Biology subjects

Ghosh, S.

Publications and source records attributed to Ghosh, S..

At least 73 records · Page 4Linked to original sources

bioRxiv2025

Tuning the H2 Production Activity of ToHydA by Molecular Simulation-Informed Protein Engineering

In the context of bio-hydrogen production, the O2-stable Group B [FeFe]-hydrogenase from Thermosediminibacter oceani has attracted significant interest due to its distinctive "TSCCCP" motif near the active site, which contains an additional (third) cysteine residue that is absent in the "TSCCP" motif of standard Group A hydrogenases. The precise role of the additional cysteine residue in H2 production has remained an open question. In this study, we sought to contribute to the understanding of this cysteines role in H2 production by combining molecular dynamics (MD) simulations, site-directed mutagenesis, biochemical assays, and Fourier-transform infrared (FTIR) spectroscopy. Remarkably, a cysteine-to-serine exchange variant (TSSCCP) demonstrated enhanced H2 production activity without compromising the O2-stability of ToHydA, offering new insights into its functional dynamics.

biophysics↗

bioRxiv2025

Multiomics Integration Reveals a Metabolic Myopathy in Cardiometabolic HFpEF

BackgroundSkeletal muscle dysfunction is a major peripheral determinant of exercise intolerance and physical disability in heart failure with preserved ejection fraction (HFpEF). Metabolic and mitochondrial dysfunction are considered to be key components of skeletal muscle dysfunction, but comprehensive profiling of metabolic pathways has not been conducted. Elucidation of dysregulated metabolic pathways is essential to determine viable targets for the treatment of exercise intolerance in HFpEF. MethodsMale ZSF1 Obese rats (HFpEF) and Wistar Kyoto (WKY) lean normotensive controls were studied at 26 weeks of age. Gastrocnemius was subjected to bulk RNA-seq, proteomics, metabolomics, and lipidomics analysis. The R package limma was used to determine differential expression in all omics layers (absolute fold-change>1.5, FDR0.05, unless otherwise indicated). Additional targeted plasma and skeletal muscle (soleus and EDL) metabolomics and lipidomics were performed on HFpEF and control rats. ResultsPathway level analysis for RNA seq and proteomics revealed significant downregulation of oxidative phosphorylation (NES -2.1, p<0.005), electron transport chain (NES -2.0, p<0.005), and TCA cycle (-1.8, p<0.05). The most upregulated pathways were PPAR signaling (NES 2.2, p<0.0001), tryptophan metabolism (NES 1.8, P<0.005), and amino acid oxidation (NES 1.8, p<0.005) pathways. Metabolomics revealed an accumulation of TCA cycle intermediate, isocitrate, and phosphate reduction. Branched-chain amino acids were significantly increased, whereas amino acids related to tryptophan metabolism were reduced and shifted towards increased serotonin accumulation. Phospholipid species were differentially regulated with increased palmitoylated phosphatidylcholines but reduced arachidonoyl-PC species. Phosphatidylethanolamines (PE) species (16:0/16:1-18:0/18:2) were increased. ConclusionOur multiomics analysis of skeletal muscle in HFpEF revealed severe mitochondrial dysfunction that was characterized by reduced complex I and II activity. Mitochondrial and peroxisomal lipid overload results in a shift in membrane phospholipid accumulation and composition. Reduced BCAA oxidation and dysregulation of tryptophan metabolism are key features of amino acid metabolism that reduce anaplerosis and promote the accumulation of toxic metabolites. Comparative analysis of other skeletal muscle disorders suggests that an acquired metabolic myopathy exists in cardiometabolic HFpEF.

physiology↗

bioRxiv2025

Mechanoepigenetic Targeting of Histone Methyltransferase EZH2 Increases Potential of Scalable Manufacturing in Human Derived Primary Mesenchymal Stromal Cells

Manufacturing clinical grade mesenchymal stromal cells (MSCs) remains a major bottleneck for cell-based therapies, as extensive in-vitro expansion on standard tissue-culture plastic (TCP) drives loss of stemness, reduced immunomodulatory activity, and diminished therapeutic efficacy. Although substrate stiffness is known to influence MSC fate through mechanotransduction, the epigenetic mechanisms linking mechanical stress to progressive phenotypic drift remain poorly defined. A mechano-epigenetic pathway is identified in this work, centered on the histone methyltransferase EZH2 - that governs chromatin remodeling and loss of stemness during serial passaging. Multi-omics, high-resolution imaging, and functional assays show that MSCs expanded on mechanically stiff TCP accumulate H3K27me3 repressive chromatin mark, lose SWI/SNF -ARID1A chromatin- remodeling foci, and exhibit an altered chromatin accessibility profile. Pharmacological inhibition of EZH2 with GSK343 selectively reduced H3K27me3, restored ARID1A-containing SWI/SNF organization, and preserved MSC morphology and expression of canonical stemness markers (CD73, CD90, CD105) even at later passage. ATAC-seq analysis revealed that GSK343 rebalanced chromatin accessibility, reopening TEAD/YAP-responsive regulatory regions while repressing accessibility at lineage- priming and senescence-associated sites. RNA-seq demonstrated that GSK343 maintained transcriptional programs associated with immunomodulation, migration, and trophic signaling, while suppressing hyperproliferative and senescence-associated genes that are characteristic of late-passage MSCs. Proteomic profiling of MSC secretomes further showed that GSK343 attenuated pro-fibrotic ECM factors and senescence- linked proteins while enhancing angiogenic and reparative mediators. Functionally, conditioned media from GSK343-treated MSCs significantly increased primary chondrocyte proliferation, demonstrating preserved therapeutic potency. Together, these findings establish EZH2 as a central mediator of stiffness-induced epigenetic drift in human MSCs and demonstrate that EZH2 inhibition can maintain the stemness without compromising their expansion ability. This work provides a foundational strategy for mechano-epigenetic engineering of MSCs and highlights EZH2 inhibition as a scalable, manufacturing-compatible approach to preserve potency for regenerative medicine applications. SIGNIFIANCE STATEMENTPrimary mesenchymal stem/ stromal cells (MSC) derived from bone marrow are currently under numerous clinical trials for regenerative medicine and cancer treatment. Despite showing tremendous potential for their differentiation, trophic and immunomodulatory properties, their efficacy in clinical endpoint is limited. The traditional cell culture condition in cell culture flask limits their clinical translation by changing the MSC phenotype with serial passaging, a critical step required for harvesting a clinically appropriate number of cells. This work discovered the precise epigenetic mechanism involving histone methylation responsible for such changes in MSC phenotype. By inhibiting the histone methylation by GSK343, this study showed that MSC phenotype can be maintained over serial passaging as demonstrated by microscopy, multi-omics, and in vitro functional assays.

bioengineering↗

bioRxiv2025

Distinguishing near- versus off-critical phase behaviors of intrinsically disordered proteins

Intrinsically disordered prion-like low complexity domains (PLCDs) drive phase transitions that underlie the biogenesis of different biomolecular condensates. The mapping of critical points is essential for generating quantitative assessments of driving forces and for distinguishing phase behaviors in the critical versus off-critical regimes. Computations play an important role in connecting the molecular-scale interactions to mesoscale phase behaviors of PLCDs and other intrinsically disordered proteins. We report results from accurate mapping of the critical regime for an archetypal PLCD. This is achieved by combining large-scale simulations with computations of Binder cumulants and the use of rigorous finite-size scaling approaches. The computed binodal, defined by knowledge of the critical point and intersection of the left arm of the binodal by the overlap and percolation lines, can be demarcated into three distinct regimes. Regime I is farthest from the critical point, with the coexisting dilute phase being akin to a gas of dispersed polymers. Regime II lies above the intersection of the overlap line and the dilute arm of the binodal. Here, coexisting dilute phases are characterized by heterogeneous cluster-size distributions with heavy tails. In Regimes I and II, dense phases are confined percolated networks. Regime III is closest to the critical point. Here, the dense phase becomes unconfined and the percolated network swells to become system-spanning. Thus, Regime III comprises two interconnected, system-spanning networks. In addition to accurately mapping the critical point, we also evaluated methods for identifying the theta temperature. We find that conventional scaling approaches lead to erroneously low estimates of the theta temperature. Instead, accurate estimation of the theta temperature requires direct calculation of the temperature dependence of the two-body interaction coefficient. We discuss the broader implications of these findings for inferring solvent quality from scaling analyses.

biophysics↗

bioRxiv2025

Intranigral injection of Alpha-Synuclein pre-formed fibrils leads to BBB compromise and Bilateral Dopaminergic Neurodegeneration in A53T Alpha-Synuclein transgenic mice.

Parkinsons disease (PD) is a progressive neurodegenerative disorder characterized by alpha-()-Synuclein neuronal aggregation and loss of dopaminergic (DA) neurons. Developing animal models that replicate PDs neuropathological phenotypes is critical for understanding its pathophysiology and evaluating potential therapeutic targets. In this study, we show that direct unilateral injection of human -Synuclein PFFs into the Substantia Nigra (SN) of mutant A53T -synuclein overexpressing mice induce bilateral phosphorylated -Synuclein (pS129) pathology in the SN. This pathology spreads to the striatum, cerebral cortex, and midbrain within 60 days and is accompanied by neuroinflammation in the midbrain and cerebral cortex. Additionally, we observed synuclein-dependent neurodegeneration, with a 50% reduction in Tyrosine Hydroxylase (TH) intensity in the SN and a 40% reduction in Striatum, both bilaterally. The model also revealed a compromised blood-brain barrier (BBB) and T-cell infiltration in the PFF injected animals, correlating with pS129 pathology and neuroinflammation. Taken together, we developed a mouse model that recapitulates multiple PD phenotypes, providing a valuable platform for testing therapeutic strategies targeting human -Synuclein pathology and for exploring CNS-peripheral immune interactions in PD. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=135 SRC="FIGDIR/small/690780v2_ufig1.gif" ALT="Figure 1"> View larger version (20K): org.highwire.dtl.DTLVardef@19c6234org.highwire.dtl.DTLVardef@1f36bbaorg.highwire.dtl.DTLVardef@25ac6borg.highwire.dtl.DTLVardef@15ad61d_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2025

A non-canonical AKT1-TERT pathway coordinates autophagy and ERphagy

The retinal pigment epithelium (RPE) sustains lifelong proteostasis under chronic stress, yet how post-mitotic cells activate autophagy when canonical kinase pathways suppress it remains unresolved. Here we demonstrate that AKT1, conventionally regarded as an autophagy suppressor, drives autophagy through a non-canonical transcriptional mechanism triggered by isoform imbalance. In RPE and age-related macular degeneration (AMD) models, AKT2 hyperactivation destabilizes mTORC2 and engages S6K-mediated IRS-1 inhibition, creating a feedforward autophagic arrest. Compensatory AKT1 activation via mTORC2 phosphorylates telomerase reverse transcriptase (TERT) at Serine 824, driving nuclear translocation that is independent of telomere maintenance. In the nucleus, TERT assembles with FOXO3 and MYC into a tripartite transcriptional complex that occupies the EIF2AK3 promoter, enabling PERK transcriptional activation. This activity converts the unfolded protein response from pro-apoptotic to cytoprotective: PERK-ATF4 signaling drives biogenesis of core autophagy machinery while simultaneously inducing selective ERphagy through the receptors TEX264 and CCPG1, which prevents pathological PERK clustering and preserves tubular ER architecture in diseased RPE. Using C. elegans phosphorylation-deficient mutants, mouse models and AMD patient induced pluripotent stem cell-derived retinal pigment epithelium (iPSC-RPE), we establish that AKT1-mediated TERT phosphorylation is an evolutionarily conserved prerequisite for FOXO/DAF-16 nuclear function and lysosomal homeostasis in post-mitotic cells. Pharmacologic targeting of AKT2 with a first-in-class dual-pocket allosteric inhibitor selectively enhances AKT1 compensation, restoring macroautophagic and ERphagy flux across disease models. These findings reveal a kinase-to-transcription axis that reprograms organelle quality control and identify the AKT1-TERT-PERK-ATF4 pathway as a therapeutic target in proteostasis-driven disease.

cell biology↗

bioRxiv2025

Integrated host phage RNA seq analysis of Mycobacteriophage D29 infection reveals dual arm balanced phage genome expression and inhibited expression of host transporter and VapC target genes.

The Mycobacteriophage D29 infects Mycobacterium species. Many of these are deadly human pathogens. In earlier studies, we performed proteomic analyses using a temperature-sensitive variant of D29, called D29ts12. This work indicated that a regulatory mechanism controls the expression of the phages right arm genes. In this study, we executed a transcriptomic analysis. We gained further insight into how phage infection affects gene expression, both in its genome and in the host genome. We identified several novel non-coding RNA contigs spanning the cos site junction. This finding indicates that this region is expressed. In addition, we observed two areas of intense transcriptional activity. One is located downstream of Pleft, in the right arm, and coincides with a major noncoding RNA species. This RNA species is essential for lytic growth, as reported earlier for mycobacteriophages of the A2 family. The other area of transcriptional activity was gene 17, which encodes the major coat protein. By comparing TPM values, we found that thermal inactivation of phage D29ts12 growth significantly decreased right-arm gene expression. Interestingly, an accompanying sharp increase in left-arm gene expression was observed. Therefore, a mechanism ensures balanced expression from the two arms. We also analysed the impact of phage infection on host gene expression. The results indicate significant changes in the expression levels of genes encoding transporters and genes regulated by the toxin VapC. This investigation is the first to use a mycobacteriophage to examine the combined transcriptome of the phage and its host.

microbiology↗

bioRxiv2025

Matrix Stiffness Dictates Doxorubicin-Induced Apoptosis by Modulating Cell-Cycle State in HeLa Cells

Drug resistance remains a major challenge in cancer treatment by contributing to recurrence and metastasis. Fractional killing, in which only a subset of cells undergo apoptosis after drug exposure, is a key contributor to this resistance and is influenced by genetic and nongenetic heterogeneity within the tumor microenvironment. Solid tumors display substantial variation in extracellular matrix stiffness, providing evidence that the mechanical context of cancer and stromal cells may play an important role in therapeutic response. Here, we investigated how substrate stiffness affects the dynamics of apoptosis and the mechanisms behind differences in the cell death response to doxorubicin (DOX). HeLa cells cultured on stiffer substrates exhibited enhanced caspase-3/7 activation and increased apoptotic cell death, whereas cells on soft substrates showed markedly reduced apoptotic signaling and improved survival. Although substrate stiffness altered cytoskeletal organization, pharmacological disruption of actin polymerization or actomyosin contractility did not influence nuclear DOX accumulation, indicating that cytoskeletal mechanics were not the primary factor in the stiffness-dependent sensitivity. Instead, flow cytometry revealed that substrate stiffness modulates cell-cycle distribution, with soft substrates enriched in the G1 population and a reduced fraction of cells in the DOX-sensitive S phase. Synchronizing cells at the G1/S phase boundary eliminated stiffness-dependent differences in apoptotic activation, demonstrating that cell-cycle state is a dominant driver of stiffness-mediated fractional killing. These findings highlight a mechanistic link between extracellular matrix mechanics and chemotherapeutic response by suggesting that microenvironment-regulated cell-cycle dynamics contribute to drug resistance in mechanically heterogeneous tumors.

biophysics↗

bioRxiv2025

PKMζ-PKC{iota}/{lambda} double-knockout reveals the atypical PKCs are crucial for hippocampal late-LTP and spatial long-term memory

PKM{zeta}, a persistently active atypical PKC (aPKC) isoform, is thought to maintain late-phase long-term potentiation (late-LTP) and long-term memory. However, PKM{zeta}-knockout mice still exhibit hippocampal LTP and spatial memory while lacking neocortical LTP, questioning whether this kinase is fundamental to enduring synaptic potentiation and memory. Tsokas et al. (2016) showed the other aPKC, PKC{iota}/{lambda}, likely compensates for PKM{zeta} during maintenance in the hippocampus of PKM{zeta}-null mice. In wild-type mice, PKC{iota}/{lambda} drives early-LTP and short-term memory, while PKM{zeta} compensates for PKC{iota}/{lambda} knockout by supporting both early- and late-phase processes. Here we show PKC{iota}/{lambda} persistently increases during maintenance in two mouse models: PKM{zeta}-conditional knockout (cKO) mice, and double-knockout mice carrying both conditional deletion of PKC{iota}/{lambda} and constitutive loss of PKM{zeta}. In the double-knockout mice, PKC{iota}/{lambda} was measured while the kinase was still present, prior to its inducible ablation, to characterize its compensatory upregulation in late-LTP before removal. To test whether this compensation was functional, we ablated PKC{iota}/{lambda} in the hippocampus of the double-knockout mice. The double-knockout eliminated late-LTP, whereas individual knockout of either aPKC alone showed normal-appearing LTP. Double-knockout also abolished spatial long-term memory without affecting short-term memory. Thus, when PKM{zeta} is absent, PKC{iota}/{lambda} persists to maintain hippocampal late-LTP and long-term memory.

neuroscience↗

bioRxiv2025

CONCERT predicts niche-aware perturbation responses in spatial transcriptomics

Spatial perturbation transcriptomics measures how genetic or chemical edits alter gene expression while preserving tissue context. Perturbation outcomes depend on a cells intrinsic state and also on how effects propagate across cellular microenvironments. We present CONCERT, a niche-aware generative model that embeds perturbation context and learns spatial kernels with a Gaussian process variational autoencoder to predict perturbation effects across tissue. We formalize three tasks: patch, border, and niche, predicting responses in nearby unperturbed regions, at tissue interfaces, and as a function of surrounding microenvironments. We evaluate CONCERT on Perturb-map lung datasets. CONCERT outperforms state-of-the-art models (dissociated counterfactuals, spatialized perturbation models, and kNN), reducing E-distance by up to 33.77% (patch), 26.05% (border), and 33.74% (niche) versus the next best, with mean absolute error down by up to 23.28% and Pearson correlation up by up to 9.10%. Two case studies go beyond benchmarking. In dextran sodium sulfate-induced colitis, CONCERT reconstructs spatial gene expression at unmeasured time points, produces longitudinal comparisons across unpaired mice, resolves intermouse heterogeneity, and recovers consistent temporal declines of inflammation-associated genes across regions. In ischemic stroke, CONCERT predicts responses under variable lesion sizes and in a 3D formulation across brain sections, capturing lesion-core and peri-lesion patterns. CONCERT performs niche-aware counterfactual prediction, reconstructs missing spatial data, and models perturbation responses across tissues.

bioinformatics↗

bioRxiv2025

Gene editing in "cell villages" enables exploring disease-relevant mutations in many genetic backgrounds

Understanding how individual genetic backgrounds shape the effects of disease-associated mutations is central to elucidating the biology of complex psychiatric disorders. We developed a scalable village editing strategy that enables simultaneous genome editing across multiple induced pluripotent stem cell (iPSC) lines, allowing systematic assessment of how polygenic context modulates the impact of specific mutations. Using pooled CRISPR editing in 15 iPSC lines spanning a range of schizophrenia (SCZ) polygenic risk scores, we generated homozygous and heterozygous knockouts in two known SCZ-associated genes: LRP1, involved in cholesterol import, and NRXN1, a presynaptic adhesion molecule. By mixing all lines prior to editing and de-multiplexing them afterward, we efficiently produced multi-donor knockout neurons at scale. Transcriptomic profiling revealed that LRP1 and NRXN1 loss produce both shared and donor-specific effects on neuronal gene expression, with variable perturbation of neurotransmitter transport and cholesterol biosynthesis pathways across genetic backgrounds. These results demonstrate that village editing enables systematic dissection of gene-background interactions in human neurons, offering a powerful framework for studying the polygenic architecture of psychiatric disease.

cell biology↗

bioRxiv2025

An induced pluripotent stem cell-based chemical genetic approach for studying spinal muscular atrophy

Spinal muscular atrophy (SMA) is a genetic disease characterized by degeneration of spinal cord motor neurons and neuromuscular junctions. Despite recent development in therapies for SMA, treatment efficacy largely relies on administration of drugs early in disease progression and is impacted by underlying patient genetics. Drug discovery for other diseases of the central nervous system (CNS) has also been hindered by heterogeneity in patient genetics and clinical presentations, as well as the need for early intervention. To address these hurdles, we utilized a chemical genetic-based screening approach to adapt the Connectivity Map (CMAP)/L1000 platform to study SMA. To do this, we differentiated moderate and severe SMA patient-specific induced pluripotent stem cells into neuronal cells utilizing a forward programming differentiation protocol, exposed each to 360 neuroactive or CNS disease-related compounds, and interrogated resulting changes in expression of >400 neural genes in a platform we term CMAPneuro. In doing so, we generated 4,559 transcriptional profiles identifying stimuli that modulate gene expression differences across SMA neurons. Finally, we make these data queryable, allowing the research community to 1.) identify CNS disease-related perturbagens that mimic or reverse differentially expressed genes, or 2.) explore the transcriptional response of a given perturbation in diverse SMA neuronal cells.

neuroscience↗

bioRxiv2025

Lipid Droplets mediated Japanese Encephalitis Virus small Extracellular Vesicle Release

Lipid droplets (LDs) and small Extracellular Vesicles (sEVs) are classically known for lipid metabolism and intercellular communication, respectively. Here, we reveal a mechanistic connection between LD dynamics and sEV-mediated non-lytic release of Japanese Encephalitis Virus (JEV) from neuronal cells. Using Neuro2A, SH-SY5Y, N9 microglia, and primary cortical neurons, we show that JEV is packaged within sEVs ([~]200 nm) through an ESCRT-independent, neutral sphingomyelinase 2 (nSMase2)/ ceramide-dependent pathway. Virions inside sEVs display a higher JEV Premature Membrane/Membrane protein (PrM/M) ratio compared to those released via the conventional secretory pathway. Although containing a higher proportion of premature virions than mature ones, sEV-associated JEV virions gain an evolutionary advantage by evading immune detection and delivering multiple virions to recipient cells, thereby increasing overall infection efficiency. Temporal profiling showed early cytoplasmic LD enrichment (from 6 hpi), followed by a surge in sEV release from 14 hpi, suggesting sequential roles for LDs and sEVs. nSMase2 inhibition decreased sEV-mediated egress without affecting viral replication, but increased cytoplasmic LD abundance, consistent with LD underutilization in multivesicular bodies (MVB) biogenesis. Our findings identify LDs as facilitators of MVB formation and nSMase2 as a key driver of sEV-mediated viral exit, revealing parallel yet coordinated pathways in JEVs stealthy egress. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=118 SRC="FIGDIR/small/686473v1_ufig1.gif" ALT="Figure 1000"> View larger version (54K): org.highwire.dtl.DTLVardef@10d28forg.highwire.dtl.DTLVardef@6fb6a0org.highwire.dtl.DTLVardef@3f0025org.highwire.dtl.DTLVardef@22b5ae_HPS_FORMAT_FIGEXP M_FIG C_FIG

microbiology↗

bioRxiv2025

Dynamic profiling of the aggresome processing pathway

Cellular homeostasis is maintained by the efficient degradation of proteins by the 26S proteasome, a key component of the ubiquitin proteasome system. When the proteasome function is impaired due to cellular stress like mutations, chemicals or epigenetic stressors, there is accumulation of unfolded or misfolded, ubiquitinated proteins leading to cellular toxicity. Cells combat this rise in cellular toxicity by sequestering ubiquinated proteins in a transient cellular structure called the aggresome. The transport of this ubquitinated protein cargo along the microtubules to form aggresomes and their clearance from cells by autophagy is called the aggresome processing pathway (APP). APP activation is a hallmark of several neurodegenerative diseases and has also been implicated in Influenza A Virus uncoating step. To study the dynamics of APP and the compositional diversity of the aggresome, we used time-dependent proximity labeling, mass spectrometry, and immunofluorescence microscopy. Here, we used vimentin (VIM) and histone deacetylase 6 (HDAC6), two established aggresome markers as baits to visualize aggresome formation and clearance in HeLa cells. We identified more than 75 novel components of mature aggresome. Further, we characterised the time-dependent protein-protein interactions in the neighbourhood of VIM and HDAC6 during APP progression. Our study identified new interactors of VIM and HDAC6 that maybe involved in aggresome clearance by autophagy. Together, our findings uncover novel APP-associated proteins that are central to advancing the understanding of this poorly characterized and highly complex pathway implicated in multiple diseases.

molecular biology↗

bioRxiv2025

Decision Voting Based Multiscale Convolutional Learning of Brain Networks With Explainability

The diagnosis of neurological disorders requires comprehensive frameworks that incorporate multimodal neuroimaging data while ensuring clinical interpretability. Recent neuroimaging research is focusing on the integration of brain structure and function to reveal some of the prominent alteration caused by a brain disorder at the system level. This work presents a fresh multiscale graph convolutional network (GCN) framework that integrates structural connectivity (SC) from diffusion tensor imaging and functional connectivity (FC) from resting-state fMRI on three different anatomical scales. We present softmax-based decision fusion for cross-modal multiscale integration in our architecture. The preprocessing pipeline improves connectivity representations by means of graph diffusion, topological sparsification, and noise augmentation. Using five-fold cross-valuation, evaluated on a schizophrenia classification dataset, our model achieves 71.59% accuracy, outperforming single-scale methods and conventional machine learning bench-marks. Explainability analysis reveals that different dysconnectivity patterns in schizophrenia patients overlapping with biomarkers reported in literature. The multiscale approach shows complementing insights: coarse scales capture global network changes while finer scales identify localized subcortical disruptions. Combining diagnostic precision with biologically interpretable modeling, this work creates a new paradigm for interpretable brain network analysis.

bioengineering↗

bioRxiv2025

Pantetheinamides that inhibit the growth of intracellular Salmonella Typhimurium

The metabolic adaptability of intracellular pathogens, such as the Gram-negative Salmonella enterica serovar Typhimurium (STm), enables their survival in nutrient-restricted host environments, while also presenting an opportunity for selective antimicrobial targeting. Herein, we report the synthesis and screening of a small pantetheinamide library aimed at inhibiting the proliferation of STm within macrophages. Two lead compounds exhibited selective activity against intracellular STm and reduced bacterial burden in a murine colitis model. Mechanistic studies suggest their primary mode of action to be the inhibition of coenzyme A (CoA) biosynthesis via the PanD-PanZ regulatory complex, which is absent in host cells. Importantly, resistance pressure to our compound was shown to be significantly stronger in nutrient-limited media compared to nutrient-rich media, supporting the targeting of nutrient stress as a strategy to delay resistance. This work highlights the value of inhibiting pathogen-specific metabolic vulnerabilities in combination with host defence mechanisms to develop novel antimicrobials.

microbiology↗

bioRxiv2025

Agent SPI-WSI: In context learning for computationally spatial pathwayinferring on whole slide histopathology images conditioned on bulk RNAsequencing using pathologist in the loop.

Bulk RNA sequencing, while cost-effective compared to high-resolution spatial transcriptomics, averages gene expression across heterogeneous cell populations and thus lacks spatial context. To address this limitation, we introduce a structured, human-guided, multi-stage computational AI agent SPI-WSI, that iteratively generates, evaluates, and refines biologically informative natural-language prompts, thereby localizing bulk-derived pathway activity within histopathology slides. Our pipeline uses in-context prompting in large language models (LLMs) to adapt dynamically to the task of prompt generation. Candidate prompts are first produced by the LLM and then subjected to a secondary pathologist critiquing by the same LLM that cross-references PubMed to ensure both biological plausibility and specificity. Each approved prompt against their image tile is scored using the vision language foundation model (CONCH). We benchmarked different LLMs, Gemini 2.0, Gemini 2.5, Claude 3.7 and Claude 4.0, and found that Claude 4.0 achieves the highest cosine similarity (0.7) between image and prompt embeddings. Pathologist-driven scoring and manual segmentation confirm that our method accurately identifies clusters of spatial pathological morphologies. In addition, we have validated the method against ground-truth spatial transcriptomic spots using inhouse and public datasets. Overall, the trend emphasized by ground-truth spatial RNA sequencing prompts is closely aligned with those from bulk prompt. This pathologist-in-the-loop workflow enables large-scale, reproducible tissue profiling and grounds AI-driven spatial annotations.

bioinformatics↗

bioRxiv2025

SEASONAL PM2.5 DIFFERENTIALLY REGULATES JAK2/STAT3 SIGNALING IN RURAL AND URBAN COHORTS

Ambient particulate matter (PM2.5) is a major environmental carcinogen linked to lung cancer, yet its molecular insights on asymptomatic non-smokers remains unclear. This study examined the effect of seasonal fluctuations of PM2.5 on oxidative stress and pro-oncogenic signaling in rural (RU) and urban (UR) cohorts from West Bengal, India. Environmental monitoring revealed higher PM2.5 and associated benzo[]pyrene in UR, during winter, induced oxidative stress (elevated ROS, 8-OHdG), reduced antioxidants (SOD, catalase, GPx), and promoted airway inflammation. Transcriptomic and bioinformatic analyses identified activation of IL-6/EGFR-driven JAK2/STAT3 signaling and its crosstalk with Ras/Raf/MAPK, leading to increased expression of downstream effectors (BCL-2, MCL-1, c-MYC, cyclin D1) and repression of tumor suppressors (BAX, p21). Notably, downregulation of JAK/STAT inhibitors PIAS2 and SOCS2 suggested persistent activation of oncogenic signaling. Linear mixed-effects models linked winter PM2.5 surges to oxidative stress, inflammation, and altered JAK2/STAT3 signaling, while regression models showed stronger responses in UR. Risk modeling predicted significantly higher lung cancer mortality in UR, underscoring the role of seasonal PM2.5 surges in JAK2/STAT3-driven carcinogenic susceptibility and the urgent need for targeted interventions.

cancer biology↗