Biology subjects

Ding, K.

Publications and source records attributed to Ding, K..

At least 19 recordsLinked to original sources

bioRxiv2026

COGNITIVE HETEROGENEITY AND AGING: AN AGE-INDEPENDENT CILIARY GENE PROGRAM TRACKS COGNITIVE RESILIENCE IN MICE

Cognitive decline with age is often viewed as inevitable, yet older animals show striking heterogeneity in functional trajectories, with many preserving cognitive integrity despite physiological decline. The biological basis of this resilience remains poorly understood, partly due to insensitive behavioral testing paradigms. Using an automated home-cage platform and hippocampal RNA sequencing, we integrated behavioral phenotyping with transcriptomic profiling to disentangle molecular programs linked to cognitive performance from chronological age. Behavioral data from 906 C57BL/6 mice across 129 standardized studies, trajectory-based modeling, variance-partitioned correlation, and stability-driven gene prioritization identified a transcriptional signature (271 genes) linked to cognitive state independent of age. Preserved cognition was associated with higher expression of genes governing ciliary structure and motility, neurite organization, and synaptic architecture, while impairment was associated with neuroimmune and metabolic-stress programs. A 38-gene cilia structural module score remained associated with resilience after controlling for age (partial rho = -0.81), with consistent effects across leave-one-out resampling. Because the constituent genes were selected on this same outcome, the module-level magnitude is not an unbiased effect-size estimate; recalibration against a null that reapplies the full selection procedure gave an empirical p of 0.026. This divergence emerged among genetically identical mice, indicating non-genetic, discrete hippocampal molecular signatures underlying individual differences. Cell-type attribution against the Allen Brain Cell Atlas localized the module predominantly to ependymal cells, with 27 of 38 genes ependymal-restricted (median detection 90.2% in ependymal cells versus 1.1% in CA1 glutamatergic neurons) and three genes (Dnah9, Dnai4, Kif9) also detected in CA1. The association was not attributable to ependymal tissue content: the module score was unrelated to Foxj1 expression (rho = -0.016) and the partial correlation was unchanged after adjustment for ependymal and choroid plexus marker panels, indicating that the module reflects how strongly the ciliary program is expressed rather than how many ependymal cells each sample contained. These findings support a conceptual shift: cognitive trajectories appear shaped by intrinsic molecular set points rather than by chronological age, with expression of a motile ciliary program at the cerebrospinal fluid interface, rather than the abundance of the cells carrying it, emerging as a candidate correlate of resilience. This approach offers a generalizable strategy for dissecting functional heterogeneity from aging and identifying pathways for hypothesis-driven tests of resilience across the lifespan.

animal behavior and cognition↗

bioRxiv2026

ScProteoAgent enables natural-language-driven single-cell proteomics analysis and interpretation

Single-cell proteomics requires computational choices to remain aligned with experimental design as research questions evolve. Here we present ScProteoAgent, which translates natural-language requests into domain-specific calculations and preserves research intent, analysis design, matrix-processing history and statistical outputs as reusable analytical state. Follow-up requests reuse applicable inputs and initiate new calculations when contrasts or experimental units change. We assembled a benchmark comprising 55 tasks and 33 reference conclusions from 11 published studies, combining progressive questions within studies with complementary biological and analytical settings. In a retrospective comparison of 88 archived outputs from eight systems, ScProteoAgent achieved the highest mean score of 93.67 under a common six-dimensional rule-based evaluation. A HeLa migration case connected phenotype composition with proteomic clusters, candidate differences and original-observation support. In liver zonation, a researcher-guided follow-up reused a saved matrix to test mouse-level spatial contrasts and compare full-matrix and observed-only results. Applications in brain development, sample preservation and hematopoiesis illustrated how interpretation depends on processing choices and experimental units. ScProteoAgent connects research questions with statistical results and their analytical context, enabling existing analyses to support more specific biological investigations.

bioinformatics↗

bioRxiv2026

Molecular Editing in Protein Degradation: From a Dual NSD2/GSPT1 Degrader to an NSD2-Selective PROTAC and a GSPT1-Selective Molecular Glue

Both NSD2 and GSPT1 represent compelling targets for anti-cancer drug discovery. Herein, we carried out single-atom modification of LLC0424, an NSD2 degrader with GSPT1 neosubstrate engagement, to yield two structurally analogous molecules with divergent modes of action: 424-ND, an NSD2-selective degrader, and 424-GD, a GSPT1-selective degrader. 424-ND drove selective NSD2 degradation and suppressed androgen receptor (AR) signaling in prostate cancer cells. 424-GD induced selective degradation of GSPT1 and upregulated the integrated stress response markers ATF4 and ATF3. Biolayer interferometry revealed distinct ternary complex cooperativity profiles in the presence of NSD2 or GSPT1, which directly correlate with the observed biased degradation activity. Molecular dynamics and metadynamics simulations showed that the compounds adopt distinct low-energy conformational ensembles with different spatial orientations, likely underlying their divergent cooperativity and selectivity. These findings demonstrate that minimal structural alteration permits precision control over degrader target selectivity, while furnishing selective chemical probes for NSD2 and GSPT1.

cancer biology↗

bioRxiv2026

Endothelial ANGPT2 insufficiency impairs retinal vascularization with ROP-like neovascular tufts

ANGPT2 is widely recognized as a critical regulator of pathological neovascularization. By analyzing scRNA-seq data from neonatal retinas, we demonstrate that Angpt2 transcripts are highly enriched in tip cells relative to other endothelial subtypes, where Angpt1/4 expression is absent. However, mechanisms underlying ANGPT2 function at angiogenic fronts remain inadequately understood. Here, we show that endothelial Angpt2 deletion severely disrupted retinal vascularization, characterized by neovascular tufts and micro-hemorrhage. Similar angiogenic defects also occurred in the brain, but were less evident in other tissues examined. Mechanistically, ANGPT2 insufficiency attenuated retinal tip cell invasion with aberrant mural cell coverage, compromising sprouting into non-vascularized tissues. Retinal RNA-seq analysis revealed that transcripts associated with endothelial migration and junction assembly were reduced in Angpt2 mutants compared to littermate controls, while upregulated genes were enriched in hypoxia-responsive pathways and mural cell development. Notably, abnormal mural-tip cell associations were detected within 48 hours post-Angpt2 deletion, displaying also a hypoxia-driven transcriptomic signature. These closely resemble the vascular pathologies observed in human retinopathy of prematurity. In contrast, Angpt1 insufficiency or Angpt4 deficiency primarily affected venous morphogenesis. Collectively, our findings imply that ANGPT2 is essential for driving tip cell invasion during sprouting angiogenesis, and that its insufficiency triggers hypoxia-driven vascular anomalies.

developmental biology↗

bioRxiv2026

Novel transplantable mouse cell line model recapitulates invasive lobular breast carcinoma (ILC) phenotype and immune microenvironment.

Invasive lobular breast carcinoma (ILC) is the most common special histological subtype of breast cancer, which accounts for 10-15% of all cases. To study the phenotype characteristics, metastatic growth kinetic and immune microenvironment of ILC, we developed an orthotopically transplantable cell line model from the spontaneous mammary fat pad tumor of CDH1-PTEN dual knockout C57BL/6 mouse with Cre-loxP system, designated CPT6. CPT6 recapitulates single-file growth pattern of human ILC, with pleomorphic features and a high mitotic index. RNA sequencing together with whole exome sequencing reveals a luminal A subtype with targetable driver mutations such as Kras G12C. As a novel orthotopically transplantable ILC model in immune competent mice, CPT6 shows robust in vivo growth and metastatic rate, and has moderate immunogenicity which appears to be T-cell independent. We also profiled the immune microenvironment of CPT6, revealing a myeloid-rich environment with dominant M2-macrophage population, which is concordant with human ILC. In summary, this model recapitulates human ILC phenotype and represents a valuable preclinical platform for evaluating immunotherapy and other therapeutic strategies for invasive lobular breast carcinoma.

cancer biology↗

bioRxiv2026

A Stress-Adaptive Lipid Kinase Axis Defines Metabolic Vulnerabilities in Neuroendocrine Prostate Cancer

Neuroendocrine prostate cancer (NEPC) persists in a profoundly hypoxic microenvironment, yet the mechanisms enabling tumor adaptation to this metabolically challenging niche remain undefined. Here, we identify the lipid kinase PIKfyve as overexpressed in NEPC, functioning as a central node in a stress-lipid kinase axis that drives adaptation to persistent endoplasmic reticulum (ER) stress. Mechanistically, NEPC requires PIKfyve-mediated lysosomal degradation and lipid recycling to maintain metabolic homeostasis under hypoxia. PIKfyve inhibition disrupts lysosomal function, leading to ER stress accumulation and activation of a compensatory, sterol regulatory element-binding protein (SREBP)-dependent de novo lipogenesis program essential for NEPC survival. This stress-lipid axis creates a synthetic vulnerability between PIKfyve and fatty acid synthase (FASN), where dual inhibition synergistically amplifies ER stress, triggers the terminal unfolded protein response, and induces tumor cell death. These findings reveal a metabolic adaptation in NEPC and provide preclinical evidence that co-targeting PIKfyve and FASN can overcome hypoxia-associated stress adaptation.

cancer biology↗

bioRxiv2026

SynFit: Synergistic Contrastive Learning for Multi-Objective Protein Fitness Prediction and Optimization

Proteins function through a complex interplay of structural and biochemical properties, and mutations can reshape these properties to generate fitness landscapes spanning multiple functional objectives. A central challenge in protein engineering is the need to simultaneously optimize multiple properties. In biocatalysis, for example, practical enzyme development routinely requires the concurrent optimization of catalytic activity, selectivity, stability, and substrate generality. However, despite recent advances in computational protein design and fitness prediction, most existing approaches treat these properties independently and do not explicitly capture the dependencies and trade-offs that govern real-world protein performance. We present SynFit, a multi-objective learning framework that integrates pretrained protein language models with experimental fitness measurements for protein fitness prediction and engineering. SynFit learns both shared and property-specific protein sequence representations through a synergistic contrastive learning strategy, enabling the identification of variants that simultaneously optimize multiple functional properties. Across a large-scale multi-fitness deep mutational scanning benchmark, SynFit consistently outperforms state-of-the-art supervised models trained on individual objectives and more accurately identifies variants that balance competing functional constraints. We further applied SynFit to multi-objective enzyme design for a new-to-nature biocatalytic enantioselective borylation reaction, providing a diverse array of novel cytochrome c sextuple variants in a single round of design with simultaneously improved catalytic activity and enantioselectivity that rival the best variants obtained through directed evolution. Together, these results establish SynFit as a general framework for multidimensional protein fitness prediction and highlight its potential to enable efficient multi-objective optimization in protein engineering, particularly in biocatalysis.

bioinformatics↗

bioRxiv2026

Specification of bone marrow sinusoids requires TIE2-mediated positive feedback involving COUPTFII and VEGFR3

The bone marrow (BM) vascular network plays crucial roles in driving bone development and supporting hematopoiesis, yet the mechanisms governing its specialized architecture, particularly sinusoidal morphogenesis, remain inadequately characterized. We show in this study that TIE2 (Tek) was highly expressed by BM sinusoidal endothelial cells (SEC) and the endothelial Tek excision led to BM sinusoidal capillarization. Particularly, the BM sinusoids displayed thinner vessel diameter with the aberrant mural cell coverage in the Tek mutants. Mechanistically, TIE2 insufficiency led to a dramatic decrease of VEGFR3 in BM-SECs while its expression in hepatic sinusoids was not obviously altered. The RNA-seq analysis showed that GO terms enriched for the downregulated genes were related to the biological processes including sinusoidal development while pathways related to arterial ECs and angiogenesis were upregulated in the bone marrow of Tek mutants. The alteration of sinusoidal VEGFR3 expression occurred within 48 h after the induced endothelial deletion of Tek. Consistently, the defective BM sinusoidal formation was validated with the induced Tek deletion in VEGFR3+ SECs. The insufficiency of TIE2 ligand ANGPT1 also led to reduced sinusoidal VEGFR3, accompanied by similar BM sinusoidal defects. Furthermore, disruption of sinusoidal morphogenesis was observed in mutant mice with the endothelial excision of Nr2f2 (COUP-TFII), displaying a decreased expression of BM sinusoidal TIE2 and VEGFR3. These findings suggest that ANGPT1/TIE2 and COUP-TFII form a reciprocal regulatory loop to coordinate BM sinusoidal specification via regulating VEGFR3.

developmental biology↗

bioRxiv2026

Restoration of E-cadherin Expression Alters Metastatic Organotropism in Invasive Lobular Breast Carcinoma Models

Invasive lobular carcinoma (ILC) is the most frequently diagnosed special histological subtype of invasive breast cancer and accounts for 10 - 15% of all cases. The pathognomonic hallmark of ILC is the genetic loss of E-cadherin (CDH1) causing the disruption of adherens junctions and resulting in discohesive, linear growth. To better understand the role of E-cadherin in ILC metastasis, we generated three ILC cell lines, MDA-MB-134-VI, SUM44PE, and BCK4, with inducible E-cadherin expression, resulting in successful restoration of functional adherens junctions. E-cadherin expression reduced growth in 2D culture, and that effect was even greater in 3D ultra-low attachment (ULA) conditions where increased cell death was consistent with the previously described role of E-cadherin in anoikis. E-cadherin expression did not rescue the lack of migration and invasion of ILC cell line models; however, it decreased haptotaxis and increased adherence to Collagen I in SUM44 cells. There was no significant effect of E-cadherin expression on primary orthotopic tumor growth, but spontaneous metastasis to the reproductive tract, brain, and GI tract was reduced. Inhibition of metastasis to the reproductive tract and brain was also seen after tail vein injection of MDA-MB-134 E-cadherin-expressing cells. In summary, overexpression of functional E-cadherin in ILC models has some, but limited, effects on 2D growth in vitro and primary tumor growth in vivo, but there are pronounced effects on 3D ULA growth and metastases in vivo, with stronger effects on metastatic sites enriched in patients with ILC, especially the reproductive and GI tracts.

cancer biology↗

bioRxiv2026

Hippocampal Glucuronyl C5-epimerase promotes stress resilience by directly engaging PI3K through a non-enzymatic mechanism

Major depressive disorder (MDD) seriously affects human physical and mental health and causes global socioeconomic burdens. Stress-induced depressive etiology is linked to dendritic atrophy in the hippocampus, however, the underlying mechanism was poorly understood. Here, we identify that glucuronyl C5-epimerase (Glce), a heparan sulfae-modifying enzyme, as a critical regulator of hippocampal dendritic integrity and stress resilience. Genetic ablation of Glce in the hippocampal excitatory neurons is sufficient to induce dendritic atrophy and depressive-like behaviors, whereas its restoration rescues these deficits. Notably, Glce levels are reduced in the plasma of depressed individuals. We further find that Glce functions independently of its enzymatic activity, instead binding directly to and activating PI3K (p85/p110), thereby triggering an Akt/CREB/BDNF signaling cascade. Together, our findings uncover a non-canonical role for Glce and establish that Glce-PI3K-BDNF axis is essential for maintaining hippocampal structure and behavioral resilience, thereby highlighting a critical role of Glce in the pathophysiology of depression.

neuroscience↗

bioRxiv2026

Engineering age-adaptive mRNA lipid nanoparticle cancer vaccines via reprogramming systemic gene expression

Lipid nanoparticle (LNP)-based mRNA vaccines have transformed cancer immunotherapy, yet their efficacy in older individuals, who represent the majority of cancer patients, remains poorly understood. Here, we uncover a critical and previously underappreciated barrier to mRNA LNP cancer vaccine performance in aged hosts: impaired systemic transgene expression. Using the SM-102 mRNA LNPs as a benchmark formulation, we show that while local immune activation and antigen presentation at the injection site and draining lymph nodes remain largely intact with age, transgene expression in peripheral organs, including the liver, lungs, and spleen, is markedly reduced. This deficit limits the magnitude and durability of CD8 and CD4 T cell responses and substantially compromises tumour control. Transcriptomic profiling further reveals that attenuated transgene expression parallels broad attenuation of antigen processing, presentation, and activation pathways in immune cells from aged animals, implicating impaired systemic mRNA translation as a central driver of the downstream immune defects and antitumour efficacy loss. Building on these insights, we identify a rationally selected LNP formulation that reestablishes distal antigen expression across age groups as an engineering strategy to revive T cell immunity, achieving full rescue of therapeutic efficacy in aged mice without additional intervention. Together, these findings establish systemic mRNA translation as a tuneable lever of vaccine performance and highlight that optimizing LNP formulations to sustain systemic transgene expression across age groups may enable next-generation, age-adaptive mRNA vaccines for cancer and other diseases of aging.

bioengineering↗

bioRxiv2026

RNAElectra: An ELECTRA-Style RNA Foundation Model for RNA Regulatory Inference

BackgroundRNA regulation programs gene expression through sequence-encoded mechanisms--including RNA structure formation, protein binding, chemical modification, and RNA-RNA targeting--whose rules span nucleotide-scale motifs and longer-range context. RNA foundation models aim to learn transferable representations from large RNA corpora, but most rely on masked language modeling (MLM), where the model loss is computed on a small subset of positions and training relies on artificially corrupted inputs that are absent at downstream inference, introducing a pretraining-downstream discrepancy. ResultsWe introduce RNAElectra, a single-nucleotide-resolution RNA foundation model pretrained on diverse non-coding RNAs from RNAcentral using ELECTRA-style replaced-token detection (RTD). RTD trains a discriminator with a loss defined over all input positions on realistically corrupted sequences, providing dense supervision that better aligns pretraining with sequence-to-function fine-tuning. RNAElectra combines nucleotide-resolution tokenization with an efficient attention design to capture local regulatory motifs and longer-range dependencies within a single reusable backbone. Using a unified, sequence-only fine-tuning pipeline without task-specific architectures or auxiliary inputs, RNAElectra demonstrates strong cross-task generalization across benchmarks and downstream evaluations spanning RNA structure and function, RNA-protein and RNA-RNA interactions, RNA modification landscapes, and quantitative regulatory readouts such as translation efficiency and mRNA stability, outperforming widely used RNA foundation model baselines on the majority of evaluated tasks. In addition to predictive performance, RNAElectra supports interpretability by enabling analysis of learned representations and sequence determinants underlying model predictions. ConclusionRNAElectra establishes RTD pretraining as a practical alternative to MLM for RNA foundation modeling, coupling dense position-wise supervision with an efficient architecture to deliver broadly transferable RNA representations. This framework provides a reusable backbone for RNA regulatory prediction and supports sequence-level RNA engineering and design.

bioinformatics↗

bioRxiv2026

Endocardial TIE1 synergizes with TIE2 to regulate the atrial internal muscular network assembly

Atrial cardiomyopathy is characterized by altered atrial structures, yet its underlying genetic basis remains inadequately explored. TIE1 mutations are reported in lymphedema patients, but whether they exhibit cardiac defects alongside lymphatic abnormalities remains unknown. Here, single-cell RNA sequencing showed high Tie1 and Tek expression in atrial endocardial cells. TIE1 deficiency severely disrupted atrial morphogenesis, with only minor ventricular defects. Bulk RNA-seq showed reduced transcripts for endothelial development and cardiac trabeculation in Tie1 mutants. Separate RNA-seq analysis of the atria and ventricles further confirmed these findings, revealing a more pronounced downregulation of trabecular genes, including Tek, in the atria upon Tie1 loss. Endothelial deletion of Tek impaired trabeculation, particularly in atria. Consistently, Tie1 loss combined with a single null Tek allele disrupted both atrial and ventricular trabeculation. Strikingly, atrial morphogenesis defects appeared within 48 hours of induced endothelial Tie1 loss coupled with Tek heterozygous deletion, implying a critical endocardial role in organizing the atrial muscular network. The synergy of TIE1 and TIE2 in remodeling atrial trabeculae was confirmed postnatally, whereas TIE1 insufficiency alone had no effect. Together, TIE1 is differentially required for atrial versus ventricular development, acting synergistically with TIE2 to regulate endocardial cell-coordinated assembly of atrial internal muscular network.

developmental biology↗

bioRxiv2026

Maternal Obesity Decreases Offspring Lifespan

Data in mice, nonhuman primates, and in humans demonstrate that exposure to maternal obesity increases the risk of multiple diseases in offspring. However, little is known about the aging effects of maternal obesity on the offspring. This study shows that maternal obesity significantly reduced the lifespan of both male and female mice born to obese dams despite being weaned onto a healthy diet at three weeks of age. This reduction in longevity was linked to an increase in age-related fibrotic pathology across multiple organs, e.g., liver, heart, and kidney. Gompertz analysis of the lifespan data showed that maternal obesity offspring have reduced lifespan due to detrimental changes established early during development rather than factors that modify aging later-in-life. These findings are translationally significant as they demonstrate that the growing prevalence of MO may lead to a decrease in overall lifespan and increase in age-related diseases in the next generation.

physiology↗

bioRxiv2026

Deconvolving mutation effects on protein stability and function with disentangled protein language models

Understanding how evolutionary constraints shape protein sequences is fundamental to deciphering the molecular mechanisms underlying protein stability and function, which has broad implications in protein engineering and therapeutics development. Recent advances in protein language models (pLMs) have enabled accurate prediction of mutation effects through evolutionary information, effectively capturing the selective pressure that governs protein sequence variation. A critical challenge, however, remains in disentangling the intertwined mutation effects on protein stability and function, as evolutionary signals conflate both stability-driven and function-driven pressures, obscuring the mechanistic basis of mutation effects and limiting their utility for rational protein engineering. In this work, we introduce DETANGO, a novel deep learning framework that explicitly deconvolves the mutation effects on protein functions by removing components attributable to stability perturbations from the pLM-predicted mutation effects. Guided by computational or experimental stability measurements, DETANGO estimates a functional plausibility score for each single-point mutation that is the component of the mutation effect not accounted for by changes in stability. Single-point mutations with low functional plausibility are predicted to be stable-but-inactive (SBI) variants, whose compromised activities are caused by direct perturbations on functional mechanisms rather than structural stability. Residues enriched for such variants are inferred to be functionally critical, as indicated by the strong evolutionary pressures to maintain protein function. Through extensive benchmarking experiments, we show that DETANGO accurately identifies SBI variants and pinpoints functionally important residues across contexts, including ligand binding, catalysis, and allostery. Moreover, extending DETANGO from individual proteins to homologous protein families reveals shared and distinctive functional patterns across protein families. Collectively, these results establish DETANGO as a biologically grounded framework for disentangling evolutionary constraints on protein stability and function, advancing mechanistic understanding of protein function, and informing rational protein engineering.

bioinformatics↗

bioRxiv2025

Lymphatics unload VEGFR3 in response to the endothelial stress upon VEGFR2 insufficiency

Vascular endothelial growth factor receptor 2 (VEGFR2) is a key target for regulating the endothelial cell lineage and angiogenesis. It is also expressed by lymphatic endothelial cells (LEC) while its participation in the process of lymphangiogenesis remains inadequately characterized. We show in this study that VEGFR2 is highly expressed in dermal initial lymphatic vessels and valves. The induced deletion of pan-endothelial Vegfr2 at the neonatal stage produced a potent suppression of dermal lymphatic growth, including the thinner lymphatic diameter, the decrease of LEC number and lymphatic valves. Mechanistically, the VEGFR2 insufficiency led to a dramatic decrease of lymphatic VEGFR3, a key regulator mediating signals for lymphatic growth and remodeling. The RNA sequencing analysis revealed that GO terms enriched for the downregulated genes included biological processes related to the EC development while pathways related to the hematopoiesis and immune responses were upregulated in the skin of Vegfr2 mutants compared with the littermate controls. This was further confirmed by the RNA-seq analysis of dermal tissues 48 hours after the endothelial Vegfr2 deletion. Consistently, an inhibitory effect on the dermal lymphatic growth was also observed by targeting Vegfr2 in PROX1+ cells, manifesting a similarly altered transcriptome signature. The alteration of lymphatic gene expression was further validated by the siRNA mediated knockdown of Vegfr2 in primary lymphatic endothelial cells, showing a transcriptional trend of LEC to hematopoietic transition. Findings from this study imply that VEGFR2 is required for the maintenance of endothelial identity and its insufficiency may trigger the alteration of lymphatic transcriptional programs to diminish the VEGFR3-mediated lymphangiogenesis.

developmental biology↗

bioRxiv2025

Inducing TRIB2 Targeted Protein Degradation to Reverse Chemoresistance in Acute Myeloid Leukemia

The myeloid oncogene TRIB2 is a key driver of acute myeloid leukaemia (AML) pathogenesis, promoting chemoresistance and blocking differentiation through ubiquitin-mediated degradation of the C/EBP transcription factor. Despite its stable and sometimes elevated expression across AML subtypes, TRIB2 remains a clinically-untargeted vulnerability. Here, we present a comprehensive investigation into TRIB2 degradation mechanisms using multimodal approaches, including CRISPR knockout, mutational protein stability, small molecule TRIB2 engagement and evaluation of a novel targeted protein degrader (TRIB2-PROTAC). We identify Afatinib, a multi-ERBB covalent inhibitor, as a rapid inducer of TRIB2 degradation, triggering AML cell death via an ERBB-independent pathway. Importantly, TRIB2 degradation synergized with cytarabine, the frontline AML chemotherapy, amplifying therapeutic efficacy. Mapping of TRIB2 ubiquitination sites revealed Lys-63 as critical for its own proteolytic turnover, and a Lys to Arg degradation-resistant mutant (KallR) conferred enhanced chemoresistance and increased leukaemic engraftment in vivo. CRISPR-mediated TRIB2 knockout validated an essential role in AML cell survival. Consistently, the novel TRIB2-PROTAC (compound 5K) achieved robust TRIB2 degradation and AML cell killing at low micromolar concentrations. These findings establish TRIB2 as a compelling therapeutic target in AML and demonstrate that leveraging the ubiquitin-proteasome system to degrade TRIB2 offers a promising strategy to overcome chemoresistance. This work provides strong preclinical rationale for the development of TRIB2-targeting therapies in AML.

cancer biology↗

bioRxiv2025

NucEL: Single-Nucleotide ELECTRA-Style Genomic Pre-training for Efficient and Interpretable Representations

Pre-training large language models on genomic sequences has become a powerful approach for learning biologically meaningful representations. While masked language modeling (MLM)-based approaches, such as DNABERT and Nucleotide Transformer (NT), achieve strong performance, they are hindered by inefficiencies due to partial token supervision, pre-training/fine-tuning mismatches, and high computational costs. We introduce NucEL, the first ELECTRA-style pre-training framework for genomic foundation models, which overcomes these challenges. Through a discriminator network identifying tokens modified by a generator, Nu-cEL achieves comprehensive token-level supervision across all sequence positions, thereby markedly improving training efficiency relative to the partial supervision of masked positions inherent in MLM frameworks. By integrating ModernBERTs architectural advancements, including hybrid local-global attention and flash attention mechanisms, NucEL establishes an optimized BERT architecture for genomic sequence modeling. Unlike traditional methods that tokenize genomic sequences into 6-mers, NucEL implements single-nucleotide tokenization, enabling fine-grained resolution and improving both efficiency and interpretability. Pre-trained on the human genome only, NucEL achieves state-of-the-art performance on benchmark datasets across diverse downstream tasks in both human and non-human species, including regulatory element identification (e.g., promoters, enhancers), transcription factor binding prediction in human and mouse, open chromatin region classification, and histone modification profiles, surpassing MLM-based models of similar size and rivaling models 25 times larger, such as NT. Ablation studies provide critical insights into tokenization and masking strategies, optimizing ELECTRA-style pretraining for DNA sequences. Attention analyses reveal NucELs superior ability to capture biologically relevant sequence motifs compared to NT, offering valuable insights into its hierarchical learning process and regulatory element modeling capabilities. This work highlights the potential of ELECTRA-style pretraining as an efficient and effective strategy for advancing genomic representation learning with broad implications for future genomic research.

genomics↗