bioRxiv Science⌕ Search

Biology subjects

Chauhan, S.

Publications and source records attributed to Chauhan, S..

At least 19 recordsLinked to original sources

Scalable Computational Phenomics of Nuclear Morphology

Interpretable computational pathology is constrained by a trade-off between scalable learned representations and biologically explicit phenotypes. Here we present NucXplore, a nuclear phenomics framework that converts each hematoxylin-and-eosin (H&E)-stained nucleus into 129 explicitly defined features spanning morphology, chromatin, intensity, texture, color, and spatial context. A Rust-based implementation accelerates extraction 17.6-fold over the Python reference implementation. Across three histopathology cohorts, NucXplore outperformed classical descriptors and larger deep-learning embeddings while retaining feature attribution. Applied to human skin, hierarchical multiple-instance models resolved eight cellular compartments and decoded age and sun-exposure states, revealing a largely shared aging program with age and cell-specific exposure-dependent recalibration. Conditional flow matching and neural ODE integration transformed cross-sectional phenotypes into model-implied trajectories, uncovering a midlife velocity minimum, late-life re-acceleration, and cell-specific changes in trajectory magnitude and direction. Source-free domain adaptation enabled preliminary transfer of age-associated predictions to an independent hospital cohort. NucXplore thus unifies interpretable representation, high-performance computing, and dynamical modeling for cellular phenotyping from routine histology.

pathology↗

tACS increases alpha burst duration in a frequency- and montage-specific manner

Transcranial alternating current stimulation (tACS) can experimentally manipulate neural oscillations in humans non-invasively. Previous studies consistently suggest that 10 Hz tACS in posterior regions leads to a general increase in alpha (7-15 Hz) oscillations. However, the specificity of this effect regarding stimulation frequency and location as well as its consequences on cognition are still unclear. We therefore aimed to investigate whether the putative effects of tACS on neural oscillations and cognition are specific to posterior stimulation in the alpha frequency range. We tested these effects using a 2 x 2 design (posterior or fronto-parietal stimulation, at 10 Hz or 6 Hz) plus a no-stimulation control group, during a sustained attention task.158 healthy participants were randomly allocated to control group (no stimulation) or one of the four active groups. All participants underwent an Electroencephalography recording and completed the Amsterdam Resting State Questionnaire before and after stimulation. Our results show that, relative to the control group, alpha power increases significantly after stimulation for the group receiving posterior 10 Hz stimulation and that this was explained by an increase in alpha oscillatory bursts duration rather than their amplitude. On the other hand, no significant differences were found between active and controls groups in behavioural performance or resting state cognition. Together, our findings demonstrate that tACS increases the duration of alpha oscillatory bursts in a montage- and frequency-specific manner and that this modulation is not accompanied by significant changes in sustained attention during stimulation or resting state cognition after stimulation.

neuroscience↗

A patient-derived xenograft model of FUS::TFCP2 intraosseous rhabdomyosarcoma reveals chemoresistance and differential sensitivity to ALK inhibitors

TFCP2-rearranged rhabdomyosarcoma is an exceptionally rare and highly aggressive malignancy driven by TFCP2 gene fusions and associated with a dismal clinical prognosis. Because standardized treatment regimens are lacking, developing representative preclinical models is critical for identifying effective therapies. Here, we present a case of a 29-year-old male with rapidly progressive, metastatic pelvic intraosseous rhabdomyosarcoma (iRMS) harboring a FUS::TFCP2 fusion and anaplastic lymphoma kinase (ALK) overexpression. To evaluate therapeutic vulnerabilities, we established a patient-derived xenograft (PDX) model that faithfully recapitulated the histologic, immunohistochemical, and molecular hallmarks of the primary tumor. High-throughput in vitro pharmacological screening of PDX-derived cells demonstrated notable resistance to standard cytotoxic chemotherapies and revealed a paradoxical and selective sensitivity profile across ALK inhibitors. The PDX-derived cells were susceptible to crizotinib, brigatinib, and ceritinib, yet resistant to the more selective second- and third-generation inhibitors alectinib and lorlatinib. Notably, next-generation ROS1/pan-TRK inhibitors (entrectinib, repotrectinib, and taletrectinib) demonstrated superior efficacy compared to the fourth-generation ALK inhibitor NVL-655. Our findings establish a validated preclinical PDX model for FUS::TFCP2 iRMS and suggest that multi-targeted tyrosine kinase inhibition may offer a more viable therapeutic strategy than narrow-spectrum ALK targeting or conventional chemotherapy.

cancer biology↗

A human monoclonal antibody from hybrid immunity exhibits exceptional neutralization breadth against antigenically divergent contemporary Omicron JN.1 lineage variants

Hybrid immunity offers stronger and more durable antibody-mediated protection against symptomatic SARS-CoV-2 infection. In the present study, we report the development of a durable antibody response in an individual with hybrid immunity who also received three doses of the prototype COVID-19 vaccine. Polyclonal plasma antibody obtained from this donor also showed extraordinary neutralization breadth against contemporary Omicron variants. One of the functional monoclonal antibodies (ATHS-C30) isolated from this individual, representing the IGHV4-30 lineage-specific B cell with strong binding affinity to JN.1 spike protein, showed extraordinary neutralization breadth, including variants of Omicron lineages that emerged beyond JN.1, such as KP.2, KP.3.1.1, KP.3.2, KP.3.3, LB.1 and XEC. ATHS-C30 was found to bind RBD with high affinity and showed distinct epitope specificity to the other neutralizing mAbs isolated from the same donor through the epitope binning assay. Molecular modelling of the CDHR3 sequence using existing structures indicated that ATHSC-30 belongs to the class 4 antibody, a feature that contributes to breadth, while epitope conservation analysis indicated that the majority of RBD-interacting residues of ATHSC-30 are evolutionarily conserved. Taken together, our study indicate that ATHS-C30 forms the basis of development of a robust and broadly neutralizing antibody response in this individual with hybrid immunity, which overcomes the antigenic variation by targeting highly conserved and cryptic epitopes in destabilizing the spike structure. ImportanceSARS-CoV-2 continues to pose a significant public health threat, particularly to immunocompromised individuals and older adults with underlying comorbidities. Hybrid immunity to SARS-CoV-2 in vaccinated individuals leads to the development of B cells that are qualitatively superior to those that are expected to develop in only vaccinated individuals. In the present study, we found that among individuals with hybrid immunity who developed robust, durable antigen-specific antibody responses, one developed antibody response capable of broadly cross-neutralising contemporary Omicron variants. This was correlated with development of antigen-specific B cell lineage, such as IGHV4-30, that produced antibodies with potent and extraordinary neutralisation breadth against contemporary Omicron lineages, such as KP.2, KP.3.1.1, KP.3.2, KP.3.3, LB.1 and XEC. This is believed to be due to heterologous antigen exposures driving the development of an antigen-specific B cell repertoire that, in turn, facilitates immune imprinting capable of overcoming the ineffectiveness of antibodies to effectively neutralize newly emerging Omicron variants.

immunology↗

HIRA-mediated H3.3 deposition preserves hepatocyte cell identity during liver aging

Age-associated functional decline is partly driven by progressive chromatin degeneration. Maintenance of chromatin integrity preserves cell identity and promotes healthy aging, but through different mechanisms in proliferating and non-proliferating cells. However, specific mechanisms of chromatin maintenance and their compensatory capacity in proliferating and non-proliferating cells are undefined. The histone chaperone HIRA deposits the histone variant H3.3 in a DNA replication-independent manner, leading to its accumulation in aging, non-proliferating cells. Here, we show that hepatocyte-specific loss of HIRA causes loss of cell identity, metabolic dysfunction, and accelerated fibrotic pathology with age. Transcriptomic and epigenomic analyses indicate that HIRA-H3.3 preserves chromatin integrity and sustains transcription of highly expressed genes, including cell identity genes. Partial hepatectomy, associated with induced proliferation, restores identity of HIRA knockout livers with compensatory deposition of canonical histones H3.1/2. Together, these results demonstrate that HIRA-mediated H3.3 deposition is essential for safeguarding cell identity and tissue function during aging of non-proliferating cells, but this function can be rescued by tissue regeneration and associated cell proliferation. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=124 SRC="FIGDIR/small/710643v1_ufig1.gif" ALT="Figure 1"> View larger version (29K): org.highwire.dtl.DTLVardef@a2a26aorg.highwire.dtl.DTLVardef@154e5aeorg.highwire.dtl.DTLVardef@b315a7org.highwire.dtl.DTLVardef@152bca4_HPS_FORMAT_FIGEXP M_FIG C_FIG

genomics↗

Chemical Dice Integrator (CDI): A Scalable Framework for Multimodal Molecular Representation Learning

The machine learning landscape for molecular property prediction is fragmented, with numerous Featurizers each capturing a narrow, specialized view of chemical structure. This heterogeneity forces a suboptimal choice of representation a priori, limiting model generalizability. We introduce the Chemical Dice Integrator (CDI), a hierarchical framework that unifies six orthogonal molecular representations, physicochemical (Mordred), topological (GROVER), visual (ImageMol), biological (Signaturizer), quantum-mechanical (MOPAC), and linguistic (ChemBERTa), into a single, coherent embedding. The framework consists of CDI-Basic, a two-tiered autoencoder that fuses these modalities, and CDI-Generalised, a Mamba State-Space Model (SSM) that learns a direct, efficient map from SMILES strings to the unified embedding space. Extensive benchmarking across 23 classification (171 tasks) and 10 regression datasets demonstrates that CDI embeddings consistently achieve superior predictive performance compared to individual Featurizers and standard feature aggregation methods. The CDI-Generalised model achieves this performance with exceptional computational efficiency, outperforming deep learning Featurizers in terms of speed and resource overhead. Furthermore, we demonstrate that the CDI embedding is chemically intuitive, allowing for the sensitive distinction of nuanced structural variants, such as chiral enantiomers and kekulized SMILES forms. By bridging multimodal chemical intelligence with scalable, sequence-based inference, CDI offers a strong foundation for molecular machine learning.

bioinformatics↗

SynGlue: AI-Driven Designer for Clinically Actionable Multi-Target Therapeutics

The rational design of protein degraders, such as proteolysis-targeting chimeras (PROTACs), requires the simultaneous optimization of multiple molecular properties, a complex challenge that limits efficient discovery. Here, we introduce SynGlue, a generative artificial intelligence (AI) framework that addresses this challenge through two core modules: data-driven, leveraging large-scale protein-ligand intelligence, and structure-guided, for physics-aware molecular design. SynGlue harness MagnetDB, a curated database of 6.37 million experimental protein-ligand interactions, and couples it with deep learning models that quantitatively predict degradation potency (DC50), maximal degradation (Dmax), and guide ternary-complex-compatible linker design. Benchmarked against 6,935 compounds, SynGlue demonstrates superior performance in relevant pharmacology prediction. To validate SynGlue, we engineered degraders for BRD4 and GSPT1. Our data-driven design for BRD4 yielded compounds with novel warhead scaffolds (<50% warhead similarity with known PROTACs), which proved to be potent degraders in vitro (DC50 = 0.19 nM) and efficacious in vivo in mouse models. Independently, our structure-guided de novo design for GSPT1 produced ultrapotent degraders (DC50 {approx} 0.0011 M) that are also effective both in vitro and in vivo, uncovering a new oncogenic dependency. By unifying data-driven and physics-aware design, SynGlue establishes a generalizable AI framework for the rapid development of clinically relevant protein degraders, with principled extension to other multi-target modalities.

bioinformatics↗

MutAIverse: An AI-Powered, Mechanism-backed Platform for Discovering Novel DNA Adducts and their precursor Genotoxins

Genotoxin exposure leads to DNA adduct formation, potentially causing mutations if unrepaired. Current DNA adductomics platforms or analytical workflows are limited by incomplete spectral libraries, reliance on experimentally validated adducts, limited cellular contexts, and inefficient computational methodologies. We introduce MutAIverse, an advanced AI-driven DNA adductomics analysis platform that overcomes these limitations by leveraging intracellular mechanistic modeling of genotoxin bioactivation to construct a comprehensive DNA adduct library and offers advanced spectral mapping. MutAIverse integrates experimentally validated and chemically valid putative DNA adducts, enabling interpretable retrograde tracking to parental genotoxins. Validation of MutAIverse against experimental MS/MS datasets demonstrated the detection of both known and novel adducts. Furthermore, application to in-house generated DNA adductomics data from tissue biopsies of smokeless tobacco-induced head and neck cancer patients revealed selective enrichment of both novel and known DNA adducts; a subset of them was validated using MS/MS analysis. Collectively, MutAIverse provides a robust, end-to-end, and interpretable platform for advanced DNA adductomics analysis.

bioinformatics↗

STING is a key driver of Japanese encephalitis virus induced inflammatory response

Interferon (IFN) and inflammation are the key early defence mechanisms that combat pathogen infection. The cytosolic DNA sensor cGAS activates immune signaling via the stimulator of interferon genes (STING) protein. Emerging evidence suggests crosstalk between innate immune DNA and RNA sensing, implicating a role of STING protein in RNA virus infection. This study characterizes STING in the context of Japanese encephalitis virus (JEV), an RNA virus of the flaviviridae family. We observe that activation of type I IFN through MAVS is essential for cGAS and STING activation. Knockdown, null mutant and inhibitor studies confirmed that STING restricts JEV replication independently of IFN{beta} signaling and autophagy. Transcriptomic analysis of STINGgt/gt bone-marrow derived macrophages (BMDMs) showed enhanced IFN response, but reduced activation of inflammatory cytokines and chemokines. Phosphorylated STING was recruited on the virus replication complex (RC), marked by the non-structural protein NS1, subsequently triggering the assembly of the NLRP3 inflammasome on the RC. STING proton channel activity was essential for NLRP3 inflammasome activation, IL-1{beta} production, and activation of pyroptotic cell death markers. Stinggt/gt mice, showed higher viremia, earlier disease onset, reduced survival, and decreased brain inflammation. These findings establish STING as a key regulator of JEV-induced inflammation and antiviral defence.

microbiology↗

MAVS Orchestrates a Potent IFN-independent Antiviral Immunity by Preserving Mitochondrial Integrity

Intrinsic antiviral defenses can restrict infection independent of paracrine interferon (IFN) signaling. While mitochondrial homeostasis is essential for immunity, its direct role in viral restriction remains incompletely understood. Here, we identify mitochondrial antiviral signaling protein (MAVS) as a central effector of a potent IFN-independent antiviral immunity orchestrated though mitochondrial safeguarding. We demonstrate that MAVS is critical for mitochondrial integrity as its loss leads to mitochondrial fragmentation, depolarization, and elevated mitophagy, alongside impaired bioenergetics and protein import. Mechanistically, MAVS maintains mitochondrial integrity by stabilizing the translocase of the outer membrane (TOM) complex and sustaining the abundance of its core components. Remarkably, this mitochondria-driven immunity restricts SARS-CoV-2 replication even under IFN-deficient conditions, and operates alongside the IFN pathways during JEV infection. Our findings redefine MAVS as a mitochondrial guardian that preserves organellar stability to enact host defense. These results highlight mitochondrial integrity as a fundamental determinant of broad antiviral immunity, integrating intrinsic and IFN-dependent mechanisms to counteract viral pathogenesis.

immunology↗

Histone H3.3 ensures cell proliferation and genomic stability during myeloid cell development

Variant histone H3.3 is thought to be critical for survival of many cells, since it is deposited in expressed genes, a feature different from core histones. For example, H3.3 deletion leads to embryonic lethality in mice. However, requirement of H3.3 in later stage of development has remained unclear. The aim of this work was to elucidate the role of H3.3 for development of myeloid lineage, important for innate immunity. We conditionally knocked out (cKO) the H3.3 genes in myeloid progenitor cells differentiating into bone marrow derived macrophages (BMDMs). Progenitor cells lacking H3.3 were defective in replication, suffered from extensive DNA damage, and underwent apoptosis. Surviving H3.3cKO cells expressed many interferon stimulated genes (ISGs) throughout differentiation. Further, H3.3cKO BMDMs possessed chromatin accessible sites, and histone posttranslational modifications consistent with the gene expression profiles, Accordingly, H3.3cKO BMDMs retained general nucleosomal structure genome wide. In summary, H3.3 is required for proliferation of myeloid progenitor cells, but is in large part dispensable for differentiation of BMDMs.

cell biology↗

Squidiff: Predicting cellular development and responses to perturbations using a diffusion model

Single-cell sequencing has revolutionized our understanding of cellular heterogeneity and responses to environmental stimuli. However, mapping transcriptomic changes across diverse cell types in response to various stimuli and elucidating underlying disease mechanisms remains challenging. Studies involving physical stimuli, such as radiotherapy, or chemical stimuli, like drug testing, demand labor-intensive experimentation, hindering mechanistic insight and drug discovery. Here we present Squidiff, a diffusion model-based generative framework that predicts transcriptomic changes across diverse cell types in response to environmental changes. We demonstrate Squidiffs robustness across cell differentiation, gene perturbation, and drug response prediction. Through continuous denoising and semantic feature integration, Squidiff learns transient cell states and predicts high-resolution transcriptomic landscapes over time and conditions. Furthermore, we applied Squidiff to model blood vessel organoid development and cellular responses to neutron irradiation and growth factors. Our results demonstrate that Squidiff enables in silico screening of molecular landscapes, facilitating rapid hypothesis generation and providing valuable insights for precision medicine.

bioinformatics↗

Multiple layers of innate immune response antagonism of SARS-CoV-2

Several SARS-CoV-2 proteins have been shown to counteract the host innate immune response, mostly using in vitro protein expression, which may not fully reflect their role in the context of viral infection. In addition, while each viral protein was characterized in a different experimental system, their relative contribution in immunosuppression remains unclear. Here we used a SARS-CoV-2 bacterial artificial chromosome with en passant mutagenesis to recover a panel of twelve infectious recombinant SARS-CoV-2 viruses, each with mutations in either NSP1, NSP2, NSP3, NSP6, NSP12, NSP13, NSP14, NSP15, NSP16, ORF3a, ORF6 or ORF8. We used the interferon-stimulated response element (ISRE)-driven luciferase assay in 293T-ACE2/TMPRSS2 cells to test the panel, demonstrating that mutations in many proteins, especially in NSP1 and NSP15, increased the type I interferon response relative to the parental wild-type virus. RNA-seq analysis of mutant-virus infected Calu-3 cells showed that the mutations in NSP1 or NSP15 lead to higher expression of multiple genes involved in innate immune response, cytokine-mediated signaling and regulation of lymphocyte proliferation. Furthermore, mutations in either NSP1 or NSP15 resulted in a greater maturation of human monocyte-derived dendritic cells in vitro. Infection of K18 hACE2 transgenic mice with either NSP1 or NSP15 mutated viruses demonstrated attentuated respiratory tract replication. Analysis of lung immune cells from infected mice by single-cell RNA-seq identified 15 populations of major myeloid and lymphoid cells with changes in the pattern of their activation associated with viral infection. The effects of mutations in NSP1 or NSP15 on these responses are consistent with differences in the immunosuppressive mechanisms utilized by the two proteins. Overall, these data demonstrate different and redundant mechanisms of innate immune antagonism by SARS-CoV-2 and suppression of activation of antigen presenting cells and T and B lymphocytes mediated by multiple viral proteins. AUTHOR SUMMARYThe mechanisms by which SARS-CoV-2 and its proteins modulate host immunity, specifically the interferon response, are still not clear. We generated twelve infectious SARS-CoV-2 viruses with mutations in individual proteins and demonstrated that many of them have interferon-antagonizing activity and immunosuppressive effects in human cells and in the K18 hACE mouse model of infection. We idemtified distinct and redundant mechanisms of immunosuppression of SARS-CoV-2 mediated by multiple individual viral proteins, with 9 out of the 12 tested proteins showing some immunosuppressive effect in at least one experimental system. The demonstrated immunosuppressive effects extend from the innate response to immune cells to pathologic changes in vivo. Importantly, this work shows, for the first time, a comparison of the effects of multiple viral proteins in the context of authentic viral infection, rather than in a surrogate system, and shows the relative contribution of each viral protein under identical experimental conditions. Overall, our data indicates that SARS-CoV-2 antagonizes multiple immune mechanisms, particularly type I interferon signaling, activation of innate immune cells and T and B lymphocyte functions with the greatest effects due to NSP1 and NSP15.

microbiology↗

Human vascular organoids with a mosaic AKT1 mutation recapitulate Proteus syndrome

Vascular malformation, a key clinical phenotype of Proteus syndrome, lacks effective models for pathophysiological study and drug development due to limited patient sample access. To bridge this gap, we built a human vascular organoid model replicating Proteus syndromes vasculature. Using CRISPR/Cas9 genome editing and gene overexpression, we created induced pluripotent stem cells (iPSCs) embodying the Proteus syndrome-specific AKTE17K point mutation for organoid generation. Our findings revealed that AKT overactivation in these organoids resulted in smaller sizes yet increased vascular connectivity, although with less stable connections. This could be due to the significant vasculogenesis induced by AKT overactivation. This phenomenon likely stems from boosted vasculogenesis triggered by AKT overactivation, leading to increased vascular sprouting. Additionally, a notable increase in dysfunctional PDGFR{beta} + mural cells, impaired in matrix secretion, was observed in these AKT-overactivated organoids. The application of AKT inhibitors (ARQ092, AZD5363, or GDC0068) reversed the vascular malformations; the inhibitors effectiveness was directly linked to reduced connectivity in the organoids. In summary, our study introduces an innovative in vitro model combining organoid technology and gene editing to explore vascular pathophysiology in Proteus syndrome. This model not only simulates Proteus syndrome vasculature but also holds potential for mimicking vasculatures of other genetically driven diseases. It represents an advance in drug development for rare diseases, historically plagued by slow progress.

bioengineering↗

BRD4 directs mitotic cell division by inhibiting DNA damage

BRD4 binds to acetylated histones to regulate transcription and drive cancer cell proliferation. However, the role of BRD4 in normal cell growth remains to be elucidated. Here we investigated the question by using mouse embryonic fibroblasts with conditional Brd4 knockout (KO). We found that Brd4KO cells grow more slowly than wild type cells: they do not complete replication, fail to achieve mitosis, and exhibit extensive DNA damage throughout all cell cycle stages. BRD4 was required for expression of more than 450 cell cycle genes including genes encoding core histones and centromere/kinetochore proteins that are critical for genome replication and chromosomal segregation. Moreover, we show that many genes controlling R-loop formation and DNA damage response (DDR) require BRD4 for expression. Finally, BRD4 constitutively occupied genes controlling R-loop, DDR and cell cycle progression. We suggest that BRD4 epigenetically marks those genes and serves as a master regulator of normal cell growth.

cell biology↗

Cell arrangement impacts metabolic activity and antibiotic tolerance in Pseudomonas aeruginosa biofilms

Cells must access resources to survive, and the anatomy of multicellular structures influences this access. In diverse multicellular eukaryotes, resources are provided by internal conduits that allow substances to travel more readily through tissue than they would via diffusion. Microbes growing in multicellular structures, called biofilms, are also affected by differential access to resources and we hypothesized that this is influenced by the physical arrangement of the cells. In this study, we examined the microanatomy of biofilms formed by the pathogenic bacterium Pseudomonas aeruginosa and discovered that clonal cells form striations that are packed lengthwise across most of a mature biofilms depth. We identified mutants, including those defective in pilus function and in O-antigen attachment, that show alterations to this lengthwise packing phenotype. Consistent with the notion that cellular arrangement affects access to resources within the biofilm, we found that while the wild type shows even distribution of tested substrates across depth, the mutants show accumulation of substrates at the biofilm boundaries. Furthermore, we found that altered cellular arrangement within biofilms affects the localization of metabolic activity, the survival of resident cells, and the susceptibility of subpopulations to antibiotic treatment. Our observations provide insight into cellular features that determine biofilm microanatomy, with consequences for physiological differentiation and drug sensitivity.

microbiology↗

Scalable projected Light Sheet Microscopy for high-resolution imaging of large samples

Light sheet fluorescence microscopy (LSFM) is a widely used imaging technique for living and large cleared samples. However, high-performance LSFM systems are often prohibitively expensive and not easily scalable for high-throughput applications. Here, we introduce a cost-effective, scalable, and versatile high-resolution imaging framework, called projected Light Sheet Microscopy (pLSM), which repurposes readily available off-the-shelf consumer-grade components and an over-the-network control architecture to achieve high-resolution imaging of living and cleared samples. We extensively characterize the pLSM framework and showcase its capabilities through high-resolution, multi-color imaging and quantitative analysis of mouse and post-mortem human brain samples cleared using various techniques. Moreover, we show the applicability of pLSM for high-throughput molecular phenotyping of human induced pluripotent cells (iPSC)-derived brain and vessel organoids. Additionally, we utilized pLSM for comprehensive live imaging of bacterial pellicle biofilms at the air-liquid interface, uncovering their intricate layered architecture and diverse cellular dynamics across different depths. Overall, the pLSM framework has the potential to further democratize LSFM by making high-resolution light sheet microscopy more accessible and scalable.

neuroscience↗

genomicBERT and data-free deep-learning model evaluation

The genome, which serves as the inherent language directing the blueprint of life, offers significant analysis prospects by combining Natural Language Processing (NLP) and machine learning (ML). Integrating biological sequences with other digital healthcare information has potential to transform data-driven diagnostics. Large language models (LLMs) can be harnessed to decode the genomic language. This endeavor encounters three critical challenges: First, long biomolecular sequences require segmentation into smaller subunits, which is non-trivial since many biological "words" remain unknown. Second, the analysis of extended DNA sequences using LLMs demands a compute-intensive infrastructure. Third, ensuring reproducibility and reusability of modeling workflows remains an unresolved issue. To tackle these challenges, we introduce an empirical DNA tokenisation approach and a versatile, semantic-aware, genome language model --genomicBERT. The model is species-agnostic and operates seamlessly at the DNA or RNA levels. By introducing a reduced and specialized DNA vocabulary, our approach minimizes computational overhead and optimizes performance. Our benchmarking demonstrates that the genomicBERT matches or surpasses the performance of contemporary tools on the same datasets under different experimental conditions. To encourage collaboration and ease of access, we introduce genomicBERT as an integral component of the openly accessible conda package, genomeNLP. Validated across diverse case studies, genomicBERT lowers the barriers to decoding genomic language, relying solely on sequence data to extract meaningful insights. HighlightsO_LIThis novel model offers a compelling solution for DNA sequence analysis by significantly reducing model size and computational costs without compromising performance, setting a new standard for efficient model development. C_LIO_LIWe demonstrate that a powerful vocabulary and tokenization method helps to derive patterns from biological sequence data while accounting for hidden semantic rules. C_LIO_LIOur method is agnostic to species or biomolecule type as it is data-driven. Hence, it can be applied to DNA and RNA C_LIO_LIWe validate the important genomicBERT tokens by mapping back to the biologically significant motifs. C_LIO_LIWe present a publicly available genome language modeling toolkit called genomeNLP, specifically designed to combine computational linguistics and genomics, enabling researchers from biology backgrounds to analyze and interpret genomic sequences effectively. C_LI

bioinformatics↗