Biology subjects

Xie, T.

Publications and source records attributed to Xie, T..

At least 19 recordsLinked to original sources

bioRxiv2026

Multiscale fracture resistance of human annulus fibrosus and its deterioration with degeneration

Soft fibrous tissues must resist fracture under repeated loading, yet how degeneration compromises this resistance remains unclear. We investigated human annulus fibrosus using fracture and fatigue testing, proteoglycan depletion, microscopy, and X-ray scattering. Non-degenerate tissue exhibited pronounced fracture anisotropy, with axial toughness nearly fourfold higher than interlamellar delamination toughness. Degeneration reduced axial and radial toughness despite comparatively preserved tensile modulus and strength. Proteoglycan depletion reduced toughness most strongly in radial and delamination loading. Neither tissue state showed measurable macroscopic crack extension over 20,000 cycles at 70% of its respective monotonic failure strain. Nevertheless, fatigue produced localized structural changes, and degenerate tissue exhibited secondary delamination and diffuse collagen damage. X-ray scattering revealed fatigue- and degeneration-dependent changes in collagen periodicity and intermolecular packing. These findings connect tissue fracture resistance to collagen-matrix interactions across scales and inform the design of crack-resistant biomaterials.

bioengineering↗

bioRxiv2026

TopoFlow: Evolutionarily Conditioned Flow Matching for Protein Conformational Ensemble Generation

Proteins function through dynamic conformational ensembles rather than a single static structure, yet accurately modeling these ensembles remains challenging. Here, we introduce TopoFlow, an evolutionarily conditioned flow-matching framework for protein conformational ensemble generation. TopoFlow uses evolutionary representations from multiple sequence alignments (MSAs) to encode global conformational heterogeneity. It combines this with structural variables from a variational autoencoder (VAE) to capture local flexibility. These complementary representations are then adaptively fused through a conditional modulation module to construct unified conditioning features that guide protein conformational ensemble generation. On the ATLAS molecular dynamics (MD) benchmark, TopoFlow improved the Jensen Shannon metrics for pairwise distances (JS-PwD), radius of gyration (JS-Rg), and time-lagged independent components (JS-TIC) by 10.73%, 5.78% and 7.75%, respectively, relative to BioEmu. Notably, this performance generalizes to experiment-based evaluation, where TopoFlow exhibits strong agreement with experimental observables, including chemical shifts (CS) and small-angle X-ray scattering (SAXS). Ablation analyses further indicated that the evolutionary representations and structural latent variables contributed complementary information to ensemble generation. Our results suggest that TopoFlow, through its integration of evolutionary and structural latent information, offers a promising strategy for generating structurally plausible and conformationally informative protein ensembles.

bioinformatics↗

bioRxiv2026

Mammalian TMC Family Proteins are Mechanically Gated Ion Channels

Every known life form senses and reacts to mechanical forces. These mechanical stimuli can be converted into electrical signals by mechanically gated ion channels, a transduction cascade pivotal to numerous physiological functions including touch, hearing, mechanical pain, circulation, gastrointestinal function, and mechanical loading in various tissues. Despite continuous efforts, numerous mechanically gated ion channels with the mechanotransduction process underlying these physiological functions remain unidentified. Here, we focused on the transmembrane channel-like (TMC) protein family expressed in the cultured cells to identify those with potential mechanosensitive activity. Remarkably, in contrast to human TMC1/2 (HsTMC1/2), human TMC3-8 (HsTMC3-8) proteins are localized to the plasma membrane when heterologously expressed in the cultured cells. Further experiments revealed that mechanical poking stimuli can effectively activate HsTMC3-8. In addition, HsTMC3-8 induced stretch-activated currents and elicited well-resolved single-channel activities in response to negative pressure stimulation. The mutants near the putative pore region altered reversal potentials (Erev) of HsTMC3-8, suggesting that TMC3-8 are likely pore-forming subunits of ion channels. In summary, we proposed that TMC proteins are the largest mammalian mechanically gated ion channel family.

neuroscience↗

bioRxiv2026

Unearthing a fungal giant: Dianjunaceae fam. nov., a novel Paleocene lineage of Xylariales harbouring Dianjunus rex gen. et sp. nov.

During mycological surveys in Yunnan Province, China, specimens of a fungus producing massive, upright stromata up to 50 cm high and individually 2.2 Kg in weight were sampled. Through an integrative taxonomic approach combining detailed morphology, multilocus phylogeny (ITS, LSU, RPB2, TUB2), and phylogenomic analyses, this fungus is proposed as the new species Dianjunus rex gen. et sp. nov., the type of the new family Dianjunaceae (Xylariales). Phylogenetic analyses robustly place Dianjunaceae as a distinct sister clade to Graphostromataceae. Divergence time estimation dates the origin of this family to the early Paleocene ([~]65 Mya), coinciding with the post-K-Pg extinction period, when an estimated 75% of all plant and animal species went extinct, and a significant ecological reorganization of life on earth happened. The stromata of D. rex represent the largest fructifications documented within the Ascomycota, significantly expanding the known morphological range of the Xylariales. The study provides a comprehensive description, including a nodulisporium-like anamorph with periconiella-like branching patterns, and discusses the taxons phylogenetic placement, and distinctive morphology. This discovery highlights the unexplored fungal diversity in East Asian forests.

microbiology↗

bioRxiv2026

Eleven deep-sea coral genome assemblies unveil insights into evolution, adaptation, and coral biodiversity

Deep-sea corals are vital in maintaining coral ecosystem biodiversity, yet their genetic characteristics remain largely unexplored. Here, we present 11 deep-sea coral genome assemblies, including four Hexacorallia and seven Octocorallia species, significantly contributing new genomic information across two orders. Our analysis reveals the historical dynamics of coral speciation and the influence of environmental factors on the evolution of coral reef ecosystems.Total of 126 horizontal gene transfer (HGT) events were detected, among which genes from the ancestor of symbiodiniaceae indicate that the ancestors of deep-sea corals may have inhabited shallow-sea environments. Notably, several of these HGTs are involved in phosphorus (PhnX/PhnW) and cholesterol (DHCR7) metabolisms within corals, indicating that HGTs may serve as an adaptive survival strategy for the coral holobionts. Deep-sea corals also rely on symbiotic bacteria to synthesize 10 essential amino acids (such as valine and tyrosine), retaining only partial amino acid synthesis capacity. In addition, we investigated the evolution of key biological rhythm genes and temperature adaptation in corals. The loss of key rhythm genes (e.g., clock and cry) in deep-sea corals and copy number difference of genes related to heat stress (e.g., Cbl-b and Rchy) revealed genetic difference between deep-sea and shallow-sea corals. Our new genome assemblies enhance the understanding of deep-sea coral evolution, biodiversity, and adaptation, providing a genetic foundation for coral conservation.

genomics↗

bioRxiv2026

Aligned recordings of neural spiking activity and licking behavior in thirsty mice

Electrophysiological data, which serve as a biological signal that bridges neural activity and behavioral tasks, provide an innovative approach to neuroscience research. In this study, we constructed a dataset that contains over 2000 neurons across 117 days recorded in 20 mice containing 28,573 trials. Data for 5 mice were collected from the Secondary Motor Cortex (M2) region 8 mice was derived from the Ventrolateral Striatum (VLS) and 7 mice were from Substantia Nigra pars Reticulata (SNR). We induced licking behavior in head-fixed mice by periodically delivering water through a spout while simultaneously recording spiking activity from three brain regions and behavior related electrical signals. This dataset ensures precise temporal alignment between neural activity and behavioral events, offering a robust foundation for investigating neural encoding mechanisms and simulation of neural activities. This dataset establishes a precise spike-to-event mapping, which enables high decoding accuracy using Multilayer Perceptron (MLP) and Support Vector Machine (SVM). It can serve as a high-quality benchmark for developing encoding and decoding algorithms in neural networks, particularly Spiking Neural Networks (SNNs).

neuroscience↗

bioRxiv2026

Autoluminescence characteristics of 25 Phalaenopsis orchids following transient transformation with eFBP2

Engineering autoluminescent plants, especially horticultural crops, has recently emerged as a promising research area, with one current approach involving the transgenic introduction of fungal bioluminescence pathway genes. Although autoluminescent plants such as tobacco and petunia have been created, it remains unclear whether all horticultural plants have the potential to be autoluminescent after genetic modification, especially those having flowers with dark colors and thick cuticles. To understand whether, and what if any, floral traits affect autoluminescence potential, we assess the autoluminescence characteristics of 25 representative Phalaenopsis cultivars after transient transformation of fungal bioluminescence pathway genes, alongside several key morphological and biochemical traits. Our results demonstrate that autoluminescence characteristics are correlated with floral color lightness, organ textures and epidermal cell types. In contrast, the content of the substrates of luciferin--caffeic acid and tyrosine, and the infiltration ease of inoculation solution into floral organs after injection, have limited effects on autoluminescence characteristics. Autoluminescence intensity can be reasonably predicted using five floral traits investigated, as 80.4% of variation can be explained by these traits. Our study not only identifies specific Phalaenopsis cultivars with high potential for developing autoluminescent lines but also provides a selection framework applicable to other horticultural crops.

plant biology↗

bioRxiv2026

Vagus Nerve Stimulation Ameliorates Epilepsy via a Gut Microbiota-Butyrate-Vagus Nerve-Brain Axis

Vagus nerve stimulation (VNS) is an established therapy for drug-resistant epilepsy, yet its mechanism remains partially understood. Here, through an integrated clinical-translational approach, we demonstrate that VNS modulates the gut microbiome and alter the level of butyrate to reduce seizure susceptibility via vagus nerve-dependent signaling. In a longitudinal cohort of epilepsy patients, VNS enriched short-chain fatty acid-producing bacteria, including Akkermansia muciniphila, Prevotella, Alistipes, and Odoribacter. Among clinical responders, fecal butyrate levels decreased concomitantly with an increase in plasma butyrate concentrations. Fecal microbiota transplantation from post-VNS patients into mice reduced seizure severity and improved gut barrier integrity. In a murine model of epilepsy, VNS similarly increased the abundance of Akkermansia and plasma butyrate concentrations, decreased fecal butyrate levels, and abolished its antiseizure effects upon vagotomy. Either colonization with Akkermansia muciniphila or exogenous butyrate supplementation was sufficient to reduce seizures and anxiety-like behavior, effects that also required intact vagal signaling. Our findings define a microbiota-butyrate-vagal pathway essential for the therapeutic effects of VNS, highlighting the gut-brain axis as a target for epilepsy treatment. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/707138v1_ufig1.gif" ALT="Figure 1"> View larger version (75K): org.highwire.dtl.DTLVardef@8243e6org.highwire.dtl.DTLVardef@37f0e6org.highwire.dtl.DTLVardef@a8b321org.highwire.dtl.DTLVardef@1887342_HPS_FORMAT_FIGEXP M_FIG C_FIG HIGHLIGHTSO_LIVagus nerve stimulation (VNS) remodels the gut microbiota and alter the level of butyrate. C_LIO_LIVNS enriches short-chain fatty acids-producing bacteria, specifically Akkermansia muciniphila, in both humans and mice. C_LIO_LITransplantation of VNS-conditioned microbiota confers seizure protection and restores gut barrier integrity. C_LIO_LIAkkermansia muciniphila or exogenous butyrate is sufficient to reduce seizures via a vagus nerve-dependent pathway. C_LI

neuroscience↗

bioRxiv2025

Capturing Multi-Scale Dynamics of Aortic Valve Calcification With a Coupled Fluid Structure and Systems Biology Model

Calcific aortic valve disease (CAVD) arises from coupled interactions between blood flow, tissue mechanics, and cellular signaling. Hemodynamic forces influence endothelial and interstitial cell behavior, while the resulting tissue remodeling alters valve motion and flow patterns. Capturing this two-way feedback requires models that integrate fluid-structure mechanics with biochemical regulation, yet such multiscale coupling remains technically challenging. Previous computational models have focused on isolated aspects of the disease: fluid-structure interaction (FSI) simulations reproduce valve deformation and flow, and systems biology (SB) models describe molecular signaling that drives fibrosis and calcification. However, without coupling, these approaches cannot predict how mechanical dysfunction initiates biochemical remodeling or how biochemical changes feed back on mechanics. Here, we present a proof-of-principle, multi-physics computational framework that couples three-dimensional FSI simulations of aortic valve dynamics with a mechanistic SB model of calcification signaling. The FSI module resolves pulsatile blood flow and leaflet deformation, yielding local wall shear stresses and tissue strains throughout the cardiac cycle. These mechanical quantities are used as inputs to the SB module, which comprises key biochemical pathways governing inflammation, TGF-{beta}/SMAD signaling, and nitric-oxide (NO)-mediated inhibition within valvular cells. Simulations predict long-term calcification trajectories for valves of varying thickness, showing that fibrosis-induced stiffening lowers shear stress, reduces NO synthesis, and enhances TGF-{beta} activation, thereby accelerating calcification. While the current one-way coupling implementation is not intended yet for clinical applications, the framework is modular and extensible, allowing for future enhancements that will advance towards this goal. These include the incorporation of additional biological pathways in the SB model and implementation of a fully two-way coupling scheme between the FSI and SB models that will increase accuracy and predictive capability of the framework. By integrating physics-based hemodynamics with systems-level biochemistry, this study demonstrates the utility of a next-generation, multi-scale modeling platform for studying cardiovascular disease that unites blood flow dynamics and biochemical signaling.

systems biology↗

bioRxiv2025

Design of TCR-mimicking binders for pHLA with high potency

The rational design of high-specificity binders to peptide-HLA (pHLA) complexes remains a major challenge in personalized immunotherapy, particularly for shared neoantigens with single-point mutations. To address this, we have developed an integrated framework that combines knowledge-based deep learning with physics-based simulation for the design of highly specific pHLA binders. Applied to p53 R175H-HLA-A*02:01, molecular-dynamics-guided electrostatic filtering yielded nanomolar binders with >10-fold selectivity over wild type. Motivated by the challenge of KRAS G12V-HLA-A*03:01, which lacks electrostatic cues, we implemented message-passing neural network (MPNN)-based optimization strategies to broaden binder design beyond charge-dependent interfaces. When reapplied to p53 R175H, these strategies improved success rates from 2/5 to 5/5, with most candidates exhibiting minimal or no binding to the wild-type complex. Chimera-based and residue-retention approaches within the MPNN optimization further expanded the sequence-structure search space while preserving critical hotspot interactions. This integrated framework thus enables the design of nanomolar, mutation-selective binders to pHLA complexes, advancing next-generation personalized cancer immunotherapies.

biophysics↗

bioRxiv2025

Distinct contributions and sensorimotor encoding of PV and FoxP2 neurons in the external globus pallidus during perceptual decision-making

The external globus pallidus (GPe), a central nucleus of the basal ganglia (BG), comprises diverse cell types, including PV and FoxP2 neurons. While their roles in movement regulation and sensory processing have been studied, their contributions to perceptual decision-making remain unclear. Using a Go/No-Go visual task in mice, we found that optogenetic activation of either population impaired behavioral performance, whereas only inhibition of PV neurons, but not FoxP2 neurons, disrupted behavior. Electrophysiological recordings revealed that PV and FoxP2 neurons exhibited distinct encoding of visual stimuli, licking movement, and trial outcomes, with a larger fraction of PV neurons positively encoding the reward-predicting (Go) stimulus and the population overall preferring the rewarded over unrewarded outcome. FoxP2 neuron inhibition had minimal effects on spontaneous firing of neurons in the substantia nigra pars reticulata (SNr), a major BG output. By contrast, PV neuron inhibition altered both spontaneous and task-related activity of SNr neurons, reducing their preference for reward-predicting stimulus or unrewarded outcome in a subpopulation-specific manner. These findings indicate that PV and FoxP2 neurons in the GPe differentially influence sensorimotor processing and SNr activity, with PV neurons playing a key role in supporting accurate performance during perceptual decision-making.

neuroscience↗

bioRxiv2025

Screening of Cellular Senescence Related Genes as Biomarkers and Therapeutic Targets for Glioblastoma by Integrated Machine Learning

Glioblastoma (GBM) is an aggressive brain tumor with limited prognostic biomarkers and therapeutic targets. This study applied an integrated machine learning (IML) framework to discover cellular senescence (CS)-related gene biomarkers and candidate therapeutic targets in GBM. Using Gene Expression Omnibus (GEO) training and validation cohorts, 113 machine learning models across 11 algorithms were integrated to pinpoint CS-associated gene signatures. Differential expression analysis combined with overlap of a CellAge senescence gene set yielded 129 CS-related differentially expressed genes (DEGs). Functional enrichment of these DEGs highlighted pathways related to cellular senescence and cell cycle regulation (Gene Ontology and KEGG). A multivariate classifier constructed via stepwise generalized linear modeling (GLM) and LASSO achieved high diagnostic performance (area under the ROC curve 0.92 in training, each independent validation set is over 0.85). Seven top-ranked genes from this model were validated, with TGF{beta}I emerging as the most robust biomarker (AUC > 0.85) and its elevated expression was associated with significantly shorter overall survival. Spatial transcriptomics and single-cell RNA sequencing localized TGF{beta}I expression to tumor cell clusters harboring high copy number variation (CNV) burdens. Immune microenvironment profiling linked TGF{beta}I expression with increased macrophage infiltration. Finally, single-cell gene set enrichment (scGSEA) and AUCell analyses indicated enrichment of ECM-receptor interaction signaling in TGF{beta}I-expressing cells. In summary, IML combined with spatial and single-cell transcriptomics identified TGF{beta}I as a potent CS-related biomarker and a promising therapeutic target in GBM.

cancer biology↗

bioRxiv2025

Dynamic Retinal Pathology in Glaucoma Progression Revealed by High-Resolution Functional Imaging in Vivo

Glaucoma, a complex and multifactorial neurodegenerative disease, is commonly associated with elevated intraocular pressure and primarily characterized by the progressive loss of retinal ganglion cells (RGCs) and their axons. Despite its prevalence, our understanding and treatment of this disease remain challenging due to the intricate interplay of various pathophysiological factors and the limited capability for in vivo functional study of the disease. In this work, we investigated the dynamic retinal pathology of glaucoma from onset to late stages through longitudinal in vivo high-resolution imaging in a silicone oil-induced ocular hypertension glaucoma mouse model. We developed an optimized adaptive optics two-photon excitation fluorescence microscopy (AO-TPEFM) technique for both morphological and functional assessments of pathological changes in the retina across three distinct glaucoma phenotypes with varying progression rates. Our AO-TPEFM technology visualized the complete process of functional and structural changes in the three most important retinal components related to glaucoma: microvascular vessels, microglia, and RGCs during disease progression. Notably, our functional imaging revealed pathological alterations in microvascular circulation, microglia state, and RGC functionalities at a very early stage of disease development when retinal morphology still appeared normal, providing critical insights into glaucoma pathogenesis. This research also demonstrates that AO-TPEFM is a powerful tool for the in vivo study of general retinal diseases.

bioengineering↗

bioRxiv2025

Genomic Touchstone: Benchmarking Genomic Language Models in the Context of the Central Dogma

The emergence of genomic language models (gLMs) has revolutionized the analysis of genomic sequences, enabling robust capture of biologically meaningful patterns from DNA sequences for an improved understanding of human genome-wide regulatory programs, variant pathogenicity and therapeutic discovery. Given that DNA serves as the foundational blueprint within the central dogma, the ultimate evaluation of a gLM is its ability to generalize across this entire biological cascade. However, existing evaluations lack this holistic and consistent framework, leaving researchers uncertain about which model best translates sequence understanding into downstream biological prediction. Here we present Genomic Touchstone, a comprehensive benchmark designed to evaluate gLMs across 36 diverse tasks and 88 datasets structured along the central dogmas modalities of DNA, RNA, and protein, encompassing 5.34 billion base pairs of genomic sequences. We evaluate 34 representative models encompassing transformers and convolutional neural networks, as well as emerging efficient architectures such as Hyena and Mamba. Our analysis yield four key insights. First, gLMs achieve comparable or superior performance on RNA and protein tasks compared to models pretrained on these molecules. Second, transformer-based models continue to lead in overall performance, yet efficient sequence models show promising task-specific capabilities and deserve further exploration. Third, the scaling behavior of gLMs remains incompletely understood. While longer input sequences and more diverse pretraining data consistently improve performance, increases in model size do not always translate into better results. Fourth, pretraining strategies, including the choice of training objectives and the composition of pretraining corpora, exert substantial influence on downstream generalization across different genomic contexts. Genomic Touchstone establishes a unified evaluation framework tailored to human genomics. By spanning multiple molecular modalities and biological tasks, it offers a valuable foundation for guiding future gLM design and understanding model generalizability in complex biological contexts.

bioinformatics↗

bioRxiv2025

Neural correlates of evidence accumulation in social-affective decision-making under perceptual ambiguity

Evidence accumulation models have been widely used to explain decision-making in sensory and cognitive domains, but how this process explains social-affective decision-making under perceptual ambiguity remains unclear. Here, we integrate computational modeling with multimodal neuroscience to investigate how perceptual ambiguity in emotion judgment influences decision dynamics. Participants viewed ambiguous stimuli--morphed images blending emotional expressions--and made binary categorizations. Using drift diffusion modeling (DDM), we show that drift rate, a core index of evidence accumulation, decreases as ambiguity increases. Scalp EEG data reveal that the late positive potential (LPP) tracks drift rate in both emotional and non-emotional decisions, but only when perceptual ambiguity is relevant to the task. Likewise, single-unite recordings from the dorsomedial prefrontal cortex (dmPFC) and amygdala neurons show that trial-by-trial firing rates in both regions encode drift rate, with opposite encoding patterns. fMRI-based connectivity analyses further reveal that the strength of amygdala-dmPFC coupling correlates with individual differences in the modulator effect of perceptual ambiguity on drift rate. Together, these findings identify a corticolimbic network that dynamically modulates evidence accumulation during social-affective decision-making under perceptual ambiguity. By linking computational parameters to neural activity across multiple modalities, this work advances our understanding of how the brain resolves emotionally ambiguous perceptual information. It provides a translational framework for studying emotion recognition and decision-making impairments in clinical populations.

neuroscience↗

bioRxiv2025

Direct electrical stimulation of the human amygdala enhances recognition memory for objects but not scenes

Research from neuroscience studies using invasive neuroanatomy-inspired direct manipulations suggests the basolateral amygdala (BLA) mediates a generalized modulation of many different types of memory. In contrast, noninvasive, psychology-inspired indirect correlations suggest that specificity exists in how the BLA prioritizes experiences in memory. We used direct electrical stimulation of the BLA to investigate the specificity of the memory enhancement in the human brain. Patients undergoing intracranial monitoring via depth electrodes viewed object and scene images, half of which were followed by BLA stimulation. Stimulation enhanced long-term memory for object but not scene images. Furthermore, BLA stimulation elicited stronger evoked responses in the anterior vs. posterior medial temporal lobe (MTL), regions that preferentially process object and scene learning, respectively. These results suggest the BLA exerts an important influence over the specificity of what information is prioritized in memory, rather than a general enhancement of all memory, and provide insight into how BLA-MTL projections contribute to the dynamics of memory prioritization.

neuroscience↗

bioRxiv2025

Controllable Gap Junctions by Vitamin B12 and Light

Gap junctions, mediate rapid signal transduction between contiguous cells, which are indispensable for multicellular organisms to coordinate cellular activities across numerous physiological processes. However, precise control of gap junctions remains elusive. Herein, we present CarGAP, a single-component chemo-optogenetic tool that utilizes the C-terminal adenosylcobalamin (AdoB12) binding domain of a photoreceptor protein (i.e., CarHC) to achieve reversible control over both vertebrate and invertebrate gap junctions with spatiotemporal precision. The vertebrate CarGAP (i.e., Cx-CarGAP), created by genetically fusing connexins with CarHC in mammalian cells, can efficiently block the gap junction channels through AdoB12-induced protein oligomerization, and subsequently reinstate them via green light-induced protein disassembly. We further introduced the CarGAP system (i.e., Inx-CarGAP) to the Drosophila ovary, enabling reversible control over the heterotypic gap junctions formed by innexin2 (Inx2) and innexin4 (Inx4, also known as zero population growth, Zpg), thereby uncovering the roles of gap junctions in stem cell-niche interactions. This study illustrates CarGAP as a generalizable chemo-optogenetic tool for interrogating the functions of gap junctions in various biological contexts.

bioengineering↗

bioRxiv2025

AI-powered integration of multi-source data for TAA discovery to accelerate ADC and TCE drug development (I): TAA Target Identification and Prioritization

The advancement of T-cell engagers (TCEs) and antibody-drug conjugates (ADCs) has been hindered by fragmented data landscapes. This paper, the first in a series, introduces an AI-driven framework specifically for tumor-associated antigen (TAA) target identification and prioritization, a critical initial step in TCE and ADC development. Our framework integrates diverse datasets-- including multi-omics repositories and information from scientific publications--to systematically enhance the discovery of TAAs. We have developed a graph retrieval-augmented generation (RAG)-enhanced language model that extracts insights from biological and clinical literature, while integrating curated public oncology-related omics databases such as TCGA, GTEx, single-cell atlases, and additional omics datasets. This approach prioritizes TAAs with high tumor selectivity and low on-target/off-tumor risk. By unifying diverse knowledge sources, our method provides a scalable, efficient, and data-agnostic strategy to address attrition challenges in both ADC and TCE drug development pipelines, focusing initially on TAA target identification and prioritization to transform the landscape of cancer therapeutics.

bioinformatics↗