Biology subjects

Shen, X.

Publications and source records attributed to Shen, X..

At least 55 records · Page 3Linked to original sources

bioRxiv2026

Mapping kinase-dependent tumor immune adaptation with multiplexed single-cell CRISPR screens

Immune dysfunction in cancer is enacted by multiple programs, including tumor cell-intrinsic responses to distinct immune subpopulations. A subset of these immune evasion programs can be systematically recapitulated through direct tumor-immune interactions in vitro. Here, we present an integrated, high-throughput single-cell CRISPR screening framework focused on the protein kinome for mapping the tumor-intrinsic regulation of T cell-driven immune pressure in glioblastoma (GBM). We combine pooled CRISPR interference and activation (CRISPRi/a) with immune-matched NY-ESO-1 antigen-specific allogeneic GBM-T cell co-culture and massively multiplexed single-cell transcriptomics to systematically quantify how genetic perturbation reshapes baseline tumor state and adaptive responses across graded effector-to-target ratios. We further leverage deep generative models for analyzing pooled CRISPR screens to decipher the effects of genetic perturbations on the mechanisms of tumor resistance. This framework resolves distinct modules of immune evasion and survival, including the regulation of the antigen-presentation machinery, interferon/NF-{kappa}B signaling, oxidative stress resilience, and checkpoint/cytokine programs, while identifying perturbations that reroute the continuous tumor transcriptional trajectory induced by T cell engagement. A secondary chemical screen in patient-derived GBM cultures identified putative kinase targets of immune evasion phenotypes (e.g., EPHA2 and PDGFRA), whose inhibition leads to the blockade of evasive programs and enhances T cell-mediated GBM killing. Together, this workflow provides a scalable blueprint for comprehensive charting of the genetic control of tumor-immune interactions.

genomics↗

bioRxiv2025

Distinct neural processing of real versus unreal optic flow in the human brain

Visual perception of self-movement relies on optic flow, a global motion pattern generated by self-movement in a rigid 3D environment. Such optic flow conveys both 2D features (e.g., radial velocity fields and centers of motion) and 3D structural information (e.g., motion perspective). However, many previous neuroimaging studies have used simplified 2D radial motion as "real" optic flow, leaving it unclear whether such "unreal" stimuli adequately capture how the brain processes ecologically valid self-movement information. Previous behavioral work has shown that sensitivity to expansion and contraction reverses depending on whether optic flow is ecologically valid (L. Li et al., 2025), raising the question of how such differences are represented in the human brain. We conducted two functional magnetic resonance imaging (fMRI) experiments to systematically compare cortical processing of real versus unreal optic flow. The two stimulus types were strictly matched in 2D global features and local motion signals, differing only in whether they preserved 3D structure consistent with self-movement. To minimize task-related influences, participants performed a task-irrelevant fixation point color-change detection task during scanning. Experiment 1 used a block design in which participants viewed expanding and contracting optic flow at different motion coherence levels. Experiment 2 presented real and unreal optic flow within the same scanning session to directly test neural sensitivity to ecological validity. Multivoxel pattern analysis (MVPA) revealed that, among multiple visual and optic flow-responsive regions, the human dorsal medial superior temporal area (MST) exhibited robust and consistent sensitivity to the ecological validity of optic flow. Specifically, under real optic flow, decoding accuracy for motion coherence in MST was significantly higher for expansion than for contraction, whereas this preference was reversed under unreal optic flow. Neural threshold estimates derived from fMR-metric functions further supported this pattern reversal, indicating systematic modulation of MST sensitivity by ecological structure. Experiment 2 further demonstrated that, regardless of motion pattern, MST reliably distinguished real from unreal optic flow, with decoding accuracy increasing as motion coherence increased. These findings indicate that MST encodes not only motion patterns themselves but also whether optic flow conforms to the ecological structure of self-movement. Together, these results indicate that MST responses are not determined solely by 2D radial motion features, but are modulated by whether optic flow preserves 3D structure consistent with self-movement. This work clarifies a long-standing conflation between simplified radial motion and real optic flow and highlights the importance of ecological structure in shaping dorsal-stream motion representations related to self-movement perception.

neuroscience↗

bioRxiv2025

Oxa1L-Mediated Co-translational Protein Insertion Maintains Mitochondrial Function in Parkinson's Disease Models

Mitochondrial respiratory chain integrity relies on the coordinated synthesis and membrane insertion of mitochondrially encoded proteins. How limitations in this process contribute to mitochondrial dysfunction under stress remains poorly understood. Here, we identify the mitochondrial insertase Oxa1L as a key buffering factor that modulates mitochondrial vulnerability under Parkinsonian stress conditions. Transcriptomic analyses and cellular stress models reveal pronounced disruption of oxidative phosphorylation pathways accompanied by reduced Oxa1L abundance. Functional gain- and loss-of-function analyses demonstrate that Oxa1L actively influences mitochondrial membrane potential, reactive oxygen species levels, and ATP production during stress. Notably, Oxa1L selectively governs the abundance of mitochondrially encoded respiratory chain subunits, indicating that co-translational membrane insertion represents a rate-limiting step in mitochondrial protein biogenesis. Structural analyses position Oxa1L at the mitoribosomal exit site, providing a spatial framework for coupling translation to inner membrane insertion. Together, our findings uncover a critical layer of mitochondrial quality control at the level of co-translational insertion and establish Oxa1L as a determinant of mitochondrial resilience under stress.

cell biology↗

bioRxiv2025

Substrates Coupling and Inhibition of the human Na+- dependent Cl-/HCO3-exchanger

Maintaining pH homeostasis is essential for neuronal excitability and for regulating cerebrospinal fluid (CSF). The Na+-dependent Cl-/HCO3- exchanger, NCBE (SLC4A10), abundant in the choroid plexus, contributes to CSF pH control by mediating bicarbonate flux, yet its molecular mechanism has remained unclear. Here we report cryo-EM structures of human NCBE in multiple functional states, revealing conserved binding sites for Na+, Cl-, and HCO3-, as well as an unexpected K+ site indicating relaxed cation selectivity. The inhibitor DIDS binds above the substrate cavity, locking NCBE in an outward-facing conformation and defining a subfamily-selective inhibitory mechanism. Functional assays and molecular dynamics simulations validate ion coordination and inhibitor effects. Comparison of cryo-EM structures with an AlphaFold3 inward-facing model supports an elevator-like mechanism and completes the conformational cycle of Na+-coupled HCO3- transporters. Together, these findings define the structural basis of NCBE and provide a framework for understanding ion coupling and pharmacological modulation within the SLC4 family.

biophysics↗

bioRxiv2025

On the stabilization of plant lipid droplets: Dynamic interplay between oleosins and phospholipids

As the central organelles of lipid and energy homeostasis in plant seeds, lipid droplets (LDs) consist of a neutral lipid core, decorated by phospholipids (prominently phosphatidylcholines, PCs) and surfactant-like proteins (mostly oleosins, OLs). So far, the dynamic interplay between PCs and OLs at the LD interface remains unclear. The presented work addresses this knowledge gap by reconstituting oil-in-water emulsions stabilized by OLs and PCs using microfluidic systems. Our results show that the resistance to droplet coalescence is primarily provided by OLs. We further reveal that the addition of PCs alters the assembly of OLs at the interface, reducing the OL network density and interfacial elasticity, thereby rendering a weaker interface. In conclusion, our study suggests complementary roles, with OLs acting as the primary stabilizers while PCs playing a destabilizing role. This contrast likely contributes to the observed metastability of LDs and can be exploited to design stimuli-responsive emulsions. HIGHLIGHTSO_LIOleosins form an interfacial network that stabilizes the oil-water interface. C_LIO_LIPhosphatidylcholines globally weaken the oleosin network and promote droplet coalescence. C_LIO_LIMicrofluidics enables controlled reconstitution and real-time analysis of lipid droplets. C_LIO_LIOleosin-phospholipid interplay explains lipid droplet metastability and can guide the design of bio-inspired emulsions. C_LI

biophysics↗

bioRxiv2025

Enhancing Generative Decoders with Stochastic Training on Biomedical Data with Missingness

Biomedical datasets are often heterogeneous and affected by noise or missing values. Deep Generative Decoders (DGD) provide a promising framework for latent representation learning, but their standard training procedure relies on sample-level stochastic gradient descent (SGD), which performs poorly with incomplete data. To address this, we introduce two stochastic training strategies -- Nested SGD and Feature Dropout -- that incorporate feature-level randomness into optimization. Evaluations on biomedical tabular datasets demonstrate that Nested SGD improves robustness under missingness the best, while Feature Dropout not only improves but also accelerates convergence with lower computational cost. These results suggest that feature-level stochasticity is a practical way to strengthen biomedical AI pipelines.

bioinformatics↗

bioRxiv2025

Deep-learning model guided discovery and characterization of bacterial unspecific peroxygenases

Unspecific peroxygenases (UPOs) are capable of catalyzing the selective oxidation of organic substrates under mild conditions, using hydrogen peroxide (H2O2) as the sole oxidant. This makes them one of the most promising biocatalysts for chemical synthesis. However, the major limitation restricting the application of UPOs to date is their difficulty in heterologous expression. Although more than 4,000 putative UPO enzymes have been recorded in databases, only about 50 of them can currently be heterologously expressed. All UPOs discovered so far originate from eukaryotes (mainly basidiomycetes and some ascomycetes), and they rely on the complex expression systems and post-translational modifications of their native hosts. This further exacerbates the challenges associated with heterologous expression of eukaryotic UPOs. In this work, we developed a deep-learning-based enzyme mining strategy and, for the first time, discovered novel UPO enzymes from bacteria, achieving successful heterologous expression in Escherichia coli. Bacterial UPOs differ greatly from fungal UPOs in sequence similarity, displaying completely distinct evolutionary trajectories. The discovery of bacterial UPOs advances our understanding of the catalytic mechanisms and expression characteristics within the UPO family, breaking the long-held assumption that UPOs can only originate from eukaryotes. Their excellent heterologous expression performance and broad catalytic versatility will further expand the application potential of UPOs.

bioinformatics↗

bioRxiv2025

Contemporary HIV-1 envelope pseudovirus panels for detecting and assessing B cell lineages with broadly neutralizing antibody potential

Although a protective HIV-1 vaccine has not yet been realized, significant progress has been made in vaccine designs that trigger B cell lineages with potential to produce broadly neutralizing antibodies (bnAbs). Advancing these strategies by optimizing vaccine boosting regimens requires early detection of maturing antibodies with neutralizing activity against native envelope glycoprotein (Env) trimers and streamlined strategies to identify antibodies as they begin to manifest desired levels of breadth and potency. Thus, we designed three types of pseudovirus screening panels based on Envs of contemporary HIV-1 isolates to facilitate detection of bnAb lineages that are on favorable trajectories during a vaccination course. The panels were selected from Tier 2 Transmitted Founder Lineage (TFL) HIV-1 Envs from placebo participants in the Antibody Mediated Prevention (AMP) efficacy trials. Using 15 bnAbs to evaluate the neutralization sensitivity of the viruses, we selected 8-member bnAb class-specific panels most sensitive to bAbs representing their class: V2-apex, V3-glycan, CD4-receptor binding site (CD4bs), Membrane-Proximal External Region (MPER), or fusion peptide (FP). Next, we combined the most sensitive viruses among the class-specific panels to create a 12-virus panel to enable optimal detection of low-titer bnAb activity across epitope specificities. Finally, as HIV-1 continues to evolve greater levels of antigenic diversity and as current global pseudoviruses bnAb panels rely on viruses collected more than twenty years ago, we showed the importance of using contemporary viral panels to assess bnAb breadth and potency and designed a 12-virus panel representative of the spectrum neutralization profiles among AMP placebo viruses. We characterized pseudoviruses bearing each selected Env using standardized human sera to confirm their Tier 2 status and biological relevance. These updated panels enable sensitive screening of neutralization activity in vaccine studies and can also provide a realistic assessment of the expected breadth and potency of maturing responses against contemporary HIV-1 Envs. AUTHOR SUMMARYA primary goal of HIV-1 vaccine research is to elicit neutralizing antibodies that can prevent infection, but HIV-1 is highly variable and so stimulating responses that can work effectively against the diverse array of HIV-1 variants remains an unsolved challenge. However, the HIV-1 vaccine field has made significant progress in terms of stimulating B cell lineages that produce antibodies known to have features able to ultimately generate potent broadly neutralizing antibodies. Inducing such precursor B cells is just the first step, as B cells need evolve in response to an antigenic stimulus through a process called affinity maturation to acquire potency and breadth against the natural diversity of circulating HIV-1 strains. The next step is to discover and optimize vaccine strategies that will selectively induce such maturation. To facilitate this work, we have designed panels of biologically relevant, contemporary HIV-1 Envelope pseudotyped viruses that will enable sensitive detection of neutralizing antibodies in human and animal vaccine studies. We also have defined a small screening panel intended to provide a quick but realistic assessment of the potential of a vaccine-stimulated antibody response or a monoclonal antibody being evaluated for clinical use to effectively counter and block currently circulating HIV-1 variants.

immunology↗

bioRxiv2025

Disruption of ADAMTSL4 Causes Ectopia Pupillae in Zebrafish via COL8A1-Driven Cell Migration

Congenital ectopia lentis (EL) poses a significant threat to visual outcomes in children. Although emerging causative genes have been identified, their biological functions remain poorly understood. In this study, we investigated the contribution of ADAMTSL4 variants to congenital EL in a cohort of 702 pedigrees. Pathogenic variants in ADAMTSL4 were identified in 23 probands (3.28%), comprising 29 distinct variants, with 14 being novel, making ADAMTSL4 the second most frequently mutated gene in the cohort. All affected individuals exhibited EL, and 26.09% also presented with ectopia pupillae (EP), also referred to as ectopia lentis et pupillae. To explore the functional consequences of ADAMTSL4 deficiency, we generated a CRISPR/Cas9-mediated adamtsl4-knockout zebrafish model, that faithfully recapitulated cardinal human disease features with an incidence comparable to that observed in affected patients. Histological and ultrastructural analysis revealed disrupted zonular fiber anchorage at the lens capsule, even in mutated eyes without overt EL or EP. Transgenic overexpression of adamtsl4 successfully reversed the ocular phenotypes, confirming the genes essential role in ocular development. Single-cell RNA-sequencing and fluorescence in situ hybridization demonstrated enriched ADAMTSL4/adamtsl4 expression in the equatorial lens epithelium, retinal pigment epithelium (RPE), iris anterior pigmented epithelium, and choroid fibroblast. Functional assays using zebrafish and human RPEs revealed that ADAMTSL4 deficiency compromised cell adhesion and promoted cell migration. Transcriptomic profiling revealed significant enrichment of extracellular matrix organization and cell adhesion pathways, with cross-species validation identifying consistent upregulation of COL8A1/col8a1b. Notably, COL8A1 knockdown in ADAMTSL4-deficient RPEs partially reversed the aberrant migratory phenotype, suggesting a functional interaction. Together, these findings establish ADAMTSL4 as a major causative gene in ectopia lentis et pupillae, highlight its role in orchestrating ocular integrity via regulation of extracellular matrix and cell behavior.

genetics↗

bioRxiv2025

Computational Discovery of Core Protein - Targeted Therapeutics Against Ginger Wilt Pathogen Ralstonia solanacearum

Bacterial wilt caused by Ralstonia solanacearum poses a severe threat to ginger production worldwide, as no consistently effective control measures exist. Here, we present a comprehensive strategy integrating pangenome analysis, functional enrichment, molecular docking, and state-of-the-art protein structure prediction to identify antibacterial agents targeting this phytopathogen. By prioritizing core, evolutionarily conserved proteins in R. solanacearum, we utilized AutoDock Vina and the recently developed AlphaFold3 to assess the binding affinity and validate protein-ligand interactions of key candidates from commercially available antibacterial pesticides. In vitro antibacterial assays confirmed strong inhibitory effects for several compounds, notably streptomycin sulfate, phellodendrine chloride, chlorogenic acid, and allicin. Our results highlight the power of combining large-scale computational screening with experimental validation for accelerating the discovery of novel antibacterial agents against highly divergent, recalcitrant plant pathogens. ImportanceThe indiscriminate use of broad-spectrum agrochemicals causes pollution and antimicrobial resistance. Our study proposes a targeted strategy against Ralstonia solanacearum by focusing on essential core proteins. This approach aims to develop specific agents that minimize harm to soil microbiota and reduce resistance selection. Our in silico framework aligns with green chemistry, offering a sustainable path for crop protection with a lower environmental footprint.

bioinformatics↗

bioRxiv2025

Longitudinal, simultaneous wide-field fluorescent Ca2+ imaging and fMRI in awake mice

Functional magnetic resonance imaging (fMRI) can be applied in mice and humans making it a key technology in translational neuroimaging research. Yet, most fMRI studies in rodents use anesthesia to limit subject motion and stress. This puts a hard boundary on the range of brain and behavioral states that can be studied in animals, and deviates from the near-universal practice of imaging people whilst awake. Recent years have seen a push towards the development of acclimation protocols for performing fMRI in mice without anesthesia, but the results have been mixed. In parallel, imaging mice without anesthesia has become routine for complementary neuroimaging methods including wide-field fluorescence calcium (WF-Ca2+) imaging. Our group works with a unique co-implementation of WF-Ca2+ and fMRI which enables the collection of complementary measures of brain function, but, until now, has necessitated the use of anesthesia. We present the first longitudinal protocol for simultaneous WF-Ca2+ and fMRI in awake mice. Our approach is comprised of a two-phase (biphasic) acclimation protocol that results in high-quality multimodal data. Comparisons of cortex-wide neuronal activity (WF-Ca2+ imaging) and blood-oxygen-level dependent fMRI signals between isoflurane-anesthetized and awake mice reveal both convergent and divergent patterns in brain function between modalities, and across time.

neuroscience↗

bioRxiv2025

Is the whole more than the sum of its parts? Considering global and local features of the connectome improves prediction of individuals and phenotype

Popular methods for analyzing the brains functional connectome examine statistical associations between pairs of atlas-defined brain regions, viewing the strength of these links as independent values. However, edges within a standard connectivity matrix, i.e., correlations between individual regions or nodes, are not independent. They are part of an interconnected system. Here, we propose that consideration of both independent, linear relationships (as in standard approaches such as linear kernel ridge regression and connectome-based predictive modeling) as well as higher order statistical associations - such as tertiary interactions between matrix components and global features of the matrix space - will enhance identification of meaningful individual differences. To test this, we adopt a geometrically grounded measure of similarity that accounts for higher-order local statistical relationships and global interactions, the Wasserstein metric. Results indicate that considering connectivity matrices as representations of their associated Gaussian distributions significantly improves both identification of individuals based on their connectivity matrices (aka, fingerprinting) and prediction of individual differences in phenotypes such as fluid intelligence and openness to experience. Thus, both pairwise local and global brain connectivity properties encode for meaningful individual differences that relate to phenotypic expressions and should be considered in brain-behavior predictive models.

neuroscience↗

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

Transdiagnostic connectome-based predictive modeling of many behavioral phenotypes reveals brain network mediators of clinical-cognitive relationships

A key assumption of the NIMHs RDoC framework is that disordered circuits in the brain should manifest in observable behaviors, including psychiatric symptomatology and cognitive deficits. However, how disordered circuitry impacts multiple behaviors remains poorly understood. Connectome-based predictive modeling (CPM) applied to functional MRI connectivity data can identify networks associated with specific behavioral measures across individuals. Prediction strength reflects how closely a measure relates to network connectivity, while derived networks provide evidence of where an individuals disordered circuits are located. Using CPM, we predicted a broad range of self-reported clinical and objective cognitive measures in a large, transdiagnostic sample with extensive fMRI data (n = 317). Prediction performance varied substantially across instruments, with objective cognitive tests yielding stronger models than self-reported clinical measures (p < 0.001). To test whether circuits underlying cognitive deficits related to symptomatology reside in regions where networks overlap, we examined the prediction strength of these sparsely shared circuits. Their connectivity strongly predicted cognitive performance and were primarily localized within the frontoparietal network and between the frontoparietal and default mode networks. These findings reveal how much various behavioral measures reflect brain networks and how circuits within the shared network space contribute to cognitive deficits associated with symptomatology.

neuroscience↗

bioRxiv2025

Mapping Human Survivability at Extreme Wet-Bulb Temperatures 32-35°C

O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=146 SRC="FIGDIR/small/677706v1_ufig1.gif" ALT="Figure 1"> View larger version (52K): org.highwire.dtl.DTLVardef@70b72dorg.highwire.dtl.DTLVardef@9fbfeorg.highwire.dtl.DTLVardef@17f05a5org.highwire.dtl.DTLVardef@4b977b_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstractC_FLOATNO C_FIG HighlightsWe report the first empirical human survival map across wet-bulb temperatures of 32-35{degrees}C. HydraAon and behavior cooling extend survival beyond the 6-hour theoreAcal limit. Females consistently tolerate extreme humid heat longer than males in controlled trials. Our findings refine heat-health models and inform equitable climate adaptaAon strategies. In briefExtreme humid heat, measured by wet-bulb temperature (Tw), is nearing the threshold of human survivability under climate change. In controlled trials with 36 adults, we provide the first empirical survival map across Tw=32-35{degrees}C, showing that hydration and behavioral adaptations extend endurance beyond the theoretical 6-hour limit, with women tolerating heat longer than men. These findings refine human heat tolerance thresholds and inform equitable climate adaptation and heat-health strategies. SCIENCE FOR SOCIETYAs climate change accelerates, extreme humid heat events--measured by wet-bulb temperature (Tw)--pose a growing threat to human survival. This study provides the first empirical benchmarks of human survivability across Tw=32-35{degrees}C, clarifying how long healthy adults can endure these conditions under full hydration and minimal activity. We find that while the theoretical 6-hour survival limit at Tw=35{degrees}C is partly supported, both men and women can withstand these conditions for 7-8 hours when hydrated, with women demonstrating greater tolerance. At Tw=34{degrees}C, projected heatstroke onset extends beyond 12-16 hours, while conditions at Tw=32-33{degrees}C remain largely compensable for more than 30 hours. These results challenge prevailing heat tolerance models, underscore the role of behavioral and physiological adaptations, and provide critical data for designing equitable heat-health policies and early warning systems that account for sex differences and vulnerable populations. Climate change-driven extreme heat events increasingly threaten human health. Here, we provide the first empirical map of human survivability across four extreme wet-bulb temperatures (Tw=32- 35{degrees}C) in controlled trials. Thirty-six healthy young adults (20 males, 16 females) were exposed to twelve dry-bulb temperature and relative humidity combinations in shaded indoor settings under full hydration. Participants remained seated and performed light office tasks until core temperature (Tcore) reached 39{degrees}C or exposures lasted 8 hours. Clinical heatstroke threshold (Tcore=40.5{degrees}C) were projected by extrapolating from our direct Tcore measurements during controlled exposure trials. At Tw=35{degrees}C, males and females reached heatstroke in approximately 7.1-7.7 and 8.3-8.6 hours, respectively--partly supporting the 6-hour theoretical limit while revealing enhanced short-duration resilience when hydrated. At Tw=34{degrees}C, projected times are 12.5-12.9 hours (males) and 15.5-16.2 hours (females). Conditions at Tw=32{degrees}C and 33{degrees}C remained (quasi-)compensable, with projected heatstroke onset after 33-35 hours. Females consistently tolerated heat stress longer than males. These physiological benchmarks are critical for improving heat-health models and embedding sex-specific vulnerability in equitable adaptation strategies amid accelerating global warming.

physiology↗

bioRxiv2025

Body temperature drives azole tolerance in Candida albicans by hindering the autophagic degradation of Erg11

Human body temperature has been shown to limit the antifungal efficacy of azoles; however, the precise mechanisms underlying this limitation remain insufficiently understood. In this study, we observed that short-term exposure (24 h) to human body temperature (37{square}{degrees}C), in contrast to 30{square}{degrees}C, significantly enhances the tolerance of Candida albicans to azoles without inducing resistance. Our findings suggest that this increased tolerance is due to a reduction in the degradation of C. albicans Erg11 at 37{square}{degrees}C compared to 30{square}{degrees}C, thereby promoting azole tolerance. This phenomenon occurs because Erg11 degradation is mediated by autophagy, which is inhibited in C. albicans at 37{square}{degrees}C. The suppression of autophagy at this temperature may be associated with elevated mitochondrial production of reactive oxygen species, leading to mitochondrial dysfunction. In response to heat stress-induced mitochondrial dysfunction, autophagy-related proteins such as Atg8 accumulate around the mitochondria, while their levels in the endoplasmic reticulum decrease, consequently inhibiting autophagy in C. albicans. These findings suggest that promoting the degradation of Erg11 in C. albicans may serve as a viable strategy to augment the fungicidal efficacy of azoles in the human host.

microbiology↗

bioRxiv2025

Dissecting the epigenome dynamics in human immune cells upon viral and chemical exposure by multimodal single-cell profiling

Environmental and pathogen exposure can lead to profound remodeling of the gene-regulatory landscape across human immune cell populations. Here, we present a single-nucleus chromatin accessibility atlas of human immune cells comprising 271,299 cells and 319,420 candidate regulatory elements from individuals exposed to HIV-1, COVID-19, Influenza virus, organophosphates, as well as healthy controls. Our longitudinal HIV cohort reveals epigenetic signatures of T cell exhaustion manifested in changes in the accessibility at binding sites for the FOXP family transcription factors. Upon severe exposure to SARS-Cov2 we identified changes in chromatin accessibility manifesting in a regulatory switch in cytokine networks characterized by the downregulation of the NF-{kappa}B motif family, alongside concordant changes in AP-1 and IRF factor networks in CD14+ classical monocytes. By integrating single-cell profiles of DNA methylation from matched samples, we created a multimodal epigenome atlas of human immune cell states using the accessibility-derived candidate regulatory elements. Both modalities exhibit complementary epigenetic signatures at transcription factor binding sites associated with cell state, as exemplified in the process of memory formation in T-cells, where BATF, AP-1, and ETS motifs exhibit significant epigenetic covariance across both epigenomic layers. Finally, by linking potentially regulatory DNA methylation signatures to changes in chromatin accessibility in monocytes, we observe that severe COVID-19 involves selective, multiomics remodeling of epigenetic profiles at TF binding sites manifested in concordant DNA methylation and accessibility dynamics at inflammation-associated regulatory TFs.

bioinformatics↗

bioRxiv2025

Leveraging Large Language Models for Redundancy-Aware Pathway Analysis and Deep Biological Interpretation

Extracting coherent, biologically meaningful insights from vast, complex multi-omics data remains challenging. Currently, pathway enrichment analysis serves as a cornerstone for the functional interpretation of such data. However, conventional approaches often suffer from extensive functional redundancy caused by shared molecular components and overlapping pathway definitions across databases. This redundancy can obscure key biological signals and compromise the interpretability of pathway enrichment results. Here, we present MAPA (Functional Module Identification and Annotation for Pathway Analysis Results Using Large Language Models [LLM]), an open-source computational framework that resolves redundancy and enhances pathway analysis result interpretation. MAPA computes functional similarity between pathways using LLM-based text embeddings, enabling comparison across different databases. It constructs pathway similarity networks and identifies functional modules via community detection algorithms. Crucially, MAPA employs LLMs for automated functional annotation, integrating Retrieval-Augmented Generation (RAG) to generate comprehensive and real-time biological summaries and reduce hallucinations. Benchmarking demonstrated MAPAs superior performance: the biotext embedding similarity showed a large effect size (Cliffs {delta} = 0.96) compared with the Jaccard index ({delta} = 0.73), and module identification achieved high accuracy (Adjusted Rand Index [ARI] = 0.95) versus existing methods (ARI = 0.23-0.33). Human expert evaluation confirmed that MAPAs annotations match expert-quality interpretations. Finally, a multi-omics aging case study illustrates that MAPA uncovers coherent functional modules and generates insights extending beyond conventional pathway analyses. Collectively, MAPA represents a significant advance in redundancy-aware pathway analysis, transforming pathway enrichment results from fragmented lists into biologically coherent narratives. By leveraging the capabilities of LLMs, MAPA offers researchers a robust, scalable tool for deriving deep mechanistic insights from complex and vast multi-omics datasets, marking a new direction for AI-driven bioinformatics.

bioinformatics↗