Biology subjects

Wang, B.

Publications and source records attributed to Wang, B..

At least 163 records · Page 9Linked to original sources

bioRxiv2025

Micro-scale spatial metagenomics: revealing high-resolution spatial biogeography of gut microbiomes

Spatial organisation is a fundamental yet poorly resolved aspect of gut microbial ecology. Conventional shotgun metagenomics provides rich functional information but relies on homogenised, macro-scale samples that obscure the micron-scale distributions critical for understanding microbial community dynamics. Here, we introduce Micro-Scale Spatial Metagenomics (MSSM), a new methodological framework that couples laser micro-dissection of tissue sections, ultra-low-input library preparation, and genome-resolved bioinformatics to reconstruct microbial communities from intestinal microsamples measuring as little as [~]500 {micro}m{superscript 2} ({approx}100 bacterial cells). We describe a fully optimised laboratory and computational pipeline that enables quantitative, strain-resolved, and functionally informed spatial profiling directly from intact gut tissue. Using chicken intestinal samples, we validated MSSM through combinatorial single-cell fluorescence in situ hybridisation (FISH) imaging and comparisons with macro-scale metagenomics, demonstrating its robustness and accuracy. MSSM captured fine-scale heterogeneity in taxonomic and functional composition across intestinal cryosections, hinting at spatially structured assemblages and segregation of metabolic capacities. Strain-level analyses uncovered coexisting Lawsonibacter lineages exhibiting distinct spatial distributions and host-specific occurrence patterns, while SNP-level microdiversity analyses showed that genetically coherent clonal populations cluster at spatial scales below [~]200 {micro}m. By enabling shotgun metagenomics at micron resolution, MSSM closes a longstanding methodological gap and provides a scalable platform for studying microbial ecosystems in situ. This approach unlocks a previously inaccessible view of microbial biogeography, offering new opportunities to investigate host-microbe and microbe-microbe interactions, and the spatial principles governing gut ecosystems. Significance statementUnderstanding how microbial communities are organised in space is essential to explaining their ecological and functional roles, yet microbiome research still relies overwhelmingly on bulk, spatially averaged measurements. We introduce micro-scale spatial metagenomics (MSSM), the first method that brings shotgun metagenomics to the microscale, enabling direct measurement of functional and taxonomic variation across regions containing as few as [~]100 cells. Unlike existing spatial approaches, MSSM reconstructs complete genomes and resolves strain-level diversity within intact tissue, allowing researchers to map metabolic potential, microdiversity, and community structure in situ. By coupling high-resolution sequencing with spatial context, MSSM reveals a previously inaccessible layer of microbial organisation, transforming how host-associated ecosystems can be studied.

microbiology↗

bioRxiv2025

A Daily Cycle of White Collar Complex Dephosphorylation Sustains Circadian Rhythmicity in Neurospora

As a photoreceptor, the transcription factor complex WCC acutely activates [~]5% of the genome in response to blue light, while as the circadian positive element in the dark WCC influences expression of about 40% of the transcriptome. Among WCC-regulated genes is frq which is both acutely light-activated through a pLRE and circadian-regulated through a C-box promoter element that is not active in constant light. The complex of FRQ, FRH, and CK1, the FFC, phosphorylates WCC at >95 sites, thereby repressing its activity and closing the circadian feedback loop in the dark. Although FFC has no described role in the light, we validated the expectation that FFC-driven WCC phosphorylation also silences C-box promoters in constant light, thereby confirming two classes of WCC targets, C-box-like that are normally repressed in the light and pLRE-like that remain light-active despite FFC-driven WCC phosphorylation. Genome-wide derepression of C-box-like promoters in frq-null fungi may explain reported non-circadian effects seen in some frq-null fungi including reduced virulence and conidiation. Reanalysis of WCC-mediated circadian activation and repression revealed that, while at dusk most WCC is phosphorylated and repressed, subsequent circadian activation is the result of transient dephosphorylation/derepression of just a small subset of this WCC pool; this small active pool drives expression of FRQ, nucleating the FFC, which rapidly re-phosphorylates the WCC pool to repress it, a phosphorylation/dephosphorylation cycle that can run for days without new WCC synthesis. The realization that both FFC and WCC are regulated primarily through phosphorylation rather than turnover leaves the circadian oscillator looking much like a phoscillator, emphasizing the primacy of post-translational regulation in timekeeping. SignificanceAt the core of circadian clocks of fungi and animals, a protein heterodimer drives expression of gene(s) whose products inactivate the heterodimer via phosphorylation. To sustain the cycle through multiple days, the activity of the heterodimer must be restored, but the means through which this happens have been unclear. In the clock model Neurospora, the White Collar Complex (WCC) is the heterodimer and FFC is the complex that inactivates it. We determined that WCC activity is restored principally by removal of the inhibitory phosphorylations and that for at least several days no new synthesis of WCC is required. The results confirm the existence of a large pool of inactive WCC in the cell and highlights the delicate balance between FCC-dependent phosphorylation/inactivation and phosphatase-dependent dephosphorylation/reactivation. Each morning, this balance allows transient activation of a fraction of the inactive WCC pool, thereby restarting the circadian cycle.

biochemistry↗

bioRxiv2025

Role of tankyrase scaffolding in the β-catenin destruction complex and WNT signaling

Aberrant WNT/{beta}-catenin signaling drives tumorigenesis and metastasis in cancer. Although enzymatic inhibitors of tankyrase (TNKS) effectively block AXIN degradation and stabilize the {beta}-catenin destruction complex (DC), they have demonstrated limited efficacy in various cancer models. Here we demonstrate that, unexpectedly, the induction of AXIN puncta represents a major barrier to achieving therapeutic efficacy. Mechanistically, catalytic inhibition of TNKS prevents TNKS turnover and drives its accumulation in the DC, wherein the scaffolding function of TNKS induces AXIN puncta formation, rigidifies the DC, and impedes {beta}-catenin turnover. Chemically induced degradation of TNKS overcomes this limitation by stabilizing AXIN without puncta formation, providing a deeper suppression of the WNT/{beta}-catenin pathway activity and the proliferation of colorectal cancer cells harboring dysfunctional APC mutations. Collectively, these findings provide an explanation for the unsatisfactory outcomes of drugging the WNT/{beta}-catenin signaling pathway by TNKS inhibitors and highlight TNKS degradation as a promising approach to treat WNT/{beta}-catenin-driven cancers.

cancer biology↗

bioRxiv2025

Elucidating the potential carcinogenic molecular mechanisms of parabens in head and neck squamous cell carcinoma through network toxicology and molecular docking

ObjectiveThis study aims to systematically investigate the molecular mechanisms through which parabens may contribute to head and neck squamous cell carcinoma (HNSCC) carcinogenesis using integrated network toxicology and molecular docking. Materials and MethodsSix commonly used parabens (ethyl-, propyl-, methyl-, heptyl-, butyl-, and benzylparaben) were selected for toxicity prediction via ProTox 3.0 and ADMETlab2.0. Their potential targets were retrieved from Swiss Target Prediction, ChEMBL, and the Similarity Ensemble Approach (SEA). HNSCC-related targets were collected from GeneCards, OMIM, and CTD, while differentially expressed genes (DEGs) were identified using TCGA-HNSC data. Functional enrichment, protein-protein interaction (PPI) network construction, hub gene identification, molecular docking, and immune infiltration analyses were performed using DAVID, STRING, Cytoscape, CB-DOCK2, and TIMER2.0. ResultsWe identified 80 common targets through which parabens may exert toxic effects in HNSCC. Among these, 12 hub genes (CCNB1, CDK1, CCNA2, CDK2, CDK4, TYMS, AURKA, CCNA1, CHEK1, CCNB2, PLK1, CDC25A) were significantly overexpressed in HNSCC tissues and were primarily enriched in cell cycle regulation, p53 signaling, and viral carcinogenesis pathways. CCNA1 was further validated as an independent prognostic factor associated with poor survival. Molecular docking revealed strong binding affinities between parabens and hub proteins. Immune infiltration analysis indicated a negative correlation between CCNA1 expression and CD8+ T cell and B cell infiltration. ConclusionsParabens may promote HNSCC progression by disrupting cell cycle regulation and immune responses via direct interactions with key hub genes. These findings provide a novel mechanistic basis for the carcinogenic potential of parabens in HNSCC and underscore the need for further experimental and epidemiological validation.

pharmacology and toxicology↗

bioRxiv2025

Modular inhibitory coding in binary networks

We developed and characterized properties of new class of binary models where, as observed in biological networks, excitatory neurons are functionally separated from inhibitory units. New patterns, represented as activation and inactivation of binary units in excitatory layer, are stored in the network through recruitment and training of inhibitory units that are group into individual modules and interact with excitatory layer. We investigate the roles the two populations play in memory storage and show that inhibitory layer plays a critical role in memory storage and management and that capacity of this new type of network scales with number of inhibitory neurons. Further, we show that performance of the network is only gradually diminished when excitatory to excitatory connections are removed, but critically depends on inhibitory to excitatory connections. These results are in line with new experimental work showing that inhibitory interneurons are playing critical role in memory storage and recall in the brain networks and may also address why generally excitatory networks exhibit sparser reciprocal connectivity as compared to connections to/from inhibitory units. We further show advantages of so designed coding shame in terms of memory capacity, its expansion with progressive storage of new memories as well as network behavior for large memory loading.

neuroscience↗

bioRxiv2025

CXCR6⁺ CD127- Tr1 Cells Balance Immunity and Persistence in Plasmodium falciparum Infection

Plasmodium falciparum (Pf) induces the clonal expansion of antigen-specific type 1 regulatory T (Tr1) cells with a capacity for long-term memory. Tr1 cells comprise nearly 90% of the Pf blood stage antigen-specific CD4+ T cell pool in children. Though, whether Tr1 cells contribute to protection from malaria remains undetermined. To address this critical gap in knowledge, we first performed scRNAseq on gated cell populations to validate CXCR6+ CD127- as new phenotypic markers to enrich for bona-fide Tr1 cells. Importantly, these Tr1 cells potently suppressed the proliferation of other CD4+ T cells in vitro via IL-10 secretion. Among children living in malaria-endemic Uganda, CXCR6+ CD127- Tr1 cells were the dominant responding subset to Pf-infected red blood cell stimulation in vitro. They also rapidly expanded following malaria and expressed IL-10 and IFNg during infection in vivo. Tr1 abundance correlated with plasma concentrations of granzyme A, IFNg, IL-10, and LAG3, suggesting that these cells act systemically. Higher CXCR6+ CD127- Tr1 cell frequencies correlated with a lower probability of symptoms given parasitemia but were also associated with delayed parasite clearance among untreated, asymptomatic children. These data suggest that Tr1 cells help mediate clinical immunity to malaria but may also facilitate parasite persistence through mechanisms of immune regulation.

immunology↗

bioRxiv2025

BioML-bench: Evaluation of AI Agents for End-to-End Biomedical ML

Large language model (LLM) agents hold promise for accelerating biomedical research and development (R&D). Several biomedical agents have recently been proposed, but their evaluation has largely been restricted to question answering (e.g., LAB-Bench) or narrow bioinformatics tasks. Presently, there remains a lack of benchmarks evaluating agent capability in multi-step data analysis workflows or in solving the machine learning (ML) challenges central to AI-driven therapeutics development, such as perturbation response modeling or drug toxicity prediction. We introduce BioML-bench, the first benchmarking suite for evaluating AI agents on end-to-end biomedical ML tasks. BioML-bench spans four domains (protein engineering, single-cell omics, biomedical imaging, and drug discovery) with tasks that require agents to parse a task description, build a pipeline, implement models, and submit predictions graded by established metrics (e.g., AUROC, Spearman). We evaluate four open-source agents: two biomedical specialists (STELLA, Biomni) and two generalists (AIDE, MLAgentBench). On average, agents underperform relative to human baselines, and biomedical specialization does not confer a consistent advantage. We also found that agents which employed more diverse ML strategies more often tended to score highest, suggesting that architecture and scaffolding may be stronger determinants of performance. These findings underscore both the potential and current limits of agentic systems for biomedical ML, and highlight the need for systematic, reproducible evaluations. BioML-bench is provided open-source at github.com/science-machine/biomlbench.

bioinformatics↗

bioRxiv2025

Molecular Architecture and Function Mechanism of Tri-heteromeric GluN1-N2-N3A NMDA Receptors

N-methyl-D-aspartate receptors (NMDARs) play a pivotal role in brain development and synaptic function. Previous studies have focused on GluN1-N2 (2A-2D) and GluN1- N3 (3A and 3B) di-heteromeric (di-) NMDARs, leaving the activation mechanism and stoichiometry of GluN1-N2-N3 tri-heteromeric (tri-) NMDARs largely unexplored. In this study, we employed a two-step affinity-tagged chromatography approach to purify recombinantly expressed GluN1-N2A-N3A tri-NMDARs and determined their cryo-EM structure. Based on the proteoliposome single-channel recording, we discovered GluN1-N2A-N3A can be activated upon co-binding of glycine and glutamate, exhibiting two distinct conductance levels. Furthermore, leveraging structural-based click- chemistry, we introduced photo-crosslinker p-azido-phenylalanine (AzF) into the N- terminal domain of GluN2A and GluN2B, enabling the crosslinking with GluN3A subunit both in vitro and in vivo. These findings provide molecular insights into the subunit arrangement, native architecture and activation mechanism of GluN1-N2-N3A tri- NMDARs and also highlight the complexity of NMDAR assembly and function in the brain.

neuroscience↗

bioRxiv2025

Cross-regulation of (p)ppGpp and c-di-GMP pathways controls a cell-cycle transition

Bacteria must adapt their physiology to changing environments, a process often orchestrated by nucleotide messengers such as (p)ppGpp and cyclic di-GMP (c-di-GMP). While (p)ppGpp promotes survival under nutrient stress by suppressing growth, c-di-GMP supports sessile lifestyles and cell differentiation. In the -proteobacterium Caulobacter crescentus, these messengers have opposite effects on the cell cycle, suggesting potential cross-regulations. Here, we uncover direct molecular cross-talk between (p)ppGpp and c-di-GMP signaling pathways. Using biochemical, structural and genetic approaches, we show that (p)ppGpp binds to the diguanylate cyclase PleD and inhibits c-di-GMP synthesis by competing with GTP at the active site. This interaction prevents c-di-GMP accumulation under nutrient-limiting conditions, linking metabolic stress to inhibition of developmental progression. Our findings reveal a new paradigm for second messenger interplay, in which one messenger directly suppresses synthesis of another. This work provides insight into how bacteria integrate metabolic and developmental cues through small-molecule cross-talk to guide adaptive behavior.

microbiology↗

bioRxiv2025

Structural and mechanistic insights into bacterial hydrazine biosynthesis

Nitrogen-nitrogen (N-N) bond-containing motifs are prevalent in both clinical drugs and natural products. Bacterial hydrazine synthetases catalyze N-N bond formation by coupling an amino acid and a hydroxylamine via a distinctive O-aminoacyl-hydroxylamine intermediate. Despite its wide occurrence, the structural and mechanistic basis of this process has remained elusive. Here, we report the first crystal structures of the O-aminoacyl-hydroxylamine synthetase component of a bacterial hydrazine synthetase, captured in binary and ternary complexes with substrates and catalytic intermediates. These structures reveal the molecular determinants of substrate recognition and, together with biochemical and computational analyses, establish a detailed mechanistic framework for the hydrazine synthetase family. Leveraging these insights, we expanded hydrazine biosynthesis through the targeted discovery of novel hydrazine synthetases and implemented a chemoenzymatic synthesis strategy. This work provides fundamental insights into hydrazine synthetases and lays the groundwork for their rational engineering as versatile biocatalysts.

microbiology↗

bioRxiv2025

DNA Sequence Trace Reconstruction Using Deep Learning

Deciphering DNA sequences is fundamental to unlocking the mysteries of life, but the high dimensionality and complexity of biological sequence data significantly hinder knowledge discovery. In particular, the challenges of sequence length, repetitive regions, and structural complexity make it difficult to directly reconstruct complete DNA sequences from raw data. Therefore, this paper proposes a DNA sequence trace reconstruction model, DNARetrace, which performs preprocessing and dataset construction, and then employs a Bidirectional Fourier-Kolmogorov-Arnold Network (Bi-FKGAT), using an extremely unbalanced loss function for link prediction, so as to reconstruct the original DNA sequence. In multi-angle experiments using both simulated and real data, DNARetrace successfully reconstructs DNA sequence traces across large-scale datasets derived from various DNA sequencing methods, overcoming the bias of current approaches toward specific sequencing platforms, and achieves competitive outcomes in DNA storage and genomics downstream tasks. We further validated the expandability of the proposed methods in DNA sequence classification and metagenomic binning tasks. In summary, DNARetrace is compatible with various sequencing scenarios; it reduces the difficulty of discovering novelty knowledge directly from high-complexity raw data, and it provides a reusable tool to accelerate DNA sequence processing and applications.

genomics↗

bioRxiv2025

Single pointwise samples of electric field on a neuron model cannot predict activation threshold by brain stimulation

BackgroundSome computational models of neural activation by transcranial magnetic stimulation overestimate the electric field (E-field) threshold compared to in vivo measurements. A recent study proposed a statistical method to account for the influence of microscopic perturbations to the E-field. The method, however, relies on the unsubstantiated assumption that thresholds can be predicted by single pointwise samples of the E-field strength along neural cables. ObjectiveWe analyzed neural responses to E-field with microscopic perturbations and demonstrate via theoretical derivation and simulations that neural activation is not determined by pointwise E-field amplitude but rather by spatial integration of the E-field along the neural cable. Therefore, the influence of microscopic E-field perturbations is negligible due to the spatiotemporal filtering by the neural membrane and axoplasm. MethodsWe derive the axial and transmembrane currents in a neural cable for an imposed E-field with microscopic perturbations. We simulate neural activation thresholds of unmyelinated and myelinated axons in two stimulation scenarios and compare thresholds for E-field activation with and without perturbations. ResultsIn the theoretical derivation, the perturbation terms average out to zero on larger spatial scales indicating that they do not influence neural activation thresholds. Simulated thresholds with the E-field spatial perturbations present had negligible differences (< 3.4%) compared to those without. ConclusionSingle point samples of the microscopic E-field on a neural cable cannot predict neural activation thresholds. Neural simulations should be used to determine any influence of the E-field spatial perturbations. The latter, however, are unlikely to account for the difference between experimental and simulated E-field thresholds.

bioengineering↗

bioRxiv2025

EphA2 and Ephrin-A1 Use the Same Interface for Both in cis and in trans Interactions That Regulate Distinct Cell Signaling and Function

The 14 members of Eph receptor tyrosine kinases (RTK) bind to membrane-tethered ligand called ephrins and mediate cell contact signaling where the receptors and ligands engage in trans on adjacent cells. Previous studies reveal that some Eph and ephrin pairs are coexpressed on the same cells, including EphA3-ephrin-A3 and EphA4/ephrin-A5, can also interact with each other in cis. However, significant discrepancies persist as to the molecular basis and functional significance of the cis interactions, owning to the difficulties to directly interrogate the interactions. Here, we utilize time-resolved live cell fluorescence spectroscopy to demonstrate direct in cis interactions between EphA2 and ephrin-A1. Structure-guided mutagenesis mapped interactions to two salt bridges between the ligand- and receptor-binding domains of EphA2 and ephrin-A1. Interestingly, the same interface is shared with in trans interaction. Consequently, EphA2-ephrin-A1 interaction in cis competes with their interaction in trans, which leads to attenuation of EphA2 canonical signaling and inhibition of cell rounding when ligand is presented in trans. EphA2 and ephrin-A1 are widely coexpressed in many epithelial tissues, and dysregulation of their expression is known to contribute to tumor initiation and progression. The detailed molecular characterization of the mutually exclusive cis and trans interactions uncovers a new mechanism underpinning their unique roles in oncogenesis. Significance StatementEphA2 exerts dual functions in tumorigenesis, depending on the binding status of its membrane-tethered ephrin-A ligands. Ligands presented in trans on adjacent cells activate EphA2 canonical signaling associated with tumor suppression, whereas loss of ligand expression promotes oncogenic noncanonical signaling of EphA2 via serine 897 phosphorylation. Combining time-resolved spectroscopy in live cells, structure-guided mutagenesis, we show strong in cis interactions between EphA2 and ephrin-A1, which shares the same interface as interaction in trans. Moreover, the cis interaction interferes with ligand binding in trans, attenuates EphA2 canonical signaling. Our results uncover a new mechanism of EphA2 regulation by its co-expressed ligand ephrin-A1 with important implications in its known roles in oncogenesis as well as other disease processes including development of cataract.

biophysics↗

bioRxiv2025

What is memorable across the Pacific Ocean shows evidence of prior learning on the memorability of new visual inputs

It has been proposed that certain objects in our environment are inherently memorable due to fundamental similarities in how the human brain stores this visual information. Here we tested this hypothesis by having two groups of participants from opposite sides of the planet remember pictures of objects shown individually from several seconds each. We found that the specific objects that were more memorable varied across groups, with fine-grained analysis of the nonoverlapping objects indicating that participants have greater facility storing objects that are familiar to them as they occur in their cultural context. Thus, the present findings demonstrate that the memorability of a stimulus is jointly determined by its features and the learning history of the human brain in which it is being stored.

animal behavior and cognition↗

bioRxiv2025

A Photo-regulatable Intein Based Trans-Splicing tool for Protein and Organelle Relocalization to Different Subcellular Compartments

The dynamics of protein subcellular localization are intricately regulated, requiring new tools for controlling protein translocation to uncover the biological significance of post-translational regulation of protein localization. Here, we developed a new method, Protein Rerouting via INtein-mediated Trans-Splicing (PRINTS), which enables precise control of protein translocation across diverse subcellular compartments. By reconstituting functional signaling peptides, PRINTS can efficiently relocalize fluorescent proteins to the 26S proteasome, nucleus, mitochondria, plasma membrane, endomembrane organelles, and liquid-liquid phase separation (LLPS) condensates. Furthermore, we incorporated the optically regulated dimerization domain CRY2clust into the PRINTS system, achieving light-mediated control of protein entry into the cell membrane and LLPS membraneless compartments. Strikingly, we observed that HNRNPA1 promiscuously recruit CRY2- or intein-containing proteins into LLPS condensate, whereas FXR1, another LLPS protein showed minimal non-specific activity. PRINTS also allow organelle relocalization, such as LLPS condensates or mitochondria to cell membrane in a light-controllable manner. Overall, PRINTS provides a versatile and robust platform for manipulating protein and organelle subcellular translocations, offering a powerful tool to investigate the regulatory coordination and crosstalk between membranous and membraneless compartments in response to physiological needs.

cell biology↗

bioRxiv2025

Scientifically Uninformed Legislation System Undermines Ecological Sustainability of the Yangtze River

Legal frameworks typically aim to protect environmental systems through regulations, treaties, and conservation measures. However, without scientifically informed enforcement, these policies may inadvertently exacerbate environmental degradation, particularly when addressing complex socio-environmental issues. We reviewed the development of Chinas environmental legislation related to ecosystem protection in the Yangtze River watershed and evaluated its effectiveness by analyzing illegal fishing cases from 2015 to 2021, prosecuted under Chinas Environmental Public Interest Litigation framework. Our findings reveal that court-ordered fish release remedies often undermine, rather than support, the sustainability of the Yangtze River ecosystem. This is primarily due to mismatches between the taxonomic identity of illegally caught fishes and those specified for release, species misidentification, and a bias toward economically valuable species in court-ordered release decisions. These unintended discrepancies between environmental protection goals and actual ecological outcomes underscore the need to integrate scientific expertise into legal practice to enhance the sustainability of socio-environmental systems.

ecology↗

bioRxiv2025

Alterations in Electroencephalography Signals in Female Fragile X Syndrome Mouse Model on a C57Bl/6J Background

BackgroundFragile X Syndrome (FXS), the most common monogenic cause of autism spectrum disorder, arises from FMR1 gene silencing and exhibits pronounced sex differences in prevalence and phenotypic severity. Electroencephalography (EEG) has emerged as a promising translational biomarker for FXS pathophysiology, yet prior research has predominantly focused on male cohorts. In the widely used C57Bl/6J (B6) mouse strain, male Fmr1 knockout (KO) models show increased absolute gamma power at both juvenile and adult stages, which may reflect cortical hyperexcitability. In contrast, little is known about female Fmr1 KO mice, except that they exhibit no gamma alterations in adulthood. This gap hinders understanding of sex-specific neurodevelopmental trajectories of EEG profile in FXS. Leveraging the genetic stability and translational relevance of the B6 strain, this study compares EEG profiles between juvenile female Fmr1 KO and wild-type (WT) B6 mice to address this critical gap. MethodsFrontal-parietal differential EEG was recorded in freely behaving mice using the Open-Source Electrophysiology Recording system for Rodents. Neural activity was analyzed across three recording conditions: in the home cage, light-dark arena, and open field arena. Computed metrics included absolute/relative power, peak alpha frequency, theta-beta ratio, phase-amplitude coupling, amplitude-amplitude coupling, and multiscale entropy to assess signal complexity. ResultsIn all recording conditions, Fmr1 KO mice exhibited reduced absolute power in theta, alpha, and beta frequency bands compared to WT controls. Relative power analysis revealed decreased alpha activity alongside increased gamma-band power, including both low and high gamma, in the KO mice. Cross-frequency coupling was disrupted, with diminished alpha-gamma phase-amplitude coupling. Amplitude-amplitude coupling between theta or alpha and gamma power displayed distinct changes in different recording conditions. Peak alpha frequency and theta-beta ratio were both reduced or unchanged in the KO mice, depending on the recording condition. Finally, EEG signal complexity remained comparable between the two genotypes across the conditions. Behaviorally, KO mice displayed hyper-exploration in the open field test, characterized by increased center time and entries. However, no overall robust correlations between EEG power in different frequency bands and behavioral parameters in the open field test were observed. Discussion and ConclusionOur results demonstrate that juvenile female Fmr1 KO mice on a B6 background exhibit EEG alterations highly consistent with those reported in FXS patients, particularly increased gamma and reduced alpha power. The robust increase in gamma activity reinforces its status as a reliable biomarker across preclinical and clinical studies, while alpha reductions and slowed peak alpha frequency implicate thalamocortical network involvement. Together, these findings highlight the translational value of this model for studying core circuit dysfunctions in FXS.

neuroscience↗

bioRxiv2025

mRNABench: A curated benchmark for mature mRNA property and function prediction

Messenger RNA (mRNA) is central in gene expression, and its half-life, localization, and translation efficiency drive phenotypic diversity in eukaryotic cells. While supervised learning has widely been used to study the mRNA regulatory code, self-supervised foundation models support a wider range of transfer learning tasks. However, the dearth and homogeneity of standardized benchmarks limit efforts to pinpoint the strengths of various models. Here, we present O_SCPLOWMC_SCPLOWRNABO_SCPLOWENCHC_SCPLOW, a comprehensive benchmarking suite for mature mRNA biology that evaluates the representational quality of mature mRNA embeddings from self-supervised nucleotide foundation models. We curate ten datasets and 59 prediction tasks that broadly capture salient properties of mature mRNA, and assess the performance of 18 families of nucleotide foundation models for a total of 135K experiments. Using these experiments, we study parameter scaling, compositional generalization from learned biological features, and correlations between sequence compressibility and performance. We identify synergies between two self-supervised learning objectives, and pre-train a new Mamba-based model that achieves state-of-the-art performance using 700x fewer parameters. O_SCPLOWMC_SCPLOWRNABO_SCPLOWENCHC_SCPLOW can be found at: https://github.com/morrislab/mRNABench.

genomics↗