Biology subjects

Zhang, T.

Publications and source records attributed to Zhang, T..

At least 199 records · Page 11Linked to original sources

bioRxiv2025

Angel Wing: a class of alternative splicing regulators in animals

Alternative splicing (AS) regulates the diversity and level of the proteome. The specificity in AS is in turn regulated by RNA binding proteins, but our understanding of how they act is far from complete. Here we identify the Angel wing (Anw) protein as a novel AS regulator. Loss of Anw in Drosophila disrupts splicing in muscle genes and subsequently muscle function. Based on a mini-gene assay in which Anw and its RNA targets are co-expressed in cultured cells, we demonstrated orthologous splicing regulation of the minigene transcripts, interaction between Anw and its RNA targets, and a remarkable functional conservation among Anw homologs. Anw forms nuclear foci, and genetic ablation of Anw domains suggests that maintaining distinctive features of these foci is important for its function. The evolution of Anw is dynamic with gene gains and losses, but preserves a cross-phyla "ultra conserved element" as an alternative exon that potentially regulates its own level by non-sense mediated mRNA decay. As the human anw homolog is a candidate gene for myasthenia gravis, our work suggests a mechanism for cellular dysfunction in this disease.

genetics↗

bioRxiv2025

Dysregulated Proline Metabolism Contributes to Subretinal Fibrosis in Neovascular AMD: Therapeutic Potential of Prolyl-4-Hydroxylase Inhibition

Subretinal fibrosis, a major cause of irreversible vision loss in neovascular age-related macular degeneration (nAMD), is driven by excessive deposition of extracellular matrix such as collagens. While proline metabolism is known to play a critical role in collagen biosynthesis and fibrosis, its involvement in subretinal fibrosis remains unclear. Here, we characterized the progression of fibrovascular lesions in JR5558 mice, observing significant molecular alterations as early as 4 weeks of age and phenotypic changes by 8 weeks. Transcriptomic and metabolomic analyses revealed elevated levels of 4-hydroxyproline, an essential component of collagen, alongside significant alterations of other fibrosis-related pathways. P4HA1, a catalytic subunit of prolyl-4-hydroxylase essential for 4-hydroxyproline biosynthesis, was prominently expressed in fibrotic lesions in retinas of JR5558 and two-stage laser-induced murine models, as well as human eyes with nAMD. Targeting P4HA1 with the small-molecule inhibitor diethyl pythiDC significantly attenuated fibrovascular lesion growth in the JR5558 murine models and reduced collagen turnover in human retinal pigment epithelium cells. Combining diethyl pythiDC with aflibercept had a stronger antifibrotic effect than monotherapies in JR5558 mice. These findings suggest a key contribution of proline metabolism, particularly proline hydroxylation, in subretinal fibrosis. Inhibiting P4HA1 with diethyl pythiDC inhibited fibrosis in the models we studied, offering a novel therapeutic strategy. Further research is warranted to explore the potential benefits of combining existing anti-angiogenic therapies with drugs that inhibit proline metabolism for the management of nAMD-associated fibrosis.

neuroscience↗

bioRxiv2025

Transfer Learning and Permutation-Invariance improving Predicting Genome-wide, Cell-Specific and Directional Interventions Effects of Complex Systems

With the advent of precision medicine, single-drug treatments may not fully satisfy the pursuit of precision medicine. However, single-drug treatments have accumulated a large amount of data and a considerable number of deep learning models. In this context, using transfer learning to effectively leverage the existing vast amount of single-compound intervention effect data to build models that can accurately predict the intervention effects of complex systems is highly worth investigating. In this study, we used a deep model based on permutation-invariance as the core module, pre-trained on a large amount of single-compound intervention data in cell lines, and fine-tuned on a small amount of complex system (like natural products) intervention data in cell lines, resulting in a predictive model named SETComp (the Concat version with ~200M parameters and the Add version with ~173M parameters). The two versions of SETComp achieved an accuracy of 93.86% and 92.70%, respectively, on the complex system-cell-gene association test set, improving by 5.82% to 27.59% compared to the baseline. When predicting the intervention effects of those complex systems the model had never encountered before, the accuracy increased by up to 24.83% compared to the baseline. In our in vitro validation, up to 88.65% of the predictions were confirmed to be correct, and the models output showed a significant positive correlation with the real-world foldchange. We further observed SETComps potential in various biomedical scenarios, achieving good performance in applications such as mechanism uncovering, repositioning, and compound synergy discovery.

systems biology↗

bioRxiv2025

β-Arrestin Condensates Regulate G Protein-Coupled Receptor Function

G protein-coupled receptors (GPCRs) are the largest class of receptors in the genome and control many signaling cascades essential for survival. GPCR signaling is regulated by {beta}-arrestins, multifunctional adapter proteins that direct receptor desensitization, internalization, and signaling. While at many GPCRs, {beta}-arrestins interact with a wide array of signaling effectors, it is unclear how {beta}-arrestins promote such varied functions. Here we show that {beta}-arrestins undergo liquid-liquid phase separation (LLPS) to form condensates that regulate GPCR function. We demonstrate that {beta}-arrestin oligomerization occurs in proximity to the GPCR and regulates GPCR functions such as internalization and signaling. This model is supported by a cryoEM structure of the adhesion receptor ADGRE1 in a 2:2 complex with {beta}-arrestin 1, with a {beta}-arrestin orientation that can promote oligomerization. Our work provides a paradigm for {beta}-arrestin condensates as regulators of GPCR function, with LLPS serving as an important promoter of signaling compartmentalization at GPCRs.

cell biology↗

bioRxiv2025

The sialylation pattern of dietary glycopeptides is vital for inhibiting pathogens and modulating gut microbiota in elderly individuals

N-glycosylation is a post-translational modification of proteins that can influence the structure and function of glycoproteins. Edible birds nest (EBN) contains highly sialylated mucins with N-glycans and O-glycans, while casein glycomacropeptide (CGMP) is modified with only O-glycans. We prepare EBN sialoglycopeptides through sequential protease treatment and their inhibitory effects on pathogens, including Candida albicans, Helicobacter pylori, and avian influenza virus (AIV), as well as their modulation of gut microbiota in elderly individuals, were compared with CGMP. N-glycan analysis revealed the presence of multiantenna N-glycans, characterized by sialylation and bisecting GlcNAc in the EBN glycopeptides. EBN glycopeptides significantly inhibited C. albicans and H. pylori. However, the removal of terminal sialic acid from EBN glycopeptides resulted in a marked reduction of this inhibitory effect. Using Maackia Amurensis Lectin II as an AIV mimic, EBN glycopeptides effectively inhibited AIV adhesion to MDCK cells. In vitro fecal fermentation using fecal samples from elderly individuals revealed that the EBN glycopeptide group exhibited remarkable probiotic effects on the gut microbiota, increasing the relative abundance of beneficial bacteria such as Lactobacillus and Roseibacterium, while decreasing the relative abundance of harmful bacteria such as Escherichia-Shigella and Klebsiella. Overall, EBN glycopeptide, rich in sialylated N-glycans and O-glycans, strongly inhibited pathogenic microorganisms and showed superior probiotic properties compared to CGMP with sole O-glycosylation.

biochemistry↗

bioRxiv2025

Loss of Kmt2c/d promotes gastric cancer initiation and confers vulnerability to mTORC1 inhibition and anti-PD1 immunotherapy

KMT2C and KMT2D (KMT2C/D) are frequently mutated in gastric adenocarcinoma, yet their function in cancer initiation remains poorly understood. In this study, based on the observation that loss-of-function mutations of KMT2C and KMT2D are enriched and co-occur in gastric adenocarcinoma, we developed genetically engineered mouse models to selectively knock out Kmt2c and Kmt2d in gastric epithelial cells with Tmprss2-CreERT2. Through histological staining and single-cell RNA sequencing, we observed that Kmt2c/d loss led to nuclear dysplasia and expansion of cells with mixed gastric lineage markers. When combined with Pten deletion, Kmt2c/d loss drove rapid development of muscle-invasive gastric adenocarcinoma as early as 3 weeks post Cre-mediated gene deletion. The adenocarcinoma exhibited decreased expression of gastric lineage markers and increased expression of intestinal differentiation markers, phenocopying human gastric adenocarcinoma. Kmt2c/d knockout reduced protein synthesis but upregulated transcription of ribosomal proteins, rendering sensitivity to mTORC1 inhibitors. Additionally, Kmt2c/d knockout increased MHC-I molecule expression and enhanced antigen presentation. Combination of mTROC1 inhibition and anti-PD1 immunotherapy significantly suppressed tumor growth in immune-competent mice. Together, these findings reveal the role of Kmt2c/d loss in gastric cancer initiation and suggest the potential therapeutic strategies for KMT2C/D-deficient gastric cancer.

cancer biology↗

bioRxiv2025

Cerebellin-2 positive (Cbln2+) Neurons in the Parafascicular Thalamic Nucleus Regulate Self-Grooming Behavior

Self-grooming represents a stereotyped repetitive behavior which is commonly observed in various neuropsychiatric disorders. Deep brain stimulation (DBS) targeting the centromedian / parafascicular complex (CM/PF) has been shown to alleviate obsessive-compulsive disorder/behavior (OCD/OCB) symptoms in human. However, little is known about the neural circuits of the PF that are involved in the regulation of self-grooming behavior in rodents. Here, we report that cerebellin-2 positive (Cbln2+) neurons in PF bidirectionally encode self-grooming. Chemogenetic activation of Cbln2+ PF neurons significantly reduces both physiological and pathological self-grooming, and chemogenetic inhibition of Cbln2+ PF neurons increases physiological self-grooming. Moreover, activating synaptic excitatory inputs to Cbln2+ PF neurons inhibits self-grooming, while synaptic inhibitory inputs enhance it. Cbln2+ PF neurons independently transmit neural signals to the dorsal and ventral striatum. Activation of PF-dorsal striatum and PF-ventral striatum pathways both inhibit self-grooming behavior. Most excitingly, we alleviate excessive grooming in OCD mice by the activation of PF neurons using non-invasive ultrasound stimulation. Together, these data reveal that Cbln2+ PF neurons are integral components of a brain network involved in self-grooming behavior and highly promising targets in the non-invasive therapy of OCD.

neuroscience↗

bioRxiv2025

FMR1 KH0-KH1 Domains Coordinate m6A Binding and Phase Separation in Fragile X Syndrome

The Fragile X Mental Retardation Protein (FMR1) regulates neurodevelopment via m6A RNA interactions, but domain-specific roles of KH0 and KH1 in RNA binding and pathology remain unclear. Through mutagenesis and AlphaFold3 modeling, we demonstrate that KH1 serves as the primary m6A-binding interface, whereas the KH0 domain (e.g., Arg138) modulates liquid-liquid phase separation (LLPS). Pathogenic mutations in KH0 disrupt RNA binding and enhance LLPS aggregation, while m6A RNA suppresses LLPS at KH0. Structural simulations reveal synergistic cooperation between KH0 and KH1 through hydrophobic and electrostatic networks. These dual-domain interactions establish a mechanistic link between m6A dysregulation, aberrant phase separation, and Fragile X syndrome pathogenesis. Our findings propose KH0 as a therapeutic target for RNA-driven neurodevelopmental disorders.

cell biology↗

bioRxiv2025

A Bivalent Molecular Glue Linking Lysine Acetyltransferases to Oncogene-induced Cell Death

Developing cancer therapies that induce robust death of the malignant cell is critical to prevent relapse. Highly effective strategies, such as immunotherapy, exemplify this observation. Here we provide the structural and molecular underpinnings for an approach that leverages chemical induced proximity to produce specific cell killing of diffuse large B cell lymphoma, the most common non-Hodgkins lymphoma. We develop KAT-TCIPs (lysine acetyltransferase transcriptional/epigenetic chemical inducers of proximity) that redirect p300 and CBP to activate programmed cell death genes normally repressed by the oncogenic driver, BCL6. Acute treatment rapidly reprograms the epigenome to initiate apoptosis and repress c-MYC. The crystal structure of the chemically induced p300-BCL6 complex reveals how chance interactions between the two proteins can be systematically exploited to produce the exquisite potency and selectivity of KAT-TCIPs. Thus, the malignant function of an oncogenic driver can be co-opted to activate robust cell death, with implications for precision epigenetic therapies.

cancer biology↗

bioRxiv2025

Structural variants contribute to the gut microbiome, blood metabolic traits and their causal relationships

Research on genetic variants affecting the gut microbiome and blood metabolites has mainly focused on single nucleotide variants (SNPs), with structural variants (SVs) remaining largely unexamined. In this study, we utilized an advanced structural variant detection pipeline on whole-genome sequencing data from 2,002 individuals with an average sequencing depth of 42 folds, identifying 138,859 high-confidence SVs, 49.57% of which were previously unreported. We performed genome-wide association studies on 22,519 common SVs with 616 gut microbial features and 121 blood metabolic traits, uncovering 30 significant links between SVs and microbial taxa/pathways and 48 connections between SVs and blood metabolites. Notable results include 19 associations reaching study-wide significance, such as an Asian-specific 6.5k deletion at the TENT2 gene, which is strongly associated with the abundance of Coprobacillus sp. 29_1 ({beta}=0.158, p=2.5x10-9), and a deletion covering HBA1/HBA2/HBM/HBQ118 linked to nine blood metabolic traits, most significantly with mean corpuscular volume (MCV; {beta}=-0.570, p=1.22x10-91). Most SVs showed a greater impact than nearby SNPs in their associations with the microbiome/metabolites. Analysis of rare SVs also identified multiple genes associated with the microbiome/metabolites. Additionally, Mendelian randomization using SVs demonstrated causal relationships, including effects from the microbiome to metabolites, such as Clostridium ramosums impact on serum uric acid and hyperuricemia, and from metabolites to the microbiome, like the influence of mean corpuscular hemoglobin concentration on MF0064: triacylglycerol degradation. This study provides a detailed SV catalog for the Chinese population and highlights the significant role of SVs in the gut microbiome, blood metabolites, and their causal relationships. HighlightsO_LIProvides a catalog of 138,859 high-confidence SVs and 49.57% of which were novel C_LIO_LIIdentify 30 genome-wide significant "SVs--microbial taxa/pathways" and 48 "SVs--blood metabolites" associations, respectively C_LIO_LIAn Asian-specific 6.5k deletion at TENT2 gene associated with the species Coprobacillus sp. 29_1 C_LIO_LIA deletion spanning HBA1/HBA2/HBM/HBQ118 associated with 9 blood metabolic traits C_LIO_LIThe majority of SVs demonstrated a more pronounced impact than neighboring SNPs in the associations with microbiome/metabolites C_LIO_LISVs help to reveal the host-microbe causal relationships such as Clostridium ramosums causal effects on serum uric acid and hyperuricemia C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=155 SRC="FIGDIR/small/643427v2_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@e660d0org.highwire.dtl.DTLVardef@ddfe4dorg.highwire.dtl.DTLVardef@62a704org.highwire.dtl.DTLVardef@c31bb5_HPS_FORMAT_FIGEXP M_FIG Graphic abstract Based on a multi-omics cohort of 2,002 individuals comprising high-depth whole genome and metagenomic sequencing data, blood metabolites, and detailed questionnaires, we identified 138,859 high-confidence SVs, uncovering 36 significant associations between SVs and microbial taxa/pathways, as well as 48 associations between SVs and blood metabolites. Mendelian randomization using SVs demonstrated causal relationships, such as Clostridium ramosums causal effect on serum uric acid and hyperuricemia. This study provides a detailed SV catalog for the Chinese population and highlights the significant role of SVs in the gut microbiome, blood metabolites, and their causal relationships. C_FIG

genomics↗

bioRxiv2025

Deciphering lung adenocarcinoma evolution and the role of LINE-1 retrotransposition

Understanding lung cancer evolution can identify tools for intercepting its growth. In a landscape analysis of 1024 lung adenocarcinomas (LUAD) with deep whole-genome sequencing integrated with multiomic data, we identified 542 LUAD that displayed diverse clonal architecture. In this group, we observed an interplay between mobile elements, endogenous and exogenous mutational processes, distinct driver genes, and epidemiological features. Our results revealed divergent evolutionary trajectories based on tobacco smoking exposure, ancestry, and sex. LUAD from smokers showed an abundance of tobacco-related C:G>A:T driver mutations in KRAS plus short subclonal diversification. LUAD in never smokers showed early occurrence of copy number alterations and EGFR mutations associated with SBS5 and SBS40a mutational signatures. Tumors harboring EGFR mutations exhibited long latency, particularly in females of European-ancestry (EU_N). In EU_N, EGFR mutations preceded the occurrence of other driver genes, including TP53 and RBM10. Tumors from Asian never smokers showed a short clonal evolution and presented with heterogeneous repetitive patterns for the inferred mutational order. Importantly, we found that the mutational signature ID2 is a marker of a previously unrecognized mechanism for LUAD evolution. Tumors with ID2 showed short latency and high L1 retrotransposon activity linked to L1 promoter demethylation. These tumors exhibited an aggressive phenotype, characterized by increased genomic instability, elevated hypoxia scores, low burden of neoantigens, propensity to develop metastasis, and poor overall survival. Re-activated L1 retrotransposition-induced mutagenesis can contribute to the origin of the mutational signature ID2, including through the regulation of the transcriptional factor ZNF695, a member of the KZFP family. The complex nature of LUAD evolution creates both challenges and opportunities for screening and treatment plans.

genomics↗

bioRxiv2025

Red Light-Activated Reversible Inhibition of Protein Functions by Assembled Trap

Red light offers strong tissue penetration and low phototoxicity, making it an attractive option for optogenetic applications. However, the available red light-inducible optogenetic tools are rather limited. Here, we present a novel red light-induced protein clustering system, namely R-LARIAT (Red Light-Activated Reversible Inhibition by Assembled Trap), that controls protein functions in a spatiotemporal manner. Our system capitalizes on the rapid and reversible binding of LDBs (nanobodies-based dimerization binders) to a bacterial phytochrome DrBphP, which uses a mammalian endogenous biliverdin chromophore to absorb red light. An anti-GFP nanobody fused with LDBs allows this method to quickly trap a wide variety of GFP-tagged proteins into light-induced protein clusters. Strikingly, our system exhibits an excellent performance in clustering efficiency with high light sensitivity and stability, it can function even when shielded by multiple glass plates. By utilizing the R-LARIAT system to trap and sequester tubulin, cell cycle progression can be blocked in HeLa cells. Therefore, the R-LARIAT system takes advantage of red light with greater tissue penetration and holds the potential to precisely control protein functions in living organisms.

cell biology↗

bioRxiv2025

Organoid-based modeling of platinum resistance identifies KRT17 as both a response mediator and biomarker for targeted therapy in ovarian cancer

Variable platinum responses drive high mortality in high-grade serous ovarian cancer (HGSOC), but to date, there is no molecular approach to define resistance levels for clinical decision-making. Here, we developed the organoid drug resistance assay (ODR-test) with patient-derived organoids from our ovarian cancer biobank and found that all HGSOC patients develop molecular resistance under exposure to carboplatin, although with varying clinical implications. Sustained phenotypic reprogramming and cellular plasticity under carboplatin pressure emerged as a conserved mechanism irrespective of the basal resistance level. Transcriptional and proteomic analyses revealed changes in cell adhesion and differentiation in post-platinum lines as adaptive responses that drive the increase in resistance. We identified Keratin 17 (KRT17) as a mediator of developing platinum resistance and validated its function by CRISPR/Cas9 and overexpression. Additionally, we found that KRT17 expression status (K-score) is a significant negative prognostic histopathological biomarker in a large cohort (N=384) of advanced HGSOC patients. In organoids, increased KRT17 levels enhanced sensitivity to PI3K/Akt inhibitors Alpelisib and Afuresertib, highlighting the potential of KRT17 as a stratification biomarker for targeted therapies.

cancer biology↗

bioRxiv2025

The stress response factor SigH mediates intrinsic resistance to multiple antibiotics in Mycobacterium abscessus

Mycobacterium abscessus (Mab) causes pulmonary diseases with limited treatment options due to its high level of intrinsic resistance to available drugs. Mab possesses complex and poorly understood drug resistance mechanisms. Identifying new drug targets and gaining a deeper understanding of drug resistance mechanisms are essential for discovering novel therapeutic alternatives. Here, we investigated the role of a putative sigma factor SigH in intrinsic multi-drug resistance in Mab. Mab SigH shares an 84% peptide sequence identity with Mycobacterium tuberculosis (Mtb) SigH, a well-known stress response protein and global transcriptional regulator. We constructed a sigH gene deletion strain of Mab ({Delta}sigH) and complemented strains by expressing either Mab sigH (CPMabsigH) or Mtb sigH (CPMtbsigH) in {Delta}sigH. The {Delta}sigH strain exhibited hypersensitivity to a broad range of antibiotics, including levofloxacin, moxifloxacin, tigecycline, tetracycline, amikacin, vancomycin, and rifabutin and all complemented strains restored the drug resistance phenotype. Additionally, {Delta}sigH showed increased sensitivity to oxidative and heat stress compared to the wild-type Mab and complemented strains. Transcriptomic analysis revealed that deletion of sigH disrupted the balance of gene expression, primarily elevating the expression of genes encoding YrbE and MCE family proteins and downregulating genes expressing ABC-type transporters, sigma and anti-sigma factors and other genes associated with antimicrobial resistance. Collectively, our findings indicate that SigH is a key regulator of global gene expression in response to environmental stresses, including antimicrobial treatment, and is crucial for the intrinsic drug resistance of Mab. SigH represents a promising target for the development of novel therapeutic strategies against Mab infections.

molecular biology↗

bioRxiv2025

CDK12/13 inhibition disrupts nucleolar morphology and promotes aberrant expression of IGS transcripts

The transcription-associated cyclin dependent kinases (CDKs) 12 and 13 play important roles in the expression of RNAs involved in DNA damage repair and ribosomal proteins, yet their roles if any, in the regulation of ribosomal DNA expression are unclear. Here we show that these CDKs play critical roles in the processing and degradation of noncoding RNAs in the nucleolus and that inhibition of their function leads to the upregulation of unprocessed ribosomal RNA (rRNA) transcripts in the intergenic spacer (IGS) regions between ribosomal subunits. These upregulated RNAs are aberrantly polyadenylated by the TRAMP complex polymerase TENT4, and are represented as polyadenylated rings surrounding the nucleolus. The accumulation of these noncoding RNAs is associated with increased RNA polymerase II (Pol II)-mediated transcription coupled with failure of degradation by the MTREX (SKIV2L2) RNA helicase, a member of the RNA exosome complex whose expression is prematurely terminated by CDK12/13 inhibition. Altogether, these findings provide new insights into Pol II-mediated transcriptional regulation by CDK12/13 within the nucleolus.

cancer biology↗

bioRxiv2025

Signaling of trans-saccadic prediction error by foveal neurons of the monkey superior colliculus

Across saccades, neurons in retinotopically organized visual representations experience drastically different images, but visual percepts remain stable. Here we investigated whether such stability can be mediated, in part, via prediction-error signaling by neurons processing post-saccadic visual images. We specifically recorded from foveal superior colliculus (SC) neurons when a visual image only overlapped with their response fields (RFs) after foveating saccades but not pre-saccadically. When we rapidly changed the target features intra-saccadically, the foveal neurons post-saccadic visual reafferent responses were elevated, even though the neurons did not directly sample the pre-saccadic extrafoveal target features. This effect did not occur in the absence of saccades, and it also scaled with the extent of the introduced intra-saccadic image feature discrepancies. These results suggest that foveal SC neurons may signal a trans-saccadic prediction error when the foveated image stimulating them is inconsistent with that expected from pre-saccadic extrafoveal representations, a potential perceptual stability mechanism.

neuroscience↗

bioRxiv2025

Le Petit Prince (LPP) Multi-talker: Naturalistic 7T fMRI and EEG Dataset

Prior neuroimaging datasets using naturalistic listening paradigms have predominantly focused on single-talker scenarios. While these studies have been invaluable for advancing our understanding of speech and language processing in the brain, they do not capture the complexities of real-world multi-talker environments. Here, we introduce the "Le Petit Prince (LPP) Multi-talker Dataset", a high-quality, naturalistic neuroimaging dataset featuring 40 minutes of electroencephalogram (EEG) and 7T functional magnetic resonance imaging (fMRI) recordings from 26 native Mandarin Chinese speakers as they listened to both single-talker and multi-talker speech streams. Validation analyses conducted on both EEG and fMRI data demonstrate the datasets high quality and robustness. Additionally, the dataset includes detailed transcriptions and prosodic and linguistic annotations of the speech stimuli, enabling fine-grained analyses of neural responses to specific linguistic and acoustic features. The LPP Multi-talker Dataset is well-suited for addressing a wide range of research questions in cognitive neuroscience, including selective attention, auditory stream segregation, and working memory processes in naturalistic listening contexts.

neuroscience↗

bioRxiv2025

Neural-physics adaptive reconstruction reveals 3D subcellular nanostructures over a large depth of field

Single-molecule localization microscopy (SMLM) enables volumetric nanoscopy by retrieving 3D molecular positions from engineered 2D fluorescence patterns. However, achieving nanoscale resolution over large axial ranges in complex biological samples remains challenging due to optical aberrations and overlapped single molecules. Here, we introduce LUNAR, a blind SMLM framework that can precisely localize high-density molecules even with inaccurate and highly distorted PSFs. LUNAR employs a self-supervised neural-physics learning strategy that jointly optimizes a physical PSF model and a neural network to infer key molecular parameters, including 3D positions, photon counts, and aberrations. Through simulations and experiments, we show that LUNAR achieves superior robustness to aberrations, high-density performance across diverse PSFs over multiple frames. We showcase its capabilities through whole-cell nanoscopy of mitochondria, nuclear pores, and neuronal cytoskeletons at large imaging depths. By uniting deep learning with physical modeling, LUNAR provides a calibration-free solution for aberration-robust volumetric nanoscopy and establishes a general framework for adaptive, data-driven 3D imaging.

biophysics↗