Biology subjects

Nguyen, H.

Publications and source records attributed to Nguyen, H..

At least 19 recordsLinked to original sources

bioRxiv2026

Design of readers for pathological PTM proteoforms in Alzheimer's disease

Alzheimer's disease (AD) is characterized by heterogeneous Tau phosphorylation and aberrant amyloid-beta (Abeta) processing, yet molecular tools that distinguish closely related molecular states of tau and Abeta with predefined specificity remain difficult to develop. Here, we designed de novo protein readers targeting three disease-associated phospho-Tau regions and N-terminal pyroglutamate-modified Abeta (N3pG-Abeta). Screening 192 designs per target identified 51 modification-selective hits among 768 candidates (6.6%). Representative readers exhibited nanomolar to subnanomolar affinities for modified peptides, discriminated authentic modifications from unmodified or substitution-based controls, and preferentially recognized modified full-length targets. We further programmed local modification specificity by interface redesign, generating pT231 readers conditional on the phosphorylation state of neighboring S235 and S237. A 2.0-angstrom crystal structure of an N3pG-Abeta reader closely matched the computational design model (interface C-alpha RMSD = 0.52 angstrom), and redesign of pT231 readers tuned apparent binding affinities from greater than 1 uM to below 1 nM. In a human 3D neuronal culture with both pathological Abeta and tau, selective readers targeting N3pG-Abeta, Tau pT231, and Tau pT217 recognized disease-associated signals, exhibiting colocalization with related antibodies and much higher signal in the AD line than the control. Tau pT231 readers further quantitatively distinguish AD patients and control human brain lysates in dot-blot assays. Together, these results establish genetically encoded readers with programmable specificity for defined Tau and Abeta modification states, enabling molecular-state-resolved detection in human disease models and biological samples, and lay a foundation for controlled pathological molecular perturbation.

bioengineering↗

bioRxiv2026

Programming N-linked glycan composition through de novo protein design

N-linked glycans are ubiquitous post-translational modifications that mediate diverse biological processes, including cell signaling and adhesion, molecular recognition, and host-pathogen interactions. Despite their importance, N-linked glycans have largely remained beyond the reach of biomolecular design due to a lack of methods capable of jointly modeling glycans and protein structure. Here, we develop a general all-atom computational pipeline for creating de novo glycoproteins in which the amino acid sequence encodes both the location and composition of N-linked glycans. The method uses three-dimensional protein structure to tune glycan accessibility to processing enzymes in the eukaryotic secretory pathway. Systematic increases in steric restriction progressively reduced processing, shifting glycans from heterogeneous complex-type structures toward increasingly underprocessed oligomannose-type species resembling those found on many viral glycoproteins. Despite this programmed restriction, the resulting oligomannose glycans remained accessible for biological recognition, inhibiting uropathogenic Escherichia coli adhesion to human bladder cells and binding mannose-binding lectin. These results establish steric control of N-glycan processing as a genetically encodable design principle and extend computational protein design to functional glycoproteins with programmed composition.

biochemistry↗

bioRxiv2026

Complex enzyme active site scaffolding by iteratively detuned catalytic guidance

De novo enzyme design requires sequences whose structures reproduce prescribed catalytic geometries. This remains challenging for active sites that couple multiple catalytic residues, cofactors and substrates. Here we introduce the Catalytic Guidance Engine (CaGE), which optimizes designed sequences to satisfy complex sets of functional constraints by alternating constraint guided structure prediction trajectories with sequence design, gradually fading the strength of the constraint information over cycles to nudge the designed sequence towards satisfying all the constraints. We train RF3 to accept atom-pair distances and other design conditions, and use classifier-free guidance to fade their influence over multiple design cycles. We also develop training-free motif guidance protocols which operate at inference time by directly adjusting selected atoms towards the design target geometry during diffusion sampling. CaGE improves catalytic-motif geometry across a diverse enzyme benchmark and yields sequences that recover complex target active sites in independent AF3 predictions. We experimentally demonstrate the power of CaGE by designing active enzymes with complex catalytic sites: a nonheme-iron design positions four active-site residues around an iron centre and a transition-state ligand, and a PLP-dependent design positions four catalytic residues around a PLP cofactor with multiple functional groups. Active site substitutions and a 2.7[A] crystal structure support the designed mechanisms and structures. Thus, CaGE provides a route to functional de novo enzymes with complex active sites.

bioengineering↗

bioRxiv2026

Programmable De Novo Design of Mesoporous Protein Crystal Frameworks

Three-dimensional protein crystals are ordered, porous macroscopic materials with potential applications in catalysis, biosensing, and biomedicine. However, most protein crystals are obtained by empirical screening, providing limited control over the lattice architecture, pore geometry or component composition that determine material function. Here, we present a modular strategy for the programmable design of highly porous, framework-like protein crystals using predefined protein-protein interactions. This strategy yielded over 30 distinct protein crystals, including single-component and multicomponent P213 and I213 lattices that grow to over 100 micrometers in size. Small-angle X-ray scattering and electron microscopy showed close agreement between experimental lattices and computational models. RFdiffusion-guided design generated isomorphous variants with matched lattice parameters, enabling coherent protein crystal alloys, epitaxial core-shell growth and reversible shell assembly. The designed crystals exhibit tunable mesoporous architectures, with limiting apertures of 2-18 nm, and support genetically encoded incorporation of fluorescent protein guests. These results establish a general route to programmable lattice engineering of protein crystals and position them as genetically encoded, compositionally tunable mesoporous materials.

synthetic biology↗

bioRxiv2026

BET BD2 inhibition facilitates SPOP-mediated degradation of chromatin-associated BRD4/BRD4-NUT, a therapeutic vulnerability in NUT carcinoma

BET bromodomain inhibitors block binding of BET family bromodomains 1 and 2 (BD1, BD2) to chromatin and have demonstrated clinical activity in NUT carcinoma (NC), a BRD-NUT fusion-driven cancer, but toxicity from BD1 inhibition has limited their effectiveness. We investigated whether selective inhibition of BRD4 bromodomain 2 (BD2) could retain antitumor activity while reducing toxicity. NC cells were uniquely sensitive to the novel BRD4-BD2 inhibitor DC-9476 and other BD2-selective inhibitors, which induced differentiation and growth arrest. A CRISPR knockout screen identified the BRD4-targeting E3 ligase SPOP as the top resistance hit. BD2 inhibition, but not BD1-selective or pan-BET inhibition, triggered SPOP-dependent proteasomal degradation of BRD4 and BRD4-NUT; SPOP loss prevented degradation and largely rescued BD2 inhibitor-induced differentiation and growth arrest. Unexpectedly, BRD4 and BRD4-NUT remained chromatin-associated during BD2 inhibition, whereas BD1 or pan-BET inhibition displaced them. Together with evidence that ectopic BRD4-NUT expression sensitizes BRD4 to degradation, these findings support a model in which BRD4-NUT megadomains create a high-density, degradation-competent SPOP substrate pool of BRD4 and BRD4-NUT upon BD2 inhibition, whereas pan-BET inhibition disperses this substrate and limits efficient degradation. In preclinical NC models, BD2-selective inhibition achieved greater tumor growth inhibition and survival benefit than pan-BET inhibition, revealing a therapeutic vulnerability.

cancer biology↗

bioRxiv2026

Ebola virus exploits host lncRNA LINC01740 to enhance ATF3 and suppress antiviral immune responses

Ebola virus (EBOV) infection causes severe hemorrhagic fever marked by dysregulated cytokine production, impaired antiviral defenses, and multi-organ failure. Macrophages are primary targets of EBOV, and viral replication profoundly alters macrophage transcriptional programs, driving hyperinflammation. Although long non-coding RNAs (lncRNAs) are increasingly recognized as regulators of immunity and viral pathogenesis, their roles in EBOV infection remain poorly understood. We performed comprehensive transcriptomic profiling of primary human monocyte-derived macrophages infected with the highly pathogenic EBOV Mayinga variant. Infection triggered extensive remodeling of both coding and non-coding transcriptomes, including hundreds of differentially expressed lncRNAs. Functional analysis of neighboring protein-coding genes of EBOV-induced lncRNAs (EVILs) revealed enrichment of pathways linked to cytokine signaling, transcriptional regulation, and cell signaling, all of which are central to Ebola virus disease (EVD) pathogenesis. Among the most strongly induced EVILs, LINC01740 and its neighboring protein-coding gene, Activating Transcription Factor 3 (ATF3), were significantly upregulated. Antisense oligonucleotide-mediated inhibition of LINC01740 reduced ATF3 mRNA and protein levels. CRISPR/Cas13d-mediated knockdown of ATF3 restored type I interferon (IFN-I) signaling and antiviral gene expression in EBOV-infected macrophages. Mechanistically, ATF3 functions as a negative regulator of IFN-I and type I interferon-stimulated gene expression, thereby suppressing antiviral immune responses in EBOV-infected macrophages. Together, these findings identify a previously unrecognized LINC01740-ATF3-IFN-I regulatory axis that EBOV exploits to promote immune suppression and viral replication.

immunology↗

bioRxiv2026

EDTA v2: enabling scalable TE annotation in animal genomes

The Extensive de-novo TE Annotator (EDTA) automates transposable element annotation in plant genomes but lacks direct LINE/SINE detection, limiting its applicability to animal genomes. We present EDTA v2, which integrates LINE and SINE detection, completely rewrites TIR-Learner for deployability and scalability, and accelerates structural detectors by up to two orders of magnitude. Tested in 30 animal genomes from the Vertebrate Genomes Project Phase I, EDTA v2 bridges the non-LTR detection gap that has prevented automated TE annotation in animals.

genomics↗

bioRxiv2026

Late-Stage Large Extracellular Vesicles Reprogram CHO Cell Metabolism in a Glutamine-Dependent Mode and Promote Antibody-Productivity to Cell-Growth Tradeoff

Extracellular vesicles (EVs) are mediators of intercellular communication, yet their impact on Chinese Hamster Ovary (CHO) cell physiology and bioprocess performance remains poorly understood. Here, we investigated whether small EVs (sEVs) and large EVs (LgEVs) that accumulate during fed-batch and perfusion cultures modulate CHO cell growth, metabolism, apoptosis, and monoclonal antibody (mAb) production. EVs isolated from early- and late-stage cultures were added to fresh CHO cultures grown with or without glutamine supplementation. Only LgEVs had a significant impact. Late-stage LgEVs markedly altered CHO-cell behavior, reducing cell proliferation, increasing apoptosis under glutamine-limited conditions, and substantially enhancing mAb productivity in a dose-dependent manner. Glutamine supplementation largely alleviated the growth-inhibitory and pro-apoptotic effects of LgEVs while preserving their positive impact on productivity, suggesting that glutamine decouples EV-mediated stress from productivity enhancement. Metabolic analyses revealed increased glucose consumption, a glutamine-dependent shift between glycine and alanine overflow metabolism, and remodeling of amino-acid utilization. Metabolic flux analysis further demonstrated enhanced glycolytic overflow and increased reliance on amino acid-supported anaplerosis. Conversely, selective removal of LgEVs from perfusion medium significantly improved cell expansion without reducing antibody production, supporting an inhibitory role for late-stage LgEVs. These LgEVs were enriched in let-7 family miRNAs and miR-21, consistent with RNAseq analyses demonstrating stress-associated enrichment of these miRNAs in CHO EVs and with functional studies showing that let-7a and miR-21reduce CHO-cell growth. Together, these observations suggest that selective miRNA loading contributes to the growth, metabolic, and productivity phenotypes elicited by late-stage LgEVs. Our findings identify LgEVs as endogenous regulators of CHO-cell physiology and potential targets for optimizing high-density fed-batch and perfusion biomanufacturing processes. HighlightsO_LIEndogenous late-stage Large Extracellular Vesicles (LgEVs) reduce CHO cell growth but boost specific mAb productivity. C_LIO_LIGlutamine supplementation rescues LgEV-mediated growth inhibition and apoptosis. C_LIO_LIMetabolic Flux Analysis (MFA) based on the dynamic behavior of amino acid and other metabolite and substrate concentrations reveals the pyruvate node as a metabolic bottleneck and the associated lactate overflow metabolism as resulting from LgEV exposure. C_LIO_LIStress-associated let-7 and miR-21 microRNAs are highly enriched on a per-EV basis in late-stage LgEVs. C_LIO_LISelective removal of LgEVs improves perfusion cell growth without impacting antibody titer. C_LI

Cell Biology↗

bioRxiv2026

Orally available designed miniproteins inhibit enterotoxigenic Bacteroides fragilis pathology by blocking toxin receptor binding

Toxigenic bacterial infections in the gut are a significant contributor to the global burden of disease. Advanced tools for protein and live biotherapeutic engineering offer potentially transformative strategies for treating such diseases, while avoiding the collateral effects of traditional antibacterials. Here we used de novo protein design to identify inhibitors of the metzincin family protease Bacteroides fragilis toxin (BFT). These inhibitors, which bind distal to the active site, interfere with toxin-mediated E-cadherin cleavage and downstream proinflammatory signaling by blocking claudin-4 receptor binding. We tested the inhibitors as disulfide-stabilized variants administered directly to the cecum or in drinking water, as well as through in situ secretion by an engineered live biotherapeutic. Across these delivery modalities, the inhibitors successfully neutralized the toxin and effectively prevented BFT-associated gut pathology, including tumor formation. These results highlight the potential of de novo designed proteins as precise, non-antibiotic interventions to mitigate bacterial toxin-driven disease in the gut.

microbiology↗

bioRxiv2026

Novel bile salt analogs reduce lipid accumulation in liver cells with potential to treat both metabolic dysfunction-associated steatotic liver disease and Clostridioides difficile infection

Metabolic dysfunction-associated steatotic liver disease (MASLD) and Clostridioides difficile (C. difficile) infection (CDI) are clinically associated, yet there is limited effective treatment for both diseases. Bile salt analogs (BSAs) have demonstrated potential in treating either MASLD or CDI. We screened a library of BSAs (n=112) previously synthesized as potential inhibitors of C. difficile spore germination, for their therapeutic potential in reducing intracellular accumulation of fatty acids in HepG2 cells as candidates for prevention and treatment of both MASLD and CDI. The screening was based on an in vitro model established by incubating HepG2 cells with free fatty acids, with obeticholic acid (OCA), a known BSA with anti-MASLD activity as a control. Gene and protein expressions were quantified to validate the treatment effect. We found that compounds C13, C24, C25, C74, C98, and C101 demonstrated significant effectiveness in both preventing the intracellular accumulation of lipids and removing pre-loaded cellular lipids. Gene expression analysis showed that C24, C25, and C74 produced a similar pattern characterized by a robust induction of FGF21 expression, while C13, C98, and C101 produced a transcription pattern that mirrors the effect of OCA. Structurally, while C13, C24, and C25 do not display drug-like properties, C74, C98, and C101 are drug-like and share a similar structure. Interestingly, C101 is a potent inhibitor of C. difficile spore germination. OCA shows a weak anti-gemination effect. Our study identified lead compound candidates for the development of novel therapeutics capable of treating both MASLD and CDI. Significance statementThe clinical association between MASLD and CDI remains an unmet need for dual acting therapeutic strategies. Given the reported potential of BSA, we screened 112 previously synthesized as potential inhibitors of C. difficile spore germination, for their therapeutic potential in reducing intracellular accumulation of fatty acids in HepG2 cells. Our study identified compounds that effectively reduce intracellular lipid accumulation and inhibit C. difficile spore germination. These results nominate lead candidates for developing dual-acting therapeutics targeting both MASLD and CDI.

pharmacology and toxicology↗

bioRxiv2026

SHP2 binds directly to SOS1 to enable RAS activation

The protein-tyrosine phosphatase SHP2 (PTPN11) regulates growth factor- and cytokine-induced RAS/ERK MAP kinase (MAPK) pathway activation, and aberrant SHP2 function causes developmental disorders and cancer1-5. It is widely believed that the catalytic activity of SHP2 is essential for pathway activation1,4,6-8. This view has shaped our interpretation of how germline PTPN11 mutations cause Noonan Syndrome (NS) and NS with Multiple Lentigines (NS-ML)2,9 and how somatic mutations contribute to myeloproliferative neoplasms and solid tumors1. Here we identify a previously undetected, protein-tyrosine phosphatase (PTP) activity-independent mechanism that revises our understanding of how SHP2 promotes RAS/ERK activation. We find that certain mutations of the nucleophilic cysteine that abolish catalytic activity still promote RAS/ERK pathway activation in normal and neoplastic mammalian cells, zebrafish embryos, and mice. Structural studies show that the SHP2 PTP domain binds directly to the Son of Sevenless 1 (SOS1) Dbl homology (DH) domain. Proximity labeling and super-resolution microscopy demonstrate that SHP2/SOS1 interaction occurs in cells and facilitates SOS1 translocation to the plasma membrane to form clusters. Our results overturn decades of dogma on SHP2 regulation of the RAS/ERK pathway and provide new insights into the mechanism of action of disease-associated PTPN11 mutations.

Cancer Biology↗

bioRxiv2026

Controlling metal-carbonate phase, form, and function through de novo protein design

Biomineralization enables living systems to construct hybrid materials by controlling the location, orientation, and polymorph of inorganic crystals with proteins and other biomolecules. Despite decades of study, the molecular principles underlying these processes remain difficult to harness in engineered materials, in part because native biomineralization proteins are often intrinsically disordered, heterogeneous, or insoluble. Here we show that de novo designed protein interfaces can be assembled into reconfigurable two-dimensional arrays which template calcite nanocrystals. By fine-tuning RFdiffusion2 on repeat protein scaffolds, we further enable the design of protein architectures which selectively form aragonite, a metastable polymorph of calcium carbonate, in nucleation conditions that otherwise result in a mixture of phases. Extending beyond inorganics found in biological systems, we show that lattice-matched protein designs template cobalt carbonate formation: a flat helical repeat protein interface promotes unconfined growth, whereas soluble D3 cage assemblies yield more homogenous cobalt carbonate nanocrystals confined to the interior of the cage. These protein-cage cobalt carbonate hybrid materials function as electrocatalysts for alkaline water splitting. Our results demonstrate the potential of deep learning-based methods to unlock the structural and functional activity of protein-mineral composites.

biochemistry↗

bioRxiv2026

Electrical stimulation precisely reproduces naturalistic spiking activity in complete and intermixed neural populations in the primate retina

Neurostimulation technologies have the potential to restore complex brain functions by evoking naturalistic patterns of neural activity. However, faithfully mimicking the brains natural neural code in diverse neuron types remains an unsolved challenge. Here we exploit focal electrical stimulation and recording to reproduce natural visual signals at cellular resolution in two intermixed and oppositely-tuned neural populations in the isolated macaque retina. We calibrated and delivered electrical stimulation sequences to precisely mimic recorded visually-evoked responses to natural images in complete populations of ~50 ON and OFF parasol cells uniformly sampling ~4{degrees} x 4{degrees} of visual angle. Distinct and opposing visually-evoked response patterns in the two intermixed populations were closely replicated spike-by-spike in 5 retinal preparations using electrical stimulation, reproducing spike trains with 51 {+/-} 17% lower variability than repeated presentations of the same visual stimulus. Focal electrical stimulation also replicated traveling waves of neural activity in response to visual motion, and highly structured artificial response patterns. These findings demonstrate for the first time that neurostimulation can match or exceed biological signaling fidelity in large, complete and intermixed neural populations.

bioengineering↗

bioRxiv2026

Leaves in Transition: Single nuclei RNA sequencing provides insights into sorghums juvenile and adult phases

The transition from juvenile to adult phase (JA) is a key developmental process in plants, driven by conserved pathways that shape growth and stress responses. In monocots such as sorghum, this transition influences traits linked to environmental resilience, yet the regulation of micro-anatomical features such as trichomes and bulliform cells remains poorly understood. Here, we use single-nucleus RNA sequencing (snRNA-seq) to generate high-resolution gene expression maps of juvenile and adult sorghum leaves. Contrary to the traditional developmental model, we find that trichomes are present in both juvenile and adult leaves, with stage-dependent differences in gene expression. We also observe that bulliform cells, typically considered adult-specific, are present in juvenile leaves. At the transcriptomic level, cell populations cluster more strongly by developmental stage than by cell type, with juvenile cells grouping with other juvenile cells and adult cells with adult counterparts, indicating a dominant effect of developmental state on gene expression. In addition, we identify enrichment of dhurrin biosynthetic genes in trichome-associated cells as well as elevated expression of aquaporin genes in these cells, pointing to a potential role for trichomes in coordinating defense and water-related processes. Together, these findings refine current models of sorghum development and suggest that key epidermal cells and their associated functions may be established earlier than previously known. Collectively, this study provides a single nucleus resolved framework for understanding how developmental stage shapes gene expression in sorghum leaves.

plant biology↗

bioRxiv2026

Structure and Dynamics of a Long-Acting Insulin Analog in Hexameric and Dihexameric States

Elucidating the structure and dynamics of insulin and its analogs has been of broad interest, while presenting challenges due to the unique structural dynamics of insulin (composed itself of two multiply cross-linked peptides A and B) and its ability to assemble in a variety of oligomeric structures under physiological conditions. Here, we present two distinct X-ray crystallographic structures of the long-acting human insulin analog detemir (INSD) resolved in hexameric and dodecameric (or dihexameric) states at 2.85 [A] and 2.70 [A] resolution, respectively, using diffraction data collected under ambient temperature conditions. Characterization of the collective dynamics of these oligomers using the Gaussian Network Model (GNM) reveals several key features: (i) Oligomerization imparts high cooperativity in structural dynamics evidenced by dissection of the cross-correlations at various hierarchical levels; (ii) detemir monomers conformational flexibility is highly suppressed within oligomeric constructs, the effect being particularly strong in the dihexamer due to the asymmetric packing of the hexamers and the presence of myristoyl groups at B peptides termini whose interactions imparts further heterogeneities; and (iii) a number of key residues retain, however, their intrinsic dynamics, to be deployed upon release from the oligomers. We distinguish in particular residues serving as hinge sites that mediates the conformational dynamics of the asymmetric units (dimers) and monomers (I2A-V3A and Y19 A -C20A, and L11B-L15B and Y26B of the respective peptides A and B), or as anchors supporting structural stability (disulfide-bridge forming cysteines, plus selected residues such as L16A, G8B and R22B-F24B. Overall, this study provides a structural-dynamic framework for gaining new insights into the dynamics of long-acting analog INSD and helps identify actionable sites for modulating insulin (analogs) dynamics toward designing more effective therapeutics.

biophysics↗

bioRxiv2026

Targeted medial prefrontal cortex stimulation prevents incubation of cocaine craving and restores functional connectivity

BackgroundRelapse remains a central obstacle in the treatment of cocaine use disorder (CUD), for which no medications have received approval from the U.S. Food and Drug Administration. Transcranial magnetic stimulation (TMS) has shown promise as a potential therapeutic intervention. However, current clinical trials often rely on a "trial-and-error" approach in target selection and experimental design. We previously developed a novel TMS platform and high-density theta burst stimulation (hdTBS) technology, enabling precise, focal stimulation of the rat medial prefrontal cortex (mPFC), including the prelimbic and anterior cingulate cortices. MethodsWe applied hdTBS intervention to a well-established rat model of cocaine relapse and craving after cessation of extended access intravenous drug self-administration and assessed brain response using resting-state functional magnetic resonance imaging (fMRI). ResultsAs expected, we observed robust time-dependent increases in cocaine seeking (incubation of cocaine craving) in control rats receiving sham stimulation over 3 weeks of abstinence accompanied by a reduction in prefrontal functional connectivity. In contrast, daily sessions of hdTBS for 7 days delivered on abstinence days 14-20 prevented the emergence of the incubation effect and restored prefrontal network functional connectivity. ConclusionsThis study provides strong preclinical evidence demonstrating that precise circuit modulation of medial prefrontal subregions causally reverses both behavioral and network-level adaptations associated with relapse vulnerability. Given the clinical accessibility and established safety profile of TMS, this work provides a mechanistically grounded framework for target selection and supports the translation of focal TMS of the mPFC for relapse prevention in CUD patient. One Sentence SummaryFocal transcranial magnetic stimulation (TMS) of the mPFC using the hdTBS procedure prevents incubation of cocaine craving and restores functional connectivity.

neuroscience↗

bioRxiv2026

Adaptive Integration of Heterogeneous Foundation Models to Find Histologically Predictable Genes in Breast Cancer

Foundation models for computational pathology have rapidly emerged as powerful tools for extracting rich biological and morphological representations from histopathology images. However, variations in model architecture, pre-training data, and optimization objectives often lead to task-dependent performance, rather than universal generalization. As a result, effective strategies for integrating their complementary strengths are essential to fully realize the potential of foundation models for robust histopathology analysis. Meanwhile, recent breakthroughs such as spatial transcriptomics provide an unprecedented opportunity to integrate genetic and histopathology information from the same patient sample, thereby maximizing both molecular and anatomical pathology insights. Specifically, each models embedding is first mapped to gene-level predictions via a dedicated prediction head, enabling model-specific feature utilization. A lightweight weighting network then adaptively aggregates these predictions to produce a unified and robust output at gene and spatial location levels. Across multiple spatial transcriptomics datasets, our approach consistently outperforms both individual foundation models and classical ensembling methods. Focusing on breast cancer, we observe substantial gains in prediction accuracy for clinically relevant PAM50 subtype markers and drug-target genes. Moreover, the proposed framework improves interpretability by revealing model-specific contributions and specialization at the gene level. Overall, our work presents an effective solution to integrating multiple foundation models for enhancing the genetic analyses of histopathology images.

bioinformatics↗

bioRxiv2026

Focal Transcranial Magnetic Stimulation of the Rat Anterior Cingulate Cortex Inhibits Incubation of Opioid Craving after Voluntary Abstinence

Relapse remains a major challenge in opioid addiction treatment, underscoring the need for innovative therapies. Progress in neuromodulation therapies has been limited by insufficient mechanistic understanding of stimulation engagement and disease-related changes in the brain. We used a novel, focal transcranial magnetic stimulation (TMS) system to deliver high-density theta burst stimulation (hdTBS) combined with resting-state fMRI to test whether anterior cingulate cortex (ACC) stimulation reduces relapse-like behavior and alters functional circuitry in a rodent model of opiate dependence. The coil focality and stimulation parameters approximate human TMS protocols, and the targeted region represents a functional homolog of the human ACC. We trained rats to self-administer oxycodone intravenously for 14 days. We then introduced an electric barrier for 13 days, which caused cessation of drug self-administration. We assessed relapse to oxycodone seeking immediately after training (early abstinence) and after electric-barrier exposure (late abstinence). We administered daily hdTBS or sham stimulation for 7 days before the late-abstinence test. Sham-treated rats showed a time-dependent increase in oxycodone seeking during abstinence (incubation of oxycodone craving) and reduced ACC functional connectivity. In contrast, hdTBS prevented the incubation of oxycodone craving and restored ACC connectivity with the dorsal and ventral striatum. Tracer-based axonal-projection data further showed that stimulation-induced effects aligned with regions receiving dense projections from the stimulation site, suggesting that the projection architecture is critical to the propagation of focal stimulation across distributed networks. These findings identify ACC-centered circuits as mechanistically informed targets for TMS-based interventions that aim to reduce opioid relapse during abstinence. One sentence SummaryPrefrontal TMS stimulation reduced relapse-like behavior and restored corticostriatal circuits, highlighting translational targets for addiction treatment.

neuroscience↗