Search bioRxiv⌕ Search

Biology subjects

Gopinath, A.

Publications and source records attributed to Gopinath, A..

At least 19 recordsLinked to original sources
bioRxiv2026

Biochemical and kinetic properties of a Type III restriction-modification enzyme Mbo45V from the host-adapted pathogen Mycoplasma bovis

Type III restriction-modification (RM) enzymes are prominent bacterial defense against bacteriophage and invading foreign DNA that also modulate the hosts epigenetic landscape. Genome analysis of the host-adapted Mycoplasma bovis PG45 that has a very small genome revealed a Type III RM locus comprising one res and three mod genes. We characterized Mbo45V, a representative enzyme encoded by this locus. The enzyme forms a heterotrimeric complex consisting of two Mod subunits and one Res subunit. Mbo45V recognizes the asymmetric sequence 5'-YAATC-3' (Y = T/C) and cleaves DNA having at least two head-to-head oriented sites [~]26-28 bp away from the recognition site. Methylation of the second adenine of the target site using cofactor S-adenosylmethionine (SAM) protects DNA from restriction, while the SAM analogue sinefungin enhances DNA binding and cleavage. Kinetic studies reveal that Mbo45V exhibits relatively weak DNA binding affinity and an unusually high Km for SAM, indicating low cofactor affinity compared to prototypical enzymes such as EcoP15I. ATPase activity is strongly stimulated by cognate DNA and is inhibited upon methylation of the substrate, suggesting a regulatory interplay between methylation and restriction functions. Comparative analysis indicates that, although Mbo45V shares core mechanistic features with prototypes from Escherichia coli, its kinetic parameters are distinct. These differences likely reflect adaptation to the stable intracellular environment of M. bovis, in contrast to the fluctuating conditions encountered by the enteric bacteria.

biochemistry↗
bioRxiv2026

MetaMuse: A Multi-Agent AI System for Biomedical Metadata Curation and Harmonization

Inconsistent and unstructured metadata in public biomedical repositories, such as the Gene Expression Omnibus (GEO), severely limits data discoverability and research reproducibility. To address this, we introduce MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW, a modular, multi-agent artificial intelligence framework designed to autonomously extract, validate, and standardize unstructured biomedical metadata. Operating through a three-stage architecture utilizing large language model agents, specialized CO_SCPLOWURATORC_SCPLOWAO_SCPLOWGENTSC_SCPLOW contextually extract candidate values for specific target metadata fields. A centralized AO_SCPLOWRBITRATORC_SCPLOWAO_SCPLOWGENTC_SCPLOW enforces cross-field logical consistency to prevent contradictory annotations. Finally, a NO_SCPLOWORMALIZERC_SCPLOWAO_SCPLOWGENTC_SCPLOW leveraging a domain-specific semantic search model (SapBERT) maps these free-text candidates to formal ontological terms. We evaluated MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW on a gold-standard dataset of manually curated GEO samples, achieving over 95% curation accuracy across key target metadata fields, and demonstrated robust scalability on a broader dataset of 400 samples. Notably, MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW avoids data hallucination by defaulting to conservative false negatives when evidence is ambiguous, thereby preserving strict data integrity. By providing a fully auditable and context-aware curation pipeline, MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW offers a scalable solution for enriching public data repositories and accelerating reproducible, data-driven scientific discovery.

genomics↗
bioRxiv2026

Adenocarcinoma cell mechanobiology is altered by the loss modulus of the surrounding extracellular matrix

Elastic and viscoelastic properties of extracellular matrices (ECM) are known to regulate cellular behavior and mechanosensation differently, with implications for morphogenesis, wound healing, and pathophysiology. Most in vitro cellular processes, including cell migration, are studied on linear-elastic substrates to mimic extracellular matrices. However, most tissues are viscoelastic and display a loss modulus (G) that may be 10-20% of their storage modulus (G) under biophysically relevant conditions. Recent research has shown that cells can distinguish between elastic and viscoelastic ECM, leading to alterations in their cellular morphology, migration rates, and contractility. Here, we present a protocol for creating PAH-based model ECMs that enables the fabrication of viscoelastic substrates with storage moduli similar to those of their elastic counterparts. To explore how G influences epithelial cell mechanobiology, we fabricated tunable viscoelastic model ECMs with G of 3 kPa, 8 kPa, and 12 kPa, and for each, independently tuned G values to approximately 300 Pa, 500 Pa, and 700 Pa, respectively. We found that A549 cells cultured on stiff elastic model ECMs migrated [~]30% slower and formed larger focal adhesions compared to their viscoelastic counterparts. Conversely, A549 cells on intermediate viscoelastic model ECMs exhibited a [~]54% reduction in migration speed, with no significant difference in focal adhesion size relative to their elastic counterparts. These findings highlight the complex interplay between substrate (ECM) elastic and viscoelastic properties in regulating epithelial cell mechanobiology and emphasize the importance of time-dependent matrix mechanics in governing epithelial responses.

biophysics↗
bioRxiv2025

GeneJEPA: A Predictive World Model of the Transcriptome

We introduce GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW, a self-supervised foundation model that learns a predictive world model of single-cell transcriptomes. Based on the Joint-Embedding Predictive Architecture, GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW predicts latent representations of masked gene sets from visible context, a shift away from reconstructing noisy expression values and toward world-model style inference over cellular state. To realize this at scale, a Perceiver encoder handles variable gene sets at fixed cost, and a tokenizer jointly represents gene identity and continuous expression using Fourier features. Trained on the Tahoe-100M atlas, GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW learns general representations that transfer across tissues and datasets. On downstream tasks, including drug response and perturbation prediction, it surpasses strong baselines and enables test-time scaling by progressively enlarging the cross-attention over the gene set, trading a small read cost for higher accuracy at inference. GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW moves toward foundation models that reason over gene-gene relations, enabling applications in annotation, prediction, and in-silico discovery.

genomics↗
bioRxiv2025

REVIVE-Flow: A Foundation Model for Blood DNAm Aging

Epigenetic clocks can predict biological age but cannot prescribe the interventions needed to reverse it. Here, we introduce REjuVenatIon Via Epigenetic Flow (Revive-Flow), a flow-matching model trained on a broad compendium of epigenetic blood studies to transport methylomes forward and backward in time. First, we learn the continuous vector field of aging as an Ordinary Differential Equation (ODE) within a stable, low-dimensional linear space. Then, the learned ODEs dynamics are integrated backward in time to define a natural, biologically-plausible rejuvenation trajectory. This path serves as a guide for a convex optimization problem that identifies the minimal, targeted CpG-level perturbation required to rejuvenate a sample. On a completely unseen test set comprising over 800 individuals from the European Prospective Investigation into Cancer and Nutrition (EPIC-Italy) cohort, Revive achieves 0.4 years rejuvenated per commanded year (R2 {approx} 0.99), with a smooth sparsity-effect trade-off. Extensive validation confirms the effect preserves inferred cell-type composition and targets biologically plausible loci enriched in genomic regions and pathways central to aging biology.

genomics↗
bioRxiv2025

Guided multi-agent AI invents highly accurate, uncertainty-aware transcriptomic aging clocks

Scientific discovery has long relied on human creativity, with computation limited to analysis. Here we report an AI-guided system, K-Dense, that accelerates hypothesis testing and delivers robust scientific discoveries. Trained on ARCHS4 (57,584 samples, 28 tissues, 1,039 cohorts, ages 1-114), the unified ensemble clock achieved R2 = 0.854 and MAE = 4.26 years--while uniquely providing calibrated confidence intervals. This self-aware design flags predictions made at transitional or extreme ages, where biological heterogeneity peaks, suggesting clinical utility for uncertainty itself. Development revealed stage-specific markers, including CDKN2A/p16 (senescence), AMPD3 (muscle wasting), MIR29B2CHG (progeroid traits), and SEPTIN3 (neurodegeneration resistance). Sliding-window analysis across 85 overlapping ranges uncovered wave-like shifts in gene importance, showing that transcriptomic aging signatures evolve continuously, not discretely. By transforming biological age assessment from static point estimates to calibrated predictions with explicit uncertainty, this approach establishes reliable and interpretable clocks. Beyond the clock itself, K-Dense demonstrates how guided AI can compress months of exploration into weeks, pointing toward a scalable framework for accelerated scientific discovery.

systems biology↗
bioRxiv2025

The Limitations of TabPFN for High-Dimensional RNA-seq Analysis

Tabular Prior-Data Fitted Networks (TabPFN) demonstrate remarkable performance on small-to-medium tabular datasets through in-context learning, but struggle with high-dimensional genomic data such as RNA-seq with tens of thousands of features. We investigate multiple approaches to adapt TabPFN for transcriptomic analysis using two benchmark datasets: Age-ARCHS4, a regression dataset derived from the ARCHS4 dataset (57,873 samples, 10,000 genes), and an Inflammatory Bowel Disease (IBD) classification dataset encompassing Crohns Disease and Ulcerative Colitis samples (2,490 samples, 10,000 genes). Our experimental design proceeds in two phases: first evaluating existing optimization methods, then testing novel adaptations including (1) self-supervised embedding learning and (2) Bulk-Former integration. We demonstrate that when constrained to equal training conditions (500 features, 10,000 samples), TabPFN outperforms classical baselines like random forest and XGBoost. However, when classical methods utilize full feature sets while TabPFN adaptations attempt to handle higher-dimensional data, all TabPFN variants consistently underperform the naive baseline. Our findings reveal fundamental limitations in current approaches to adapting TabPFN for genomic applications, showing that architectural modifications paradoxically degrade performance, while intelligent metadata-based subgrouping emerges as the most effective strategy for deploying TabPFN on biological data.

bioinformatics↗
bioRxiv2025

Parkinson's Paradox: Alpha-synuclein's Selective Strike on SNc Dopamine Neurons over VTA

In synucleinopathies, including Parkinsons disease (PD), dopamine neurons in the substantia nigra pars compacta (SNc) exhibit greater vulnerability to degeneration than those in the ventral tegmental area (VTA). While -synuclein (Syn) pathology is implicated in nigral dopamine neuron loss, the mechanisms by which Syn affects neuronal activity and midbrain dopamine network connectivity prior to cell death remain unclear. This study tested the hypothesis that elevated Syn expression induces pathophysiological changes in firing activity and disrupts network connectivity dynamics of dopamine neurons before neuronal loss. We employed two mouse models of synucleinopathy: preformed Syn fibril (PFF) injection and AAV-mediated expression of human Syn (hSyn) under the control of the tyrosine hydroxylase (TH) promoter, both targeting the VTA and SNc. Four weeks post-injection, brain sections underwent histological, electrophysiological, and network analyses. Immunohistochemistry for TH, hSyn, and phospho-Ser129 Syn assessed Syn expression and dopaminergic neuron alterations. Neuronal viability was evaluated using two complementary approaches: quantification of TH+ or FOX3+ and TUNEL labeling. Importantly, these analyses revealed no significant changes in neuronal counts or TUNEL+ cells at this time point, confirming that subsequent functional assessments captured pre-neurodegenerative, Syn-induced alterations rather than late-stage neurodegeneration. Electrophysiological recordings revealed a differential effect of hSyn expression. SNc dopamine neurons exhibited significantly increased baseline firing rates, whereas VTA dopamine neurons remained unchanged. These findings indicate a region-specific vulnerability to Syn-induced hyperactivity of dopamine neurons. Further analysis revealed impaired homeostatic firing rate regulation in SNc, but not VTA, dopamine neurons, demonstrated by a reduced capacity to recover baseline firing following hyperpolarization. Collectively, our results demonstrate that, prior to neurodegeneration, elevated Syn expression differentially disrupts both basal firing activity and network stability of SNc dopamine neurons, while sparing VTA dopamine neurons. By identifying neurophysiological changes preceding dopaminergic neuron loss, these findings provide critical insights into the pathophysiological mechanisms predisposing SNc neurons to degeneration in Parkinsons disease. Significance StatementA central question in Parkinsons disease research is why dopamine neurons in the substantia nigra pars compacta (SNc) are more vulnerable than those in the ventral tegmental area (VTA). This study reveals that alpha-synuclein (Syn) pathology differentially impacts dopamine neuronal activity and network connectivity, causing changes in the SNc before neuronal loss occurs, but not in the VTA. These findings provide a mechanism to explain the differential resilience of these neighboring dopamine neuron populations and provide insights into Parkinsons disease progression. The methodologies developed in this study establish a foundation for investigating network topology in deep brain structures and its role in neurodegenerative disorders.

neuroscience↗
bioRxiv2024

Brain-body circuit by which midbrain dopamine neuronal activity modulates splenic immunity

Midbrain dopamine neurons are well-known to shape central nervous system function, yet there is growing evidence for their influence on the peripheral immune system. Here we demonstrate that midbrain dopamine neurons form a circuit to the spleen via a multi-synaptic pathway from the dorsal vagal complex (DVC) through the celiac ganglion. Midbrain dopamine neurons modulate the activity of D1- and D2-like dopamine receptor-expressing DVC neurons. In vivo activation of midbrain dopamine neurons induced dopamine release in the DVC and increased immediate early gene expression in both the DVC and celiac ganglion. Activation of this midbrain-to-spleen circuit reduced spleen weight and decreased naive CD4+ T-cell populations without affecting total T-cell numbers. These findings unveil a functional midbrain- DVC-celiac ganglion-spleen pathway, through which midbrain dopamine neurons modulate splenic immunity. These novel insights into the neural regulation of the immune system have important implications for diseases involving altered dopamine neurotransmission and highlight potential targets for immunotherapeutic interventions.

neuroscience↗
bioRxiv2024

Transfer Learning Of Gene Expression Using Reactome

AO_SCPLOWBSTRACTC_SCPLOWIn clinical research, translating findings from model organisms to human applications remains challenging due to biological differences between species as well as limitations of orthologous, and homologous, gene comparisons, which is fraugt with information loss as well as many-to-many mapping. To address these issues, we introduce a novel Universal Gene Embedding (UGE) model that leverages transformer-based few-shot learning for species-agnostic transfer learning with heterogeneous domain adaptation. The UGE model, trained on a dataset of gene expression from ten organs across rats and mice, establishes a unified biological latent space that effectively represents genes from any organ or species. By focusing on reactomes--comprehensive profiles of gene expression responses to drugs--the UGE model enables functional gene mapping across species based on the similarities of these profiles. Our contributions include a gene reactome vector prediction module, a robust framework for mapping drug-induced gene expression patterns across species, strategies for optimizing experimental design, and enhanced gene mapping precision. These advancements provide a new tool for genetic research and a new paradigm for cross-species insights, potentially revolutionizing our understanding of gene function, drug responses, and the translation of findings from model organisms to human clinical applications.

systems biology↗
bioRxiv2024

Gene Expression Reactomes Across Species Do Not Correlate with Gene Structural Similarity

Mammalian animal model species for preclinical drug testing share a large number of genes with similar names and/or protein primary structures, but gene expression patterns in different tissues and in response to different drugs have not been previously systematically characterized. Here, we experimentally measure the gene expression responses (reactomes) of each gene in the mouse and rat genomes in 10 different organs to 3 different drugs. Surprisingly, we observe that gene reactomes across species generally do not correlate with structural similarity. Thus, we propose an alternative functional cross-species gene mapping approach, based on organ-specific gene expression response to drug dosing.

molecular biology↗
bioRxiv2023

Method for Identifying Proteomic Biomarkers of Health

Classification of proteomic samples from sick and non-sick individuals is important for developing high-quality diagnostics of diseases. Creating a shortlist of proteins useful for diagnostics is challenging. In this manuscript a simple algorithm of creating a multidimensional biomarker of health based on scanning proteomics data is provided. The algorithm is applied to several existing publicly available datasets and demonstrates a 9-protein indicator of atopic dermatitis, and 6-protein indicator of heart failure.

bioinformatics↗
bioRxiv2023

BIOFORMERS: A SCALABLE FRAMEWORK FOR EXPLORING BIOSTATES USING TRANSFORMERS

AO_SCPLOWBSTRACTC_SCPLOWGenerative pre-trained models, such as BERT and GPT, have demonstrated remarkable success in natural language processing and computer vision. Leveraging the combination of large-scale, diverse datasets, transformers, and unsupervised learning, these models have emerged as a promising method for understanding complex systems like language. Despite the apparent differences, human language and biological systems share numerous parallels. Biology, like language, is a dynamic, interconnected network where biomolecules interact to create living entities akin to words forming coherent narratives. Inspired by this analogy, we explored the potential of using transformer-based unsupervised model development for analyzing biological systems and proposed a framework that can ingest vast amounts of biological data to create a foundational model of biology using BERT or GPT. This framework focuses on the concept of a biostate, defined as a high-dimensional vector encompassing various biological markers such as genomic, proteomic, transcriptomic, physiological, and phenotypical data. We applied this technique to a small dataset of single-cell transcriptomics to demonstrate its ability to capture meaningful biological insights into genes and cells, even without any pre-training. Furthermore, the model can be readily used for gene network inference and genetic perturbation prediction.

bioinformatics↗
bioRxiv2023

High Frequency Longitudinal RNAseq Reveals Temporally Varying Genes and Recovery Trajectories in Rats

When living organisms are exposed to potentially toxic chemicals, they respond via changes in gene expression. Traditional differentially expressed gene analysis based on before/after blood samples does not reveal the responses temporal dynamics, and often produces false positives and negatives. Here, we performed longitudinal daily RNA sequencing on rats dosed with tetracycline, isoniazid, carbon tetrachloride, or valproate. We identified 4,302 temporally variable genes (TVGs) with statistically strong change in expression following dosing. Projecting TVG expressions into a 3-dimensional principal component (PC) space reveals consistent trajectories for recovery following dosing, and enables separation of healthy from recovering states with 91% to 99% area under the receiver operator curve. Finally, we observed that Fast Recovery vs. Slow Recovery rats exhibited distinct temporal expression patterns in the TVGs, suggesting that individual variations could be potentially captured via longitudinal RNAseq analysis.

molecular biology↗
bioRxiv2023

Multi-population dissolution in confined active fluids

Autonomous out-of-equilibrium agents or cells in suspension are ubiquitous in biology and engineering. Turning chemical energy into mechanical stress, they generate activity in their environment, which may trigger spontaneous large-scale dynamics. Often, these systems are composed of multiple populations that may reflect the coexistence of multiple species, differing phenotypes, or chemically varying agents in engineered settings. Here, we present a new method for modeling such multi-population active fluids subject to confinement. We use a continuum multi-scale mean-field approach to represent each phase by its first three orientational moments and couple their evolution with those of the suspending fluid. The resulting coupled system is solved using a parallel adaptive level-set-based solver for high computational efficiency and maximal flexibility in the confinement geometry. Motivated by recent experimental work, we employ our method to study the spatiotemporal dynamics of confined bacterial suspensions and swarms dominated by fluid hydrodynamic effects. Our computational explorations reproduce observed emergent collective patterns, including features of active dissolution in two-population active-passive swarms, with results clearly suggesting that hydrodynamic effects dominate dissolution dynamics. Our work lays the foundation for a systematic characterization of natural and synthetic multi-population systems such as bacterial colonies, bird flocks, fish schools, colloidal swimmers, or programmable active matter.

biophysics↗
bioRxiv2023

Lateral gating mechanism and plasticity of the BAM complex in micelles and E. coli

The {beta}-barrel assembly machinery (BAM) mediates folding and insertion of the majority of OMPs in Gram-negative bacteria. BAM is a penta-heterooligomeric complex consisting of the central {beta}-barrel BamA and four interacting lipoproteins BamB, C, D, and E. The conformational switching of BamA between inward-open (IO) and lateral-open (LO) conformations is required for substrate recognition and folding. However, the mechanism for the lateral gating or how the structural details observed in vitro correspond with the cellular environment remains elusive. Here we addressed these questions by characterizing the conformational heterogeneity of BamAB, BamACDE and BamABCDE complexes in detergent micelles and or E. coli using pulsed dipolar electron spin resonance spectroscopy (PDS). We show that the binding of BamB does not induce any visible changes in BamA and the BamAB complex exists in the IO conformation. The BamCDE complex induces an IO to LO transition through a coordinated movement along the BamA barrel. However, the extracellular loop (L6) is unaffected by the presence of lipoproteins and exhibits a large segmental dynamics extending to the exit pore. PDS experiments with BamABCDE complex in intact E. coli confirmed the dynamic behavior of both the lateral gate and the L6 in the native environment. Our results demonstrate that the BamCDE complex plays a key role for the function by regulating lateral gating in BamA.

biophysics↗
bioRxiv2023

Size dependent diffusion and dispersion of particles in mucin

Mucus, composed significantly of glycosylated mucins, is a soft rheologically complex vis-coelastic material lining respiratory, reproductive, and gastrointestinal tracts in mammals. Mucus may present as a gel or as a fluid, and serves as a barrier to the transport of microbes including, harmful particles, and inhaled atmospheric pollutants. Studies on mucin gels have provided insight into swelling kinetics, and the diffusion and permeability of molecular constituents such as water. The transport and dispersion of micron and sub-micron tracer particles in mucin gels and solutions differs from the motion of molecular water since the much larger tracers may interact with microstructure and larger features of the mucin network. Here, using brightfield and florescence microscopy, high speed particle-tracking, and passive microrheology, we study the thermally driven stochastic movement of 0.5 - 5.0 {micro}m tracer particles in 10% mucin solutions at neutral pH, and in 10% mucin mixed with unmodified limestone rock dust, modified limestone, and crystalline silica. Particle trajectories, mean square displacements, and the displacement probability distributions, are used to assess tracer diffusion and transport. Complex moduli are concomitantly extracted using microrheology techniques. We find that under the conditions analyzed in our experiments the mucin solution present as a highly viscous, weakly viscoelastic fluid rather than as a viscoelastic gel. For small to moderately sized tracers with diameter than 2 {micro}m, effective diffusion coefficients follow predictions of classical Stokes-Einstein relationship. Diffusivity in rock-dust laden mucin is surprisingly larger than in bare mucin. Probability distributions of squared particle displacements indicate that heterogeneity, transient trapping, and electrostatic interactions impact tracer transport, especially for larger tracers. Our results prompt further exploration of physiochemical and rheological mechanisms mediating particle transport in soft, viscoelastic biopolymer networks and materials.

bioengineering↗
bioRxiv2023

Facile determination of the Poisson's ratio and Young's modulus of polyacrylamide gels and polydimethylsiloxane

Polyacrylamide hydrogels (PAH) and polydimethylsiloxane (PDMS) are two soft materials often used in cell mechanics and mechanobiology, in manufacturing lab-on-a chip applications, among others. This is partly due to the ability to tune their elasticity with ease, in addition to various chemical modifications. For affine polymeric networks, two (of three) elastic constants - the Youngs modulus (E), the shear modulus (G), and the Poissons ratio ({nu}) - describe the purely elastic response to external forces. However, the literature addressing the experimental determination of {nu} for PAH (also sometimes referred to as PAA gels in the literature) and PDMS is surprisingly limited when compared to the literature reporting values of E and G. Here, we present a facile method to obtain the Poisons ratio and Youngs modulus for PAH and PDMS based on static tensile tests, and cross-correlate these values with those obtained via a second independent method, shear rheology. We show that: i) the Poissons ratio may vary significantly from the value for incompressible materials ({nu} = 0.5), and ii) find a high degree of agreement between shear rheology and macroscopic static tension tests for PAH but not PDMS. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=121 SRC="FIGDIR/small/540222v1_ufig1.gif" ALT="Figure 1"> View larger version (24K): org.highwire.dtl.DTLVardef@120116forg.highwire.dtl.DTLVardef@58acbcorg.highwire.dtl.DTLVardef@1b48683org.highwire.dtl.DTLVardef@eb3fda_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗