Biology subjects

Dong, S.

Publications and source records attributed to Dong, S..

At least 19 recordsLinked to original sources

bioRxiv2026

The RNA-binding protein TRIM71 Promotes Supporting Cell Reprogramming and Hair Cell Regeneration by Activating Sall4

Hair cell (HC) loss in the auditory organ (cochlea) is irreversible and represents the leading cause of deafness in humans. The absence of HC regeneration is primarily due to the limited plasticity of cochlear supporting cells (SCs). Using a novel doxycycline-inducible TRIM71 transgenic mouse model, this study demonstrates that overexpression of the RNA-binding protein TRIM71 promotes spontaneous cochlear HC regeneration in neonatal mice. Experiments in cochlear organoids indicate that TRIM71 reprograms SCs into progenitor-like cells and that this reprogramming is further enhanced by coactivation of follistatin (FST). FST augments TRIM71-induced increases in chromatin accessibility at progenitor and HC-specific gene loci, facilitating their transcriptional activation in SCs. Notably, FST also enhances TRIM71-induced reactivation of Sall4, which, like Trim71, is highly expressed in early otic progenitors. Using cochlear organoids and explants, we show that the early otic transcription factor SALL4 is required for TRIM71-induced HC formation. By contrast, SALL4 overexpression inhibits HC formation by maintaining reprogrammed SCs in an undifferentiated, proliferative state. These findings establish TRIM71 as an activator of HC regeneration in vivo, provide mechanistic insights into TRIM71's role in SC reprogramming, and identify SALL4 as its essential effector.

cell biology↗

bioRxiv2026

Human panepigenome represents epigenomic diversity

The human pangenome captures genetic diversity beyond a single linear reference, yet conventional DNA methylome maps largely assume that the underlying CpG substrate is fixed. Here we present a first draft of the human panepigenome on DNA methylation, generated from 440 haplotype-resolved long-read methylomes spanning 26 globally distributed populations. This map jointly captures CpG presence, absence and methylation across diverse human haplotypes, revealing 12.8 million CpGs absent from GRCh38 and an average of 1.2 million variant-associated CpGs (var-CpGs) per methylome. Var-CpGs captured population-associated epigenomic variation beyond that represented by shared reference CpGs, and structural variants could create or remove entire CpG islands, frequently through mobile element insertions. Promoter var-CpG methylation was associated with transcript expression, with most associations persisting after adjustment for the underlying CpG-altering variant, whereas a subset showed evidence of mediation or variant-by-methylation interactions. CpG-altering variants were also enriched among molecular QTLs in HPRC2 and across GTEx tissues, and among clinically and pharmacogenomically annotated loci. Together, these results establish a haplotype-resolved framework in which human epigenomic diversity reflects not only variation in methylation level but also genetic variation in presence or absence of CpG substrates, providing a foundation for panepigenomic studies across tissues, populations and disease contexts.

genomics↗

bioRxiv2026

Locus-specific transposable element expression drives human hematopoietic stem cell disease pathophysiology

VEXAS syndrome (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) is a severe, inflammatory syndrome resulting from mutated UBA1 leading to hematopoietic stem cells (HSC) expansion. Although UBA1-mutant HSC show complex phenotypes including proteostasis defects, sustained inflammation and clonal expansion of myeloid biased progeny, the pathogenic mechanisms at the HSC level are unknown from these gene-centric studies alone. By focussing on the non-coding genome and using advanced functional genetic methods, we found that VEXAS HSC, compared to controls, had altered expression of individual transposable elements (TE) and are key regulators of VEXAS pathogenesis. Locus-specific TE quantification identified two L1 elements, L1-10 and L1-15, active in both normal and VEXAS HSC that drive myeloid commitment by co-opting SPI1 and IRF1 transcription factors (TF) via networks common to other myeloid-biased conditions. Lipid nanoparticle (LNP)-mediated CRISPRi of L1-10 and L1-15 in UBA1-mutant HSC also caused reversion of VEXAS-associated functional phenotypes in vitro and in vivo. Functionally, pharmacologic inhibition of UBA1 with TAK-243 led to L1-10 and L1-15 RNA accumulation, while enhancement of UBA1 activity with Auranofin reversed this effect. Our study provides direct evidence that VEXAS-specific TE govern HSC clonal dominance, thereby uncovering a regulatory axis underlying HSC biology and disease mechanisms, opening a therapeutic strategy directed towards the repetitive genome.

genetics↗

bioRxiv2026

A Preparation-Free Mixture-of-Experts Framework for Protein-Ligand Affinity Prediction

Protein-ligand affinity (PLA) prediction is central to AI-driven drug discovery, but precise interaction-based methods require costly conformation preparation and data encoding, limiting their throughput. To reconcile accuracy with efficiency, we first investigate whether pre-trained molecular representation models can replace complex encoders. A unified and diverse assessment of sequence-, graph-, and image-based representations reveals both strong overall performance and family-wise variability, delivering the first practical guidance for encoder selection in PLA tasks. Next, to achieve high computational efficiency without sacrificing expressiveness, we adopt the mixture-of-experts (MoE) strategy from large language models. Systematic ablation studies uncover key design principles for deploying MoE in molecular prediction. The resulting model, HydrAffinity, is an interaction-free, dynamic sparse model that uses pre-trained encoders and MoE for parameter-efficient learning. It outperforms all interaction-free methods and matches state-of-the-art interaction-based methods on CASF-2016. Routing analysis confirms that MoE develops distinct, family-specific activation patterns, providing interpretable evidence of dynamic parameterization across protein classes. Zero-shot tests on DUDE-Z and LIT-PCBA further show strong EF5% performance, making HydrAffinity a practical, scalable solution acting as an effective early-stage pre-filter. Graphic abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=79 SRC="FIGDIR/small/740495v1_ufig1.gif" ALT="Figure 1"> View larger version (26K): org.highwire.dtl.DTLVardef@fb7a56org.highwire.dtl.DTLVardef@1cc154org.highwire.dtl.DTLVardef@1d874eeorg.highwire.dtl.DTLVardef@1e4db22_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2026

Fast and Accurate Photon-Transport Modeling based on Foundation-Model-Encoded Implicit Neural Surrogate towards Optimized Near-Infrared Brain Stimulation

Clinical treatment planning for near-infrared (NIR) neuromodulation requires patient-specific dosimetry to optimize light fluence (LF) delivery to cortical targets. The gold-standard Monte Carlo photon-transport forward solver is accurate but computationally expensive and non-differentiable for personalized inverse design across subjects. Here, we present PhomiNeuro, a foundation-model-encoded differentiable surrogate for time-resolved LF modeling and pathology-guided inverse design. A pretrained 3D medical imaging foundation model (VISTA3D) was domain-adapted to 285 training head models annotated with optical properties, and then coupled to an implicit neural representation (INR) that predicts LF at arbitrary spatiotemporal coordinates. Regularized by a time-dependent diffusion-equation residual, it demonstrated superior fidelity in 84 held-out participants, and ablations identified VISTA3D-derived anatomical priors as dominant and physics regularization as complementary. PhomiNeuro enables fast and robust queries of LF and its gradients with respect to illumination parameters for targets guided by individual structural magnetic resonance imaging and amyloid positron emission tomography, achieving 466-fold per-iteration speedup and 167-fold end-to-end optimization speedup. Explanatory analyses of LF delivery variability, including sex differences and amyloid status, identified structural variations as the primary drivers. These results position PhomiNeuro as a highly extensible translational framework toward personalized treatment planning and digital twin development in precision neuromodulation.

bioengineering↗

bioRxiv2026

Photopatterned spatiotemporal organisation and in situ differentiation of 3D human cortical networks

Modelling human cortical microcircuitry in vitro requires platforms that recapitulate both the compositional complexity and spatial architecture of developing neural tissue. Current organoid and assembloid models often rely on the bulk fusion of pre-differentiated, region-specific cells, lacking the capacity for emergent spatial co-differentiation and microenvironment-driven multiscale organisation. There is also a lack of neural and neuronal-glial models with photo-architectured network geometries. To address these limitations, we present a volumetric in situ differentiation system using a triculture of precision reprogrammed human iPSC-derived glutamatergic neurons, GABAergic neurons and astrocytes embedded throughout ultra-soft photocrosslinkable hydrogel microenvironments. The deterministic and spatially controlled method allows us to engineer macro-scale, interconnected human neural networks directly onto functional microelectrode array interfaces using projection photopatterning for high-throughput screening. Unlike fusion-based organoids and assembloids, our platform enables simultaneous, spatially distributed lineage differentiation and maturation, and extensive topography-guided neurite outgrowth bridging localised cellular hubs to recapitulate various aspects of neurodevelopmental patterning and synaptic integration in 3D. The model enables topographic patterning of neuronal-glial networks as well as 3D cell-embedded bioprinting with the developed triculture system. Both modes of cellular growth are studied and demonstrated here. Longitudinal electrophysiological tracking over a month of culture reveals a transition from immature, quiescent states to asynchronous, information-dense microcircuits characterised by an expanded state-space manifold and physiological excitatory-inhibitory balance. By replicating the mechanics of native brain parenchyma, the model presents a highly reproducible, scalable and flexible platform for the study of cortical microcircuitry development, neurodegenerative decline, and inter-regional network assembly.

bioengineering↗

bioRxiv2026

Spatiotemporal unfolding of prefrontal neural response divergence in adolescent depression during naturalistic experience

Depression alters how emotional experience is interpreted and regulated, yet how depression-related neural divergence unfolds during ongoing experience remains poorly understood, leaving unresolved whether such divergence reflects stable trait-like abnormalities or state-dependent responses that emerge only at particular moments. This question is especially consequential in adolescence, when depression often emerges amid heightened social sensitivity and continuing maturation of prefrontal regulatory systems. Here, we used naturalistic film watching fMRI data of 36 adolescents with depression and 36 matched controls to map where, how, and when neural responses diverge between diagnostic groups. Neural polarization analysis localized group-level divergence to two prefrontal hubs, the dmPFC-rACC and left dlPFC, where individual similarity to the depression-group response pattern was associated with depressive symptom severity across the cohort. These local divergence sites were not spatially isolated. Seed-based inter-subject functional connectivity showed that they were embedded within distributed stimulus-locked coupling patterns spanning prefrontal control, salience, and subcortical reward circuits. Time-resolved phase synchrony further revealed that this divergence was not continuous, but arose at specific narrative moments, with the dmPFC-rACC diverging during socio-emotional events and the left dlPFC during cognitively demanding ones. Together, these findings move beyond static localization to reveal depression-related neural divergence as a spatially organized, network-embedded, and temporally gated response pattern shaped by the unfolding content of naturalistic experience.

neuroscience↗

bioRxiv2026

Inbreeding depression by polygenic load following a severe population bottleneck

It is poorly understood how populations survive extreme bottlenecks despite severe inbreeding. We investigate the genomic architecture of inbreeding in the once critically endangered Seychelles warbler (Acrocephalus sechellensis) using 37 years of individual-based monitoring and 1.8 million SNPs. Linkage disequilibrium patterns reveal a historical effective population size of [~]270 that plummeted to [~]13 coinciding with human colonisation events of the Seychelles archipelago. Contemporary genomes are over one-third inbred from recent inbreeding measured by runs of homozygosity (ROH) formed since human colonisation of the Seychelles, 50 generations ago (mean FROH < 50 generations ago = 0.38), and we identify significant inbreeding depression across cross key fitness traits. On average, a 17.9% reduction in lifespan and a 15.1% reduction in lifetime offspring production were associated with each 10% increase in individual inbreeding. Genome-wide scans reveal this depression is caused by a polygenic load. Our results suggest that a history of small population size could have facilitated the selective purging of severe genetic load, while mildly deleterious alleles escaped selection via drift. A partitioned genetic load architecture potentially enabled species recovery despite the inbreeding depression that followed near-extinction.

evolutionary biology↗

bioRxiv2026

Integrative Transcriptomic and Functional Analysis Reveals Fatty Acyl Elongases Involved in Sex Pheromone Biosynthesis in Rice Leaffolder, Cnaphalocrocis medinalis (Lepidoptera: Pyraloidea)

Elongases are essential enzymes in the biosynthesis of sex pheromones in many lepidopteran species. Together with desaturases, they determine the carbon skeletons of many pheromone precursors, thereby contributing to the production of species-specific chemical signals. However, to date, such fatty acyl elongase gene has not been functionally characterized. The rice leaffolder, Cnaphalocrocis medinalis, utilizes a blend of C18 monounsaturated aldehydes and alcohols as its sex pheromone, implying a critical elongation step from C16 precursors. In this study, we performed pheromone gland transcriptome analysis and identified 45 candidate biosynthetic genes. Functional assays in Nicotiana benthamiana showed that the {Delta}11 desaturase Cmed070400 produces (Z)-11-hexadecenoic acid, which serves as the substrate for elongation. Multiple elongases catalyzed its conversion to (Z)-13-octadecenoic acid, with Cmed092440 showing the highest activity. These findings provide the first experimental evidence for elongase-mediated formation of C18 pheromone precursors in C. medinalis. The identification of a minimal set of functionally active enzymes further enables reconstruction of this pathway in plant systems, offering a basis for sustainable production of pheromone precursors for pest management applications.

zoology↗

bioRxiv2026

A Seychelles warbler genomic toolkit

Understanding evolutionary processes is greatly facilitated by high-quality data on genetic variation. We report the development of a genomic toolkit for a recently bottlenecked, long-term studied species, the Seychelles warbler (Ptimerl dezil; Acrocephalus sechellensis). This toolkit comprises a reference genome assembled into 31 chromosomes, together with functional annotations and reference-panel-free imputation of whole-genome sequences from 1,935 individuals. The genomic data have been used to assign the sequenced individuals into a genetic pedigree. Individual genomic data are associated with a suite of phenotypic metadata, amassed from three decades of fieldwork in this closed, long-term monitored population. We compared sex and parentage assigned using the genomic data with the previously recorded sex and parentage metadata to identify and correct 41 sample DNA samples labelled with the wrong identity. This population resource enables a wide range of analyses, that include, but are not limited to phylogenetics, metabarcoding, recombination rates, linkage patterns, adaptation, heritability, demographic history, selection, and inbreeding estimates. We wish to encourage interest from researchers seeking to collaborate on future analyses and data collection. Overall, our methods demonstrate the potential of next generation sequencing and statistical tools to provide dense genomic datasets at large sample sizes for wild populations.

genomics↗

bioRxiv2026

Transition metal-triggered immunity via an Arabidopsis NLR pair

Plants are challenged by multiple biotic (e.g. pathogens) and abiotic (e.g. heavy metals) stressors. Some transition metals can enhance plants defense against pathogens, but the mechanism remains unclear. Here, we demonstrate that an Arabidopsis head-to-head gene pair of intracellular nucleotide-binding leucine-rich repeat (NLR) receptors, both expressed in the endodermis of roots, antagonistically control transition metal-triggered immunity. One NLR, STM2 binds transition metal ions, such as Cd2+, Cu2+ and Zn2+, with its LRR domain to enhance its NAD+ hydrolytic activity and immune responses via the EDS1/PAD4/ADR1 module, triggering enhanced resistance to bacterial wilt Ralstonia solanacearum. The other NLR, STM1 suppresses STM2 to protect plants from transition metal-triggered immunity and growth inhibition in the presence of excess metals. STM1 also dampens resistance to the pathogen. Our study defines an NLR activated by transition metals and reveals a trade-off between susceptibility to pathogens and sensitivity to transition metals that are pervasive in soil.

plant biology↗

bioRxiv2026

Density-dependent facilitation of livestock by small mammal ecosystem engineers

Small mammals and large herbivores have co-evolved in grasslands for millions of years, yet how they interplay remains unclear. Although large herbivores can significantly affect the smaller ones, the reversal effect is largely unknown. On the Qinghai-Tibetan Plateau, plateau pikas (Ochotona curzoniae) are often considered pests that compete with livestock at high densities. Using field experiments, we show that the presence of pikas facilitates weight gains of yaks (Bos grunniens) by improving vegetation composition at a moderate density level. Compared to the pika-present treatment, pika removal dramatically increased cover of the poisonous Stellera forbs by two-fold, reducing the abundance and protein content of palatable grasses and sedges, yak foraging efficiency, and yak weight gain by up to 42%. Notably, we found a humped relationship between yak weight gains and pika burrow densities in the pika-present plots; the facilitative effect of pikas on yaks was highest at about 200 burrows/ha, but shifted to a competitive effect at pika densities exceeding 400 burrows/ha. These results provide the first empirical evidence that maintaining a moderate density of small mammalian herbivores can benefit growth performance of livestock by improving vegetation composition. Our study highlights the significance of moderate populations of ecosystem-engineering small mammals in sustaining pastoral productivity in rangelands.

ecology↗

bioRxiv2026

Molecular dynamics of Brodmann Area 22 in development and autism

Challenges in verbal communication are a prominent feature of autism. However, gene regulatory programs in speech-related cortical regions remain poorly characterized. In parallel, it remains unclear whether the heterogeneous genetic factors underlying autism converge on shared neurobiological mechanisms. To address these gaps, we generated paired transcriptomic and epigenomic data from post-mortem human brain tissue across 100 donors. Here, we show that transcriptional differences in the speech-related Brodmann Area 22 in individuals with neurodevelopmental conditions, including autism, are strongest among those with a known genetic diagnosis. A similar but attenuated signature is observed in those without a genetic diagnosis. These transcriptional differences are most pronounced in neurons, with glutamatergic L4/5 intratelencephalic neurons affected across multiple modalities. Finally, multimodal analysis implicates altered RFX3-dependent networks as a central hub in autism, particularly among L4/5 intratelencephalic neurons in non-verbal individuals. Together, our study identifies regulatory architecture linking chromatin state, transcriptional output, and variation in verbal ability in autism.

neuroscience↗

bioRxiv2026

Fully T2T pedigree assemblies reveal genetic stability and epigenetic plasticity of human centromeres across inheritance and cell-fate transitions

Centromeres are essential chromosome components yet remain poorly understood due to their highly repetitive sequence architecture. Using fully-phased telomere-to-telomere diploid assemblies from a three-generation pedigree integrated with long-read epigenomes from matched peripheral blood mononuclear cells, induced pluripotent stem cells, and neural progenitor cells, we generate allele-resolved single basepair resolution maps of centromere genetic and epigenetic dynamics across inheritance, reprogramming, and differentiation. We show that centromeric dip regions (CDRs), which define the functional core of centromeres, are positionally stable across generations and cell-fate transitions. In contrast, CDR epigenetic architecture is highly dynamic. Reprogramming markedly attenuates CDR hypomethylation, which is partially restored during differentiation in parallel with global hypomethylation of active alpha-satellite arrays and coordinated changes in nucleosome organization and protein occupancy. Centromeric remodeling is insulated from X-chromosome status, including Xa, Xi, and erosion. Finally, de novo mutations arising during reprogramming are enriched in centromeric regions but depleted within functional centromeric cores.

genomics↗

bioRxiv2026

A scalable approach to resolving variants of uncertain significance

Over 90% of missense variants across [~]4,000 disease-associated genes are variants of uncertain significance (VUS). Experimental variant effect measurements provide critical evidence about pathogenicity and inform disease biology, but most variants lack data and clinical translation has been limited. The Impact of Genomic Variation on Function Consortium generated experimental data for 62,215 variants across ten genes using multiplexed assays and 1,407 variants across 163 genes using arrayed assays, curated 193,139 additional community-generated variant effect measurements across 30 additional genes, and developed automated calibration methods for translating experimental data and variant effect predictions into clinical evidence. To reduce current VUS, we developed a scalable workflow using only experimental and predictive evidence, enabling reclassification of 75% of the 16,115 VUS in these genes as pathogenic or benign with <1% error. To minimize future VUS, we analyzed >90,000 unobserved variants; 62% had enough evidence to be "preclassified" as pathogenic or benign. We validated our data, evidence and classifications using All of Us and created interactive resources to enable clinical use of the calibrated data. Thus, for 40 genes, representing 1% of the clinical genome, we resolve most existing VUS and future variants, illustrating how systematic use of scalable evidence can empower genomic medicine.

genomics↗

bioRxiv2026

An AI-Native Biofoundry for Autonomous Enzyme Engineering: Integrating Active Learning with Automated Experimentation

The engineering of enzymes with novel functions is a cornerstone of synthetic biology but remains bottlenecked by the fragmentation between computational design and physical execution. While "self-driving" laboratories promise to resolve this, existing systems often rely on rigid, device-specific scripts that lack the flexibility to handle complex, evolving scientific tasks. Here, we report an AI-native autonomous biofoundry that fundamentally redefines laboratory automation through a "cloud-edge synergistic" architecture. The platform features an Agent-Native control system powered by Large Language Models (LLMs) and the Model Context Protocol (MCP), which bridges the semantic gap between abstract scientific intent and heterogeneous hardware execution. This architecture enables non-experts to orchestrate the entire Design-Build-Test-Learn (DBTL) cycle via natural language. By integrating deep phylogenetic mining, zero-shot protein language models (ESM-2), and supervised active learning, our system efficiently navigates rugged fitness landscapes. As a rigorous proof of concept, we applied this platform to evolve a Family B DNA polymerase for CoolMPS sequencing, a task requiring the incorporation of non-natural 3-blocked nucleotides. In just three autonomous rounds, the platform achieved a hit rate of >66% and identified variants with a 37% reduction in sequencing error rate compared to a commercial reference. This work demonstrates that AI-native infrastructures can not only accelerate trait evolution by orders of magnitude but also provide a scalable, brand-agnostic paradigm for the future of automated scientific discovery.

bioengineering↗

bioRxiv2026

Chromatin context-dependent deacetylation by the asymmetric Rpd3L

The regulation of gene expression requires precise control of chromatin-associated complexes that respond to diverse structural and epigenetic cues. The Rpd3 Large (Rpd3L) complex is a Sin3 histone deacetylase (HDAC) complex that dynamically adapt to chromatin states to reinforce transcriptional silencing, yet the mechanisms governing the catalytic activation in chromatin context-dependent manner remain unclear. Here we present the cryo-EM structure of Rpd3L bound to both mono- and di-nucleosome substrate at near-atomic resolution, uncovering a substrate-guided allosteric activation mechanism. Rpd3L adopts an asymmetric architecture, in which the proximal catalytic module anchors the first nucleosome, while the Sin3 PAH domains engage linker DNA to reposition a second nucleosome. This spatial configuration brings the distal catalytic module into proximity with chromatin and unlocks its latent deacetylase activity. Biochemical and mass spectrometry analyses confirm that dual nucleosome engagement selectively enhances Rpd3L activity and broadens substrate specificity. Together, these findings establish a hierarchical mechanism by which Rpd3L interprets histone modifications and nucleosome organization to modulate its enzymatic output at promoter regions. Our study provides a framework for understanding higher-order chromatin repression mechanisms by chromatin-regulation complexes and co-repressors.

molecular biology↗

bioRxiv2026

Leveraging Human Pangenome for Improved Somatic Variant Detection

Somatic variant detection is technically challenging due to low variant allele fractions, the confounding presence of germline variation, and reference bias. Linear references such as GRCh38 miss sample-specific variation, causing misalignments and incorrect variant calls. Although telomere-to-telomere donor-specific assemblies (DSAs) accurately represent individual genomes, their application is limited by cost and technical barriers. Alternatively, the graph-based human pangenome provides a scalable framework to improve read alignment and perform genome inference. Here, we benchmarked somatic variant detection using GRCh38, graph-based pangenomes, and pangenome-inferred DSAs with a HapMap mixture dataset and the COLO829 melanoma cell line. Pangenome-guided alignment improves read mapping and somatic variant calling accuracy. Furthermore, personalized pangenomes partially reconstruct donor-specific genomic content, improving accuracy, reducing germline contamination, and enabling detection of events in loci absent or poorly represented in GRCh38. These findings demonstrate that graph-based and personalized pangenomes are effective strategies for enhancing somatic variant detection compared with GRCh38.

genomics↗