Biology subjects

Wu, K.

Publications and source records attributed to Wu, K..

At least 37 records · Page 2Linked to original sources

bioRxiv2026

Slc25a34-Mediated Mitochondrial-to-Cytoplasmic AMP Transport Activates Brown Adipose Tissue Thermogenesis

Metabolites are emerging as signaling molecules that mediate cellular function, extending beyond their well-established roles in metabolic pathways. Members of the solute carrier (SLC) family mediate metabolite transport across cellular compartments, raising the possibility that these proteins may sense environmental stimuli and regulate cellular biological processes by triggering signaling cascades linked to metabolite transport. This study investigated the response of the SLC25A family, a unique set of inner mitochondrial membrane-localized transporters, to cold as an environmental stimulus in mediating metabolic reprogramming; and whether this reprogramming, driven by the metabolites transported by SLC25A proteins, subsequently promotes the activation of thermogenesis in brown adipocytes. After screening members of the SLC25A family for their responsiveness to cold stimuli and brown adipose tissue (BAT) activation, we found that Slc25a34 was robustly induced under these conditions. We further demonstrated that Slc25a34 mediates the transport of adenosine monophosphate (AMP), derived from de novo glucose synthesis, from mitochondria to the cytosol. This transport potentiates AMP-activated protein kinase (AMPK) signaling and glycolytic flux in brown adipocytes, both of which facilitate BAT thermogenesis during cold exposure. Intriguingly, cold exposure directly promoted the activation of peroxisome proliferator-activated receptor gamma (PPAR{gamma}), which transcriptionally upregulated Slc25a34 expression. More importantly, genetic ablation of Slc25a34 impaired BAT thermogenesis. Thus, our study reveals a novel cold-induced metabolite-sensing pathway, where Slc25a34-mediated AMP transport between mitochondria and the cytosol serves as a critical signal for activating BAT thermogenesis. These findings provide compelling evidence that metabolite transport across cellular compartments acts as a key driver of cellular physiology, thereby offering novel insights into metabolite-based therapeutic strategies for metabolic diseases. HighlightsO_LISlc25a34 is cold-responsive and transcriptionally regulated by PPAR{gamma}. C_LIO_LISlc25a34 functions specifically to mediate the mitochondrial-to-cytosolic transport of AMP in brown adipocytes. C_LIO_LIMitochondrially sequestered de novo synthesized AMP acts as a signaling reservoir, and its Slc25a34-mediated efflux to the cytosol activates AMPK and glycolysis, supporting BAT thermogenesis. C_LI

cell biology↗

bioRxiv2026

SIRT5-dependent regulation of ASL controls arginine metabolism and T cell function

NAD supplementation blunts Th1 and Th17 inflammation, in part, through arginine metabolism-dependent regulation of mitochondrial energetics, redox balance and signal transduction. Whether the NAD+-dependent sirtuin deacylases contribute to this regulation is unknown. Here, we show that both SIRT1 and SIRT5 transcript levels are induced in CD4+ T cells in human participants following oral supplementation of the NAD+ precursor nicotinamide riboside (NR). Among the sirtuin family members, SIRT5 rather than SIRT1 emerged as the predominant regulator of arginine and fumarate metabolism. Genetic depletion or pharmacologic inhibition of SIRT5 attenuated NR-mediated increases in arginine and fumarate and abolished the anti-inflammatory -effects of NR on Th1 and Th17 cytokine production. In contrast, the responses to exogenous arginine or citrulline supplementation were preserved, indicating that SIRT5 functions upstream of arginine biosynthesis. Metabolomic profiling further demonstrated that SIRT5 is required for NR-induced remodeling of the arginine biosynthetic pathway. Mechanistically, SIRT5 physically interacted with arginosuccinate lyase (ASL), promoted ASL-dependent arginine accumulation, and regulated ASL post-translational acylation, including glutarylation and malonylation. Loss of SIRT5 disrupted NR-mediated redox homeostasis, antioxidant gene expression, and cytokine suppression. Collectively, these findings identify SIRT5 as a critical mediator of NAD precursor-induced metabolic remodeling that links ASL-dependent arginine metabolism to redox balance and effector function in human CD4 T cells.

immunology↗

bioRxiv2026

Epigenetic priming with valeric acid unlocks multi-stress resistance in plants by targeted histone deacetylase inhibition

Progressive climate change is driving increasingly devastating crop losses through droughts, heatwaves and other adverse weather. Combined with shrinking arable land and spreading plant diseases, this makes it crucial to enable crops to grow efficiently in unfavorable, dynamic environments. Because stressors demand prompt, highly coordinated responses, plant stress adaptations rely heavily on epigenetic regulation. Histone deacetylases (HDAs), particularly class I, were recently shown to repress these responses. As constitutive defense is energetically costly, tools enabling temporal modulation of such mechanisms are highly sought after in crop biotechnology. This study evaluated whether valeric acid (VA), a five-carbon carboxylic acid, can inhibit HDA and activate plant defense responses. Here we show that VA is a potent HDA inhibitor that confers resistance to multiple abiotic and biotic stresses in Arabidopsis, and further validate the abiotic component in maize and tomato. Despite its simple chemical structure, structural and transcriptomic evidence shows that VA acts by selectively inhibiting two major stress-repressing deacetylases, HDA19 and HDA6. By targeting these epigenetic switches, VA activates natural defense and acclimation responses. Time-course transcriptomic analyses further revealed that priming with VA induces transcriptional memory, which, together with VA-induced metabolic rewiring, enables rapid and efficient responses to future stressors. Most importantly, despite activating energetically demanding defenses, VA promotes vegetative growth and increases yield under normal conditions, thereby breaking the growth-defense trade-off. These findings establish VA as the first epigenetic biostimulant of its kind, capable of improving plant performance under both abiotic and biotic stress while simultaneously increasing crop yield. As the epigenetic mechanisms underlying its action are evolutionarily conserved, VA priming emerges as a promising universal strategy to mitigate the climate-driven global crisis of crop losses.

plant biology↗

bioRxiv2026

Signaling input from the polar flagellum uncouples quorum sensing from cell-density regulation in Vibrio species

Quorum sensing (QS) is traditionally recognized as a signaling mechanism that monitors cell density to coordinate bacterial group behaviors. Although the architecture of many QS pathways and their dependence on cell density are well established, the impact from other physiological cues on QS response is largely unclear. We demonstrate here that flagellar integrity profoundly influences QS responses across multiple Vibrio species. Systematic disruption of flagellar components in Vibrio cholerae revealed that loss of the hook cap protein FlgD induces QS sRNAs Qrr1-4 production independent of cell density and the canonical QS regulator LuxO. This bypass mechanism also overrides normal cell density-dependent regulation of several QS-controlled phenotypes. Notably, {Delta}flgD mutation restores intestinal colonization of {Delta}luxO mutants in an infant mouse model, indicating that flagellar defects reshape V. cholerae QS during infection. This bypass mechanism is evolutionarily conserved as {Delta}flgD mutation induced LuxO-independent qrr expression in multiple Vibrio species. The FlrBC two-component system, essential for flagellar gene expression, is required for the activation of qrr expression in {Delta}flgD mutants. FlgD specifically interacts with the FlrB histidine kinase but not the FlrC response regulator, and a hyperactive FlrC variant is sufficient to drive qrr expression even in cells with FlgD. These findings suggest that not only FlrBC senses cytoplasmic FlgD levels to monitor flagellar completeness to direct flagellar gene expression, this signaling system also functions as a link to bypass the canonical cell-density control of QS by integrating cellular structural information to coordinate group behaviors.

microbiology↗

bioRxiv2026

ANCHOR: haplotype-aware allelic and isoform inference from single-cell long-read RNA sequencing with de novo variant calling

Long-read RNA sequencing enables haplotype- and isoform-resolved allelic analysis of transcriptomes, yet extending this capability to single cells and distinct cell types remains computationally challenging due to sparse coverage, sequencing errors, incomplete variant information, and reference-biased transcript assignment. Here we present ANCHOR, a haplotype-aware framework for single-cell long-read RNA sequencing that performs de novo expressed-variant discovery, molecule-level haplotype assignment and isoform-resolved allelic quantification. ANCHOR combines a signed-graph variant caller, pair hidden Markov modelling and beta-binomial UMI aggregation to infer parental allele counts for genes and splice-resolved isoforms, without requiring a pre-existing phased genotype or deep learning. In human single-cell long-read RNA benchmarks, ANCHOR improved variant-calling performance over tested long-read RNA callers at single-cell and low-to-moderate coverage, and its beta-binomial model reduced depth-driven false positives in allele-specific expression testing. Applied to newly generated single-cell long-read RNA-seq data from reciprocal mouse crosses during gastrulation, ANCHOR resolved cell-type- and isoform-specific parent-of-origin imprinting and identified an antagonistic maternally biased Sgce isoform. ANCHOR provides a general framework for allele-and isoform-resolved analysis of diploid single-cell long-read transcriptomes.

bioinformatics↗

bioRxiv2026

GalaxyVS: Exploring 100-Billion Compounds in Seconds

We present GalaxyVS, a hardware-software co-designed virtual screening framework built to explore the 100-billion commercially accessible chemical space in seconds, deployed at the National Supercomputing Center in Tianjin. Built upon the dense vector retrieval paradigm of DrugCLIP, GalaxyVS bypasses the structural dependencies and computational overhead of classical docking to enable rapid screening against experimentally determined as well as geometrically feasible pockets on AlphaFold-predicted structures. To scale this paradigm to the 100-billion level, the system must overcome the significant computational burden of offline representation encoding, critical memory and I/O bottlenecks during online retrieval, and the risks of diversity collapse and precision loss within final screening results. Utilizing the heterogeneous supercomputing infrastructure, GalaxyVS accelerates the offline encoding through deep operator adaptations and resolves online retrieval bottlenecks via disk-native vector indexing coupled with in-memory staging to ensure both broad accessibility and high throughput. Concurrently, a two-stage refinement protocol effectively mitigates diversity collapse and ensures high-fidelity affinity ranking. Consequently, GalaxyVS achieves a daily scoring throughput of 1.5 x 1016 target-ligand pairs, representing a six-orders-of-magnitude leap over previous supercomputing records. Driven by this throughput, we screened nearly 100,000 protein structures across six species against the 100-billion compound library in just 16 hours. The resulting comprehensive cross-species interaction landscape, GalaxyDB, will be openly released at https://galaxyvs.drugclip.com.

bioinformatics↗

bioRxiv2026

Dynamic partitioning shapes the in vivo organization of the E. coli RNA degradosome

The bacterial RNA degradosome is a central mediator of RNA turnover, yet how its components are organized and regulated in living cells remains unclear. Using live-cell single-molecule imaging, we quantified the spatial distribution and dynamics of the four major Escherichia coli degradosome components: RNase E, RhlB, PNPase, and enolase. These proteins occupied distinct membrane-associated and cytoplasmic pools whose relative abundance varied among proteins and changed with physiological state. In particular, PNPase underwent pronounced redistribution between membrane-associated and cytoplasmic states in response to altered RNA availability and growth conditions, whereas RNase E and RhlB remained largely membrane associated. To determine the functional consequences of this organization, we examined the degradation of lacZ reporter transcripts with different translation initiation strengths. PNPase and RhlB preferentially promoted degradation of weakly translated transcripts, whereas strongly translated transcripts were largely insensitive to their loss. Together, these results reveal that the bacterial RNA degradosome is not a static, uniformly assembled molecular machine. Instead, RNA decay is organized through dynamic intracellular partitioning of degradosome components, linking physiological state and translation status to transcript degradation.

microbiology↗

bioRxiv2026

Evolutionary analysis of vertebrate KCNH voltage-gated potassium channels and their expression in zebrafish embryos

Voltage-gated potassium channels (Kv) are a large family of potassium channels composed of 40 members across 12 subtypes. The KCNH genes encode 3 subfamilies of voltage-gated potassium channels: Kv10 (EAG, ether a go go), Kv11 (ERG, EAG-related gene), and Kv12 (ELK, EAG-like K). Kv channels play prominent roles in the neuronal and cardiovascular systems. Mutations in Kv channels have been linked to many human diseases, such as epilepsy, heart arrhythmias, and cancers. Significant progress has been made in understanding protein structures, physiological functions, and the pharmacological modifiers. However, the evolutionary history and gene expression of vertebrate KCNH genes during embryonic development remain largely unknown. We systematically identified and cloned 14 kcnh genes in zebrafish. Then, we examined vertebrate KCNH channel evolution by phylogenetic and syntenic analyses. Our data revealed that the three subtypes of the KCNH gene family have already evolved in invertebrates, long before the emergence of vertebrates. The number of vertebrate KCNH genes increased, most likely due to whole-genome duplications (WGDs). In addition, we examined zebrafish kcnh gene expression during early embryogenesis by in situ hybridization. Each subgroups genes showed similar but distinct gene expression domains with some exceptions. Most of them were expressed in neural tissues. Notably, kcnh6a showed robust expression in the developing heart, consistent with its conserved role in cardiac repolarization. Additionally, a few kcnh genes were transiently expressed in nonneural tissues, such as somites and the notochord, suggesting they may have a unique role in embryonic development. Our phylogenetic and developmental analyses of KCNH channels shed light on their evolutionary history and potential roles during embryogenesis, in line with their physiological functions and human channelopathies.

developmental biology↗

bioRxiv2026

Genos-m: a foundation model for human-associated microbial genomes

Human-associated microbial genomes encode extensive strain-level diversity and niche-specific gene repertoires that are critical to host health. However, these complex sequence features remain difficult to capture using general-purpose DNA foundation models, highlighting the need for dedicated representation learning tailored to the human microbiome. Here, we introduce Genos-m, an open-source foundation model for human-associated microbial genome representation. Genos-m was pretrained on approximately 1.2 trillion nucleotide tokens from a curated microbial genome corpus, including human-associated prokaryotic isolates, high-quality metagenome-assembled genomes (MAGs) and bacteriophages, supplemented with GTDB species-level representative genomes to broaden prokaryotic taxonomic breadth. The model uses a sparsely activated Mixture-of-Experts (MoE) Transformer architecture, with 4.7 billion total parameters, approximately 330 million activated parameters per forward pass and a maximum context length of one million base pairs. We evaluated frozen Genos-m representations across short-sequence and gene-level tasks, biosynthetic gene cluster (BGC)-based regional sequence tasks, whole-genome strain phenotype prediction, and zero-shot transfer on prokaryote-related RNAfitness assays. Across these benchmarks, Genos-m consistently ranked among the leading comparison models, with the best performance in five of eight gene-fitness regression tasks and in BGC type classification. Using sparse autoencoders, we identified sparse features in Genos-m hidden activations that aligned with annotated ORFs, intergenic regions, and tRNA and rRNA loci. In downstream applications, Genos-m-derived genome-informed species representations in-corporated into a human microbiome self-supervised learning model improved colorectal cancer (CRC)-control classification over conventional species-abundance random forest models. Genos-m also generated stable sample-level embeddings from as few as 10,000 metagenomic reads, retaining gut microbial community structure that distinguished geographic origin and aligned with enterotypes defined from full-depth taxonomic profiles. Together, these results support Genos-m as a reusable representation model for microbial genomes and metagenomes, with conclusions bounded by the reported datasets, task definitions and evaluation protocols. Genos-m model weights, inference code, and usage documentation are publicly available on GitHub (https://github.com/BGI-HangzhouAI/Genos-m) and Hugging-Face (https://huggingface.co/BGI-HangzhouAI/Genos-m).

bioinformatics↗

bioRxiv2026

BLOC1S1 regulates autolysosomal and exosomal dynamics during CD4⁺ T cell differentiation

Although the endolysosome system is central to intracellular recycling, signal transduction, and intercellular communication via exocytosis, its role in immunoregulation remains incompletely defined. We recently identified that CD4+ T cell-specific depletion of BLOC1S1, a component of multiprotein complexes regulating endolysosomal biology, predisposes toward type 2 (Th2) immunity. We therefore hypothesized that the study of BLOC1S1-deficient CD4+ T cells would expand our understanding of endolysosomal dynamics in Th2 function. Here, we demonstrate that CD4+ T cell BLOC1S1 deficiency resulted in aberrant lysosomal distribution, accumulation of endosomal vesicles, and increased exocytosis, which collectively correlated with enhanced Th2 immune responses. The phenotype was associated with upregulation of key components of the exocytosis machinery, including RAB11 and VAMP7. Functional inhibition of these vesicle trafficking proteins following siRNA knockdown of RAB11 and VAMP7 significantly attenuated Th2 cytokine secretion in BLOC1S1-deficient CD4+ T cells, highlighting their essential role in exosome-mediated cytokine export. Furthermore, exosomes derived from BLOC1S1-deficient CD4+ T cells promoted Th2 polarization in recipient cells, indicating a mechanism of intracellular amplification. Together, these findings identify BLOC1S1 as a critical regulator of lysosomal dynamics and exocytic vesicle fusion, thereby linking intracellular trafficking mechanisms to Th2 immune regulation.

immunology↗

bioRxiv2026

Actin disassembly triggers CNS myelin compaction and wrapping

Compact myelin enables rapid and precise impulse conduction in the vertebrate nervous system. During CNS development, oligodendrocytes wrap spirally around axons while compacting membranes, yet how cytoskeletal remodeling is coupled to these events remains unclear. Actin disassembly is required for wrapping, but whether wrapping is driven by iterative actin-based protrusion or follows a transition to actin-independent mechanisms is unresolved. Here we integrate tract-resolved developmental profiling, live-cell compaction mapping, and in vivo genetic perturbation to define how actin disassembly is coupled to myelin wrapping. We find that actin filaments undergo pronounced, sustained disassembly before active wrapping begins, with little evidence for persistent actin within sheaths during wrapping. We develop a live-cell assay that maps membrane compaction in cultured oligodendrocytes and show that compaction zones are depleted of actin filaments and expand when actin disassembly is promoted. Consistent with this model, oligodendrocyte-specific actin disassembly in vivo accelerates the appearance of thicker myelin early in development. Together, our results support a model in which actin disassembly promotes myelin wrapping by enabling membrane compaction, and provide a platform to dissect how compaction is regulated in development and disease.

neuroscience↗

bioRxiv2026

Ecological determinants of disease and immunity in myelodysplastic syndromes

Myelodysplastic syndromes (MDS) are clonal hematopoietic malignancies characterized by ineffective hematopoiesis, dysplastic morphology, and risk of progression to acute myeloid leukemia. While genomic alterations intrinsic to malignant MDS disease-initiating cells have been well-characterized, molecular assessment of the bone marrow in situ has been limited. Here we present single cell spatial assessment of gene expression, T cell receptors, as well as MDS-defining mutations and RNA isoforms within fixed, decalcified human bone marrow core biopsies (41 MDS, 15 controls) paired with single cell immunogenomic analysis of bone marrow aspirates (35 MDS, 6 controls). Bone marrow spatial analyses of >7.47x106 cells identified hematopoietic and non-hematopoietic populations not readily captured in dissociated tissue. We developed computational methods to compare ecological niche structures, revealing enriched hematopoietic niches and reorganization of T cell immunity in MDS patient bone marrow. In situ genotyping of mutant cells revealed increased TGF{beta} expression in malignant megakaryocytes suppressing local T cell cytotoxicity. By contrast, TGF{beta} signaling was disrupted in mutant cells due to aberrant splicing of multiple TGF{beta} signaling components. This study provides a spatially resolved landscape of human MDS bone marrow and uncovers mechanisms by which malignant cells simultaneously promote intrinsic clonal persistence while rewiring the microenvironment for immune escape.

cancer biology↗

bioRxiv2026

CACNA1C as a Prognostic Biomarker and Therapeutic Target in High-Grade Serous Ovarian Cancer: Clinical Validation and Molecular Dynamics of Nifedipine Blockade

High-Grade Serous Ovarian Cancer (HGSOC) is the most lethal gynecological malignancy due to aggressive growth, widespread metastases, and high intra-tumoral heterogeneity. Poor prognosis is largely due to late diagnosis, hence there is an urgent need to identify novel biomarkers for screening, diagnosis, and monitoring. Here, we propose the voltage-dependent calcium channel hCaV1.2 encoded by CACNA1C as a potential biomarker and therapeutic target in HGSOC. Using IHC analysis for ten ovarian cancer patients, cytotoxicity assay, TCGA gene expression and survival analyses, homology modeling, molecular docking, Calcium channel membrane assembly and molecular dynamics simulations, we tested CACNA1Cs role in HGSOC progression and the effect of blocking on cancer cell survival. We show that nifedipine (NIFE), a calcium channel blocker (CCB), had a tumor suppressive effect based on binding models predicted by three-dimensional computer assisted molecular modeling and in vitro validation using human HGSOC cell line. Using The Cancer Genome Atlas ovarian public cohort, we found CACNA1C mRNA expression strongly correlated with poor patient survival for late-stage and metastasis than primary. We also show strong correlation of CACNA1C protein expression using immunohistochemistry correlating with COH ovarian carcinomas patients disease progression. This research demonstrates that targeting HGSOC via CCBs may be therapeutically beneficial. By establishing further in vitro, in vivo, and clinical trials using FDA approved NIFE may be repurposed to target CACNA1C for HGSOC. Novelty and ImpactHigh-grade serous ovarian cancer (HGSOC) remains lethal due to late diagnosis and drug resistance. This study identifies CACNA1C (Cav1.2) as a novel prognostic biomarker and therapeutic target in HGSOC, showing that elevated expression correlates with metastatic/recurrent disease and poor survival. Using molecular dynamics and in vitro models, we demonstrate that the FDA-approved calcium channel blocker nifedipine binds stably to Cav1.2 and suppresses tumor cell growth more effectively than cisplatin. These findings support repurposing nifedipine for biomarker-driven HGSOC therapy. Translational RelevanceLate diagnosis and progressive relapses significantly contribute to the poor prognosis of ovarian cancer. Identification of a tumor biomarker that can be used for screening, diagnosis, and monitoring is critical for improving clinical outcome. Our findings demonstrate that CACNA1C is a viable diagnostic marker for HGSOC and that its blockade with CCBs reduces tumor progression, highlighting their therapeutic potential.

cancer biology↗

bioRxiv2026

Prolonged cold exposure enhances regeneration potential in Arabidopsis

Prolonged cold exposure over winter impacts plant growth and development but its role beyond flowering regulation remains underexplored. In this study, we show that extended cold enhances regenerative capacity, promoting both callus formation and shoot regeneration in Arabidopsis. This enhancement is mediated by the cold-induced AP2/ERF transcription factors C-REPEAT/DRE-BINDING FACTOR 1 (CBF1), CBF2 and CBF3 which interact with the histone acetyltransferase HISTONE ACETYLTRANSFERASE OF THE GNAT FAMILY 1 (HAG1). The CBFs recruit HAG1 to the loci of key regeneration regulators, such as WUSCHEL-RELATED HOMEOBOX 5 (WOX5), to promote their expression via histone acetylation. Our findings thus uncover an epigenetic mechanism by which prolonged cold primes plants for enhanced regeneration, highlighting how environmental cues influence developmental plasticity in plants.

plant biology↗

bioRxiv2026

Attention-Guided Multimodal Neuroimaging Fusion Network for Modeling Brain Aging Pattern

Brain age prediction from neuroimaging data provides critical insights into neurodevelopmental trajectories and neurodegenerative processes. However, effectively leveraging complementary structural and functional brain information for accurate prediction remains a major challenge. In this study, we propose an Attention-guided Multimodal brain Age prediction Network (AMAge-Net), a novel framework that integrates resting-state functional MRI (fMRI) and structural MRI (sMRI) to enhance brain age estimation. In AMAge-Net, functional features are captured from fMRI through a hierarchical Graph Attention Network, while structural features are learned from sMRI via a 3D DenseNet architecture. To enable effective cross-modal integration, AMAge-Net incorporates a Multi-Head Cross-Attention mechanism followed by a Gated Fusion Module, allowing the model to dynamically prioritize the most informative features from each modality, thereby improving interpretability and predictive accuracy. Evaluation on the Cam-CAN dataset (652 participants, aged 18-89) demonstrates that AMAge-Net outperforms state-of-the-art unimodal and multimodal baselines, achieving a mean absolute error (MAE) of 5.09, root mean square error (RMSE) of 6.52, R2 of 0.87, and Pearson correlation (PCC) of 0.94. The proposed model further demonstrates robust generalization, achieving an MAE of 4.29, RMSE of 5.59, R2 of 0.58, and PCC of 0.77 on the independent OASIS-3 dataset. Comparative and ablation studies further confirm the effectiveness of the proposed fusion strategy and modality-specific encoders. Beyond predictive performance, AMAge-Net highlights interpretable brain regions that provide insights into the mechanisms of functional and structural brain aging, while gender-specific analyses reveal distinct aging trajectories between males and females. These findings establish AMAge-Net as a powerful and interpretable approach to brain age estimation, advancing efforts to characterize healthy aging and detect early deviations associated with neurological and psychiatric disorders. Author summaryEstimating the biological age of the brain from imaging data offers a window into normal development, healthy aging, and the early stages of disease. A major challenge is how to combine information from structural scans, which show brain anatomy, and functional scans, which capture brain activity. Here, we present a new computational framework that integrates both types of data to improve the accuracy and interpretability of brain age prediction. Applied to two independent, large-scale lifespan magnetic resonance imaging datasets of individuals spanning early adulthood to late life, our framework produced highly accurate predictions and consistently outperformed existing methods. Beyond predictive performance, the model highlighted brain regions that appear especially important for age-related changes, and it revealed distinct aging patterns between men and women. These findings provide a powerful and interpretable tool for studying how the brain changes across the lifespan, with potential applications in detecting early deviations linked to neurological and psychiatric disorders.

neuroscience↗

bioRxiv2026

SMAD4 loss drives chromosomal instability during tumourigenesis via translational reprogramming

Chromosomal instability (CIN), arising from errors in chromosome segregation during cell division, is a hallmark of cancer. Whilst CIN can result from several mitotic defects, the mechanisms that initiate CIN to drive tumourigenesis remain incompletely understood. Here, we show that loss of SMAD4 reprograms translation to induce CIN, resulting in tumour formation. Multi-omics analysis of tumourigenesis models driven by loss of SMAD4 complemented by functional studies demonstrate that loss of SMAD4 in pre-neoplastic cells leads to dysfunctional mitosis and an altered global translation landscape. We show that CDK11B is translationally downregulated in SMAD4-\- cells, and re-expression of the mitosis-specific isoform of this protein (CDK11B-p58) rescues the mitotic defects. Analysis of patient tumours reveals a strong correlation between markers of CIN and SMAD4 status, indicating the clinical relevance of this phenotype. Collectively, we reveal a previously unrecognised role for SMAD4 as a gatekeeper for CIN-mediated tumourigenesis via regulation of translation.

cancer biology↗

bioRxiv2026

Systematic Evaluation Defines the Limits of Ferroptosis in Cancer Therapy

Ferroptosis is a cell death mechanism characterized by the accumulation of iron-catalyzed lipid peroxides in membrane lipid acyl chains and subsequent loss of membrane integrity.1 Despite thorough investigation of its mechanisms in cultured cells, induction of ferroptosis has unresolved clinical utility in cancer therapy. Here, we systematically evaluate ferroptosis induction via multiple mechanisms, in both cell and tumor models, using focused genetic screens, genetic loss-of-function systems, and pharmacological perturbations. Through this analysis we identify cancer cell line subsets with distinct responses to canonical ferroptosis inducers and suppressors and define the underpinnings of each. Inhibition of central in vitro ferroptosis suppressors GPX4, GCLC, or SLC7A11 across these multiple models fails to impact established tumor growth. In contrast, deficiency in the cytosolic thioredoxin reductase and pharmacologic GCLC inhibition potently induces tumor regression and triggers a form of non-ferroptotic cell death regulated by cystine availability and translation. These analyses further reveal that the principal essential function of environmental cystine in cultured cells is to support selenoprotein function, identified through investigating our finding that {beta}-mercaptoethanol supports exponential growth in cystine-free conditions. Thus, while ferroptosis activation may be efficacious alone or in combination with other therapies in specific tumor contexts, cell culture systems greatly overestimate the potential anti-cancer effects of ferroptosis induction via the GPX4 axis.

cancer biology↗

bioRxiv2026

Beyond Lipschitz: Ranking Binding Affinity in Hyperbolic Space

Despite the importance of protein-ligand affinity ranking in drug discovery, existing deep learning models struggle to distinguish hard inactives that are structurally similar to active compounds but biologically inactive. We theoretically show this failure stems from the Lipschitz continuity constraint in Euclidean space, which makes neural networks locally insensitive to subtle yet critical structural perturbations. To overcome this, we propose AlphaRank, which demonstrates that scoring affinity as the negative hyperbolic geodesic distance allows subtle tangential variations to bypass Lipschitz constraints, intuitively corresponding to pivotal binding mode factors such as conformational fit. Meanwhile, AlphaRank employs joint optimization to simultaneously ranking affinity scores among actives in a proximity-aware manner and separate actives from decoy inactives through a geometric cone constraint. Experiments demonstrate that AlphaRank outperforms state-of-the-art models like Boltz-2 in both affinity ranking and active-inactive discrimination, while providing a interpretable representation space.

bioinformatics↗