway members as nanoparticle trafficking modulators, and demonstrated concordance between genetic and pharmacologic perturbations in modulating the liposomal nanoparticle interactions in pediatric DMG cells. These findings establish a biology-first screening approach for identifying previously unappreciated regulators with potential relevance to nanoparticle-based therapeutic strategies in pediatric brain tumors.

bioengineering↗

bioRxiv2026

PEC: a robust algorithm to reconcile pedigree and SNP-chip data on the basis of LD block, haplotype information, and Mendelian conflicts

MotivationPedigree errors frequently occur in livestock populations due to long-term manual record-keeping, which reduces the efficiency of breeding programs. Although several pedigree correction methods exist, their practical application is often limited by complicated procedures, high computational cost, and insufficient accuracy. Therefore, an effective and efficient solution for pedigree error correction is needed. ResultsWe developed a new algorithm and software, PEC, to accurately and efficiently correct pedigree errors. The method matches haplotype fragments between candidate parents and offspring using estimated linkage disequilibrium patterns and subsequently checks for Mendelian conflicts to adjust the pedigree. Using simulated pig datasets, we compared PEC against SeekParentF90 and AlphaAssign in terms of accuracy, memory usage, and computation time. PEC demonstrated superior performance across all metrics. Furthermore, application of single-step genomic best linear unbiased prediction (ssGBLUP) in a real pig population showed that PEC corrected pedigrees significantly improved the accuracy and unbiasedness of genomic evaluations, highlighting the importance of pedigree error correction. AvailabilityThe PEC software is freely available at https://github.com/TXiang-lab/JPEC.

genetics↗

bioRxiv2026

Identification of Genomic Loci Associated with Cellular Rhythms in Diversity Outbred Mice

Inter-individual variation in human molecular and behavioral circadian rhythms motivates genetic dissection in model systems with human-like diversity. We quantified cellular clock phenotypes from primary skin fibroblasts of several hundred Diversity Outbred (DO) mice-each carrying a unique mosaic of eight founder genomes-by longitudinal bioluminescence recordings of a Bmal1-luciferase reporter (LumiCycle). Canonical rhythm parameters (period, phase, amplitude, damping) were extracted and exhibited broad variability (heritability {approx}13-35%), exceeding the ranges of founder strains. We performed genome-wide QTL mapping with R/qtl2 (linear mixed models with sex and experimental group covariates, kinship control, permutation-based significance, 1.5-LOD support intervals). A suggestive QTL for amplitude localized to chromosome 12 (LOD 6.9; [~]7.5-12.0 Mb), with founder effects indicating higher amplitude for C57BL/6J and lower for PWK/PhJ. Among 21 protein-coding genes in this interval, Apob (apolipoprotein B), a clock-regulated determinant of lipoprotein assembly, emerged as a strong candidate for amplitude control. A phase QTL mapped to chromosome 1 (support interval [~]1.36 Mb) with divergent founder effects (C57BL/6J, NOD/ShiLtJ, WSB/EiJ: delayed; PWK/PhJ, CAST/EiJ: advanced) and prioritized candidates including Epha4 (an Eph receptor tyrosine kinase implicated in photic entrainment) and Acsl3. Integrative analysis in GeneWeaver connected QTL gene sets to prior loci for voluntary alcohol consumption and circadian period on proximal chromosome 12, and highlighted overlaps with GWAS signals for adolescent idiopathic scoliosis and schizophrenia, suggesting shared pathways between circadian regulation, metabolism, and neurobehavioral traits. Together, these findings define reproducible genomic loci for cellular clock phenotypes in a highly recombinant population, nominate tractable candidate genes (Apob, Epha4/Acsl3) for mechanistic follow-up, and illustrate how high-diversity mouse genetics bridges cellular circadian variation with complex disease biology.

genetics↗

bioRxiv2026

Generalist large language models complement tailor-made predictors for tumor genomics interpretation

General-purpose large language models (LLMs) are trained on large corpora to acquire broad knowledge, but whether LLMs can replace, or augment, task-specific models is unclear. We evaluated LLMs on three real-world, clinically important tumor genomic interpretation tasks, in order of increasing difficulty: (i) distinguishing tumor from non-tumor mutations (n=34,415 variants), (ii) distinguishing driver from passenger mutations (n=13,469 variants), and (iii) inferring cancer type from tumor sequencing reports across multiple assays and institutions (n=102,791 samples). The best general-purpose LLMs performed as well as the benchmark tailor-made predictor for task (i). Ensembling tailor-made models with zero-shot LLMs improved their performance for tasks (i) and (ii). For task (iii), LLMs outperformed or supplemented tailor-made models on out-of-distribution data. Without fine-tuning, current LLMs already can be useful in clinical genomic interpretation by adding complementary expertise to tailor-made, state-of-the-art predictors.

genomics↗

bioRxiv2026

Selective Treg recruitment to bone remodeling niches is required for digit tip regeneration

Adult mammals have limited capacity for tissue regeneration, where most injuries resolve through fibrotic scarring rather than functional tissue restoration1-4. Studies in regenerative vertebrate species, including amphibians, teleost fish, reptiles and mammals, have established that the innate immune system plays instructive roles in regeneration5-11, yet the role of adaptive immune cells and how the immune response distinguishes regenerative from non-regenerative injuries, remain poorly understood. The mouse digit tip provides a rare mammalian model of complete multi-tissue regeneration where distal amputation through the terminal phalanx (P3) triggers complete multi-tissue regrowth, whereas a more proximal amputation of the same bone results in fibrotic scarring12-16. Using an intravital multiphoton imaging approach capable of longitudinally tracking bone remodeling and immune cells in live mice17, we find that regulatory T cells (Tregs) are selectively recruited to regenerating but not scarring digit tips. Tregs localize first to sites of osteoclast-mediated bone resorption and persist at the bone surface when an expanding stromal progenitor pool, known as the blastema, initiates digit regrowth. Acute depletion of Tregs impairs bone resorption and subsequent bone regrowth. Mice lacking T and B cells or CD4+ and CD8+ T cells show similar bone remodeling defects, suggesting a dominant role for Tregs within the adaptive immune compartment in promoting mammalian digit tip regeneration. Treg depletion impairs regeneration through an IL-10-independent mechanism, pointing to a non-canonical effector program. Lastly, pharmacological blockade of the chemokine receptor CXCR4 reduces Treg recruitment to the bone compartment, diminishes bone-associated macrophage accumulation, and attenuates bone degradation in regenerative amputations. Together, these findings identify Tregs as essential regulators of bone remodeling during mammalian digit tip regeneration.

immunology↗

bioRxiv2026

Structural Plausibility Without Binding Specificity: Limits of AI-Based Antibody-Antigen Structure Prediction Confidence Scores

Antibody-antigen binding prediction remains a central challenge for AI-driven therapeutic discovery, particularly in discriminating cognate interactions from structurally plausible but incorrect pairings. We present a controlled, AI-method- and antibody-format-agnostic evaluation framework that measures binding specificity under realistic conditions. Using 106 experimentally determined single-chain antibody (nanobody)-antigen complexes and 11,342 shuffled non-cognate pairings, we benchmarked publicly-available state-of-the-art structure prediction methods (AlphaFold3, Boltz-2, Chai-1). Although the methods tested often generated geometrically plausible complexes, internal confidence metrics (ipTM) frequently failed to discriminate correct from incorrect pairings. Increased sampling improved structural refinement but not pairing discrimination, indicating that computational resources are better allocated across independent seeds and explicit negative controls. We conclude that internal confidence scores are not inherently calibrated to binding specificity and require validation against realistic decoys. To enable community benchmarking and method development, we release [~]1.8 million AI-generated complex structures and guidance for the benchmarks ahead.

bioinformatics↗

bioRxiv2026

Calcium directs actin assembly via allosteric activation of formin INF2

Calcium signals spatiotemporally orchestrate cytoskeletal dynamics, typically through Rho GTPase-mediated signaling cascades, yet the direct molecular transducers that convert local calcium transients into spatially controlled actin assembly have remained elusive. Here, we uncover a structure-based mechanism by which calcium-bound calmodulin (Ca2+-CaM) directly activates formin INF2 to drive actin assembly. We demonstrate that Ca2+-CaM binds to the diaphanous inhibitory domain (DID) of INF2 with nanomolar affinity, inducing allosteric conformational changes that sterically disrupt INF2 autoinhibition. The unexpected bipartite Ca2+-CaM binding interface on INF2 enables ultrasensitive decoding of local calcium microdomains, such as ER-mitochondrial contacts, where activated INF2 promotes actin polymerization to facilitate mitochondrial fission. We further show that the Charcot-Marie-Tooth disease-associated INF2 R91G mutation enhances Ca2+-CaM binding via optimized interfacial dynamics, suggesting a gain-of-function disease mechanism. Our work establishes the Ca2+-CaM-INF2 axis as a direct molecular transducer linking spatial calcium signals to actin-dependent organelle dynamics, defining a Rho-GTPase-independent paradigm for calcium-cytoskeleton communication, with broad implications for INF2-linked pathologies.

biochemistry↗

bioRxiv2026

Cancer drives atherosclerotic plaque vulnerability by inducing pathological angiogenesis

Cardiovascular disease (CVD) is the leading cause of death for many cancer survivors, a phenomenon traditionally attributed to shared risk factors and cardiotoxic chemotherapies. Here, we hypothesized that cancer may also directly promote atherosclerosis. Propensity-matched analyses confirmed significantly elevated cardiovascular event rates amongst cancer patients, independent of comorbidities such as smoking. Atheroprone mice implanted with colorectal tumors demonstrated accelerated features of plaque vulnerability, driven by pathological angiogenesis and intraplaque hemorrhage. Mechanistically, tumor-secreted TNF- induced the pro-angiogenic factor LRG1 across multiple murine models and human plaques. Therapeutic interventions targeting these pathways, including with FDA-approved cytokine inhibitors or tumor resection, prevented plaque destabilization in mice and reduced coronary revascularization rates in patients. Together, these findings suggest that cancer may causally promote CVD and unveil novel translational strategies for cancer survivors.

pathology↗

bioRxiv2025

Streptavidins play a multifunctional role within a biotin-pathway antibiotic network encoded in a biosynthetic supercluster

Streptavidin, the well-known biotin-binding protein extensively used in biotechnology, is naturally co-produced in Streptomyces avidinii alongside stravidins--inhibitors of biotin biosynthesis. Here, we uncover and activate a conserved genomic region flanked by two streptavidin genes, revealing multiple biosynthetic gene clusters that produce diverse biotin-related metabolites, including stravidins, acidomycin, -methyl-KAPA, -methyldesthiobiotin, and the novel non-proteinogenic amino acid ANDA. We show -methyldesthiobiotin arises from -methyl-KAPA, illustrating how methylated analogues interfere with distinct steps of biotin biosynthesis. Contrary to the view that streptavidin functions solely by sequestering biotin, our biochemical, structural, and bioactivity analyses demonstrate that it also binds acidomycin, neutralizing its antibacterial activity to protect the producer while likely facilitating compound secretion. The crystal structure of the streptavidin-acidomycin complex reveals the molecular basis for this dual functionality. Our findings establish a multifunctional streptavidin that integrates biotin sequestration and self-resistance to balance an antibiotic network targeting the biotin pathway in microbial competition.

microbiology↗

bioRxiv2025

Ordering from the edge: Chiral multicellular pattern formation directed by actin self-organisation in cells

The mechanisms underlying both the establishment of mirror symmetry and deviations from it in development of bilateral multicellular organisms remain insufficiently understood. Actin cytoskeletons of individual cells exhibit intrinsic chirality, and a strong correlation exists between single-cell actin fibres chiral organisation and the collective alignment of cells confined to rectangular adhesive islands (2D-microtissues). Here, we demonstrate how multicellular chiral patterns can be inferred from the chiral behaviour of actin fibres in individual cells. By analysing chiral actin systems in cells with elliptical and semi-circular shapes, representing inner and boundary positions within 2D-microtissues, we defined the rules of chiral motile behaviour and formulated two models of cell alignment: (i) chiral rotation of inner cells and (ii) chiral tilting of boundary cells relative to island edges. In both models, neighbouring cells are also mutually aligned. Systematic variation of island area and aspect ratio, combined with dynamic observations, revealed the primary role of boundary cells. Chiral order first emerged at tissue boundaries and then propagated inward. This outside-in mechanism also explains why mirror-symmetric cell groups are difficult to achieve with uniformly chiral cells, but can be obtained either by reversing chirality in one subgroup or by enlarging the microtissue to minimise boundary influence.

biophysics↗

bioRxiv2025

Simple, Fast and Highly Efficient One-or Two-step Proteomic Preparation Enables Deep Profiling of Microgram-level FF and FFPE Tissues

Large-scale tissue proteomics requires workflows that are efficient, rapid, and repeatable across diverse samples. Herein, we present Simple Workflow for Integrated and Fast Tissue-preparation (SWIFT), which enables complete processing of fresh- frozen (FF) and formalin-fixed, paraffin-embedded (FFPE) tissues in either one- or two-step formats, while maintaining deep proteome coverage with high repeatability from low to microgram-level tissues. For FF tissues, an incubation process integrating lysis, reduction, alkylation, and digestion generates peptide samples directly from tissues in [≤] 1.5 h. For FFPE tissues, concurrent deparaffinization, rehydration, and de- crosslinking is achieved within 0.5 h, followed by one-step peptide preparation. Furthermore, our workflows eliminate desalting and offline cleanup steps, thereby reducing variability and total processing time. Using our methods, we identified up to [~]10,000 protein groups and [~]150,000 peptides across multiple mouse organs on timsTOF Pro. Repeatability was high (pairwise Pearsons r > 0.96 across six experimental replicates), with dynamic ranges spanning 6-7 orders of magnitude. Organ-enriched protein analysis identified functionally distinct proteins unique to each tissue Paired FF and FFPE analyses revealed preservation-induced shifts, with FFPE tissues showing reduced detection of membrane-associated and respiratory proteins, including mitochondrial Complex I. Together, our fast and simplified workflows enable deep tissue proteomics for large-scale clinical and translational studies in a cost- effective and widely-accessible manner.

biochemistry↗

bioRxiv2025

Dual-mode intramolecular agonist-dependent mechanoactivation of the adhesion GPCR ADGRG1/GPR56

Mechanical stimuli instruct cardinal cellular decisions pertaining to cell fate, proliferation, morphology, and movement1,2. How adhesion G protein-coupled receptors (aGPCRs), a large family of mechanosensors with more than 30 members in humans3-5, transduce mechanical cues into metabotropic commands has been a matter of debate due to the lack of suitable approaches to analyze receptor activation during mechanotransduction in live cells6. A central question is whether aGPCR activation strictly requires mechanical dissociation of the autoproteolyzed N- and C-terminal fragment (NTF and CTF) complex to expose the tethered intramolecular agonist (TIA; 'Stachel'). Here, we investigated human ADGRG1/GPR56 (G1), an aGPCR with roles in brain development7, skeletal muscle8, and platelet function9, to study the events during aGPCR mechanotransduction by monitoring receptor dissociation and G protein recruitment in real time during cell migration and substrate deformation. Magnetic-tweezer measurements at a physiologically relevant loading rate showed that isolated G1 GAIN domains can partially unfold and refold at forces below those required for NTF-CTF dissociation. In cells, G1 dissociation occurred at retracting fibers during cell migration10 over an adhesive substrate through its native extracellular matrix ligand, tissue transglutaminase 2 (TG2), and was enhanced by cell stretching. Simultaneous live recording of G protein recruitment and pharmacological assays during mechanical stimulation of G1 reveal a dual-mode activation mechanism: mechanical force elicits robust metabotropic signaling without requiring NTF dissociation and generates an intracellular response commensurate with the fully dissociated receptor. Both signaling modes depend on an intact Stachel element11,12 of the receptor under our experimental conditions. Our findings determine mechanotransduction principles governing aGPCR activation at the cell-ECM interface. We reveal that G1 does not function as a simple binary switch, but rather as a highly sophisticated mechanosensor that dynamically encodes localized extracellular tension into distinct, highly scalable metabotropic outputs.

cell biology↗

bioRxiv2025

Pan-Cancer Single-Cell Profiling Uncovers the Biological Characteristics of Cancer-Testis Genes

Cancer-testis genes (CTGs) are attractive immunotherapeutic targets owing to their restricted testicular expression and aberrant activation in cancers. However, their regulatory mechanisms, spatial organization, and clinical utility remain incompletely understood. Here, we leveraged large-scale single-cell and spatial transcriptomic data to perform a comprehensive pan-cancer analysis of CTGs. We established a high-confidence pan-cancer CTG catalog and uncovered a heterogeneous epigenetic regulatory landscape in which X-linked CTGs are predominantly governed by DNA methylation, whereas autosomal CTGs are more strongly associated with chromatin regulators. Building on the observation that CTG activation is a robust pan-cancer hallmark of malignancy, we developed a computational framework that enables rapid malignant cell annotation with performance comparable to established copy number variation-based methods. Spatial transcriptomic analyses revealed that CTG expression in head and neck squamous cell carcinoma is preferentially enriched at the invasive tumor front. Clinically, we highlighted the underappreciated therapeutic potential of CTGs and prioritized CT83 and DCAF4L2 as promising candidate targets for T-cell receptor-engineered T-cell therapy in triple-negative breast cancer and liver cancer, respectively. Our study advances mechanistic understanding of CTG biology and provides a valuable resource for the development of CTG-based immunotherapies.

cancer biology↗