Biology subjects

Pan, X.

Publications and source records attributed to Pan, X..

At least 19 recordsLinked to original sources

bioRxiv2026

Human neutrophils release extracellular RNA that modulates Aspergillus fumigatus gene expression across kingdoms.

Extracellular RNAs (exRNAs) are emerging as potent bidirectional interkingdom effectors of host-pathogen interactions, particularly in plant and insect systems. Here, using small RNA-sequencing and genetic manipulation, we provide evidence consistent with a function of the human neutrophil exRNA-ome in modulating gene expression across kingdoms in the fungal pathogen Aspergillus fumigatus. The neutrophil extracellular space contains a diverse repertoire of RNA, which was corroborated using a droplet-digital stem-loop RT-PCR protocol to measure differential abundance of several tRNA-derived RNAs between two extracellular fractions. A closer inspection of the extracellular miRNA pool revealed enrichment of antimicrobial Let-7 family sRNAs and immunomodulatory miR-223, among others. Secretion of these sRNAs was largely impervious to infection status, suggesting that neutrophils produce and maintain an antimicrobial exRNA environment against A. fumigatus that requires further targeting. Cross-kingdom miRNA target prediction and exRNA delivery experiments revealed that putative fungal targets are modulated independently of the fungal argonaute proteins and in a donor-dependent manner. Although heterogeneous, for donors that exhibited a silencing effect, mutated target sequences in the fungus did reverse silencing. We hypothesize that the exRNA landscape produced by neutrophils contributes to their multifaceted role in antifungal defense.

microbiology↗

bioRxiv2026

MERFISH-based pooled genetic screening reveals cell non-autonomous phenotypes in intact tissue

Genetic perturbations can reshape both the state of a targeted cell and the behavior of neighboring cells. Measuring these effects systematically in intact tissues requires pooled screening methods that jointly identify perturbations, molecular phenotypes, and spatial context of individual cells. Here, we report an imaging-based pooled screening method that detects perturbation identities and endogenous RNAs without enzymatic amplification. Relative to our previous enzymatic-amplification-based workflow, amplification-free MERFISH detection with direct hybridization increased recovery of perturbation barcodes by ~10-fold and endogenous transcripts by 2.5-fold, improving perturbation assignment as well as molecular profiling of both perturbed and neighboring cells. Applying this method to a genetic mosaic mouse liver screen targeting ~200 genes in hepatocytes, we profiled 1.5 million cells in intact tissues and mapped both cell-autonomous responses of perturbed hepatocytes and cell non-autonomous changes in neighboring immune and vascular populations. These included localized accumulation of pDC-like cells near hepatocytes with impaired V-ATPase function and angiogenic responses in vascular cells near Vhl-perturbed hepatocytes with hypoxia activation. Our method provides a simplified, scalable approach to linking genetic perturbations with cell state, multicellular response, and tissue organization.

genomics↗

bioRxiv2026

Auditing Protein-Protein Interaction Signals with Sparse Autoencoder Fingerprints

Protein language models have become a dominant foundation for sequence-based protein-protein interaction (PPI) prediction, but their generalization remains limited under stringent evaluation, and benchmark accuracy alone cannot reveal whether a PPI predictor learns partner-specific biological signals or exploits contextual shortcuts. Here we introduce AuditPPI, an interpretable framework that transforms sparse-autoencoder (SAE) features from a frozen protein language model into order-invariant pair fingerprints, recasting PPI prediction as an auditable tabular-learning problem. AuditPPI achieves competitive predictive performance across diverse PPI benchmarks while enabling multiscale auditing of the information supporting its predictions. Protein-level audits show that partner-independent participation signals remain substantial in conventional benchmarks, including protein-disjoint benchmark, but are largely insufficient for pair-level discrimination in topology- and degree-controlled benchmark. Pair-level analyses further show that removing protein identity overlap does not eliminate contextual structure: protein-disjoint benchmarking retains strong subcellular co-localization signals and benchmark-dependent feature matching captured by SAE co-activation and absolute-difference. Structural analyses, however, provide little evidence that these predictive signals are specifically grounded in PPI interfaces: feature enrichment at interfaces for protein-disjoint benchmark was no longer detectable after controlling for surface exposure, and contact-specific enrichment was observed for only a small fraction of testable SAE feature pairs. AuditPPI therefore separates predictive success from mechanistic fidelity and provides a practical framework for evaluating whether sequence-based PPI predictors capture partner-specific biological signals instead of contextual shortcuts.

bioinformatics↗

bioRxiv2026

Inhibition of macrophage neuraminidase 1 protects against immune thrombocytopenia by limiting platelet clearance

Immune thrombocytopenia purpura (ITP) is an autoimmune disorder characterized by a reduction in circulating platelet levels, primarily due to generation of autoantibodies to platelet surface antigens followed by their spleen macrophage-mediated clearance. Emerging evidence implicates neuraminidase (sialidase) enzymes including neuraminidase 1 (NEU1) in platelet clearance and ITP severity; however, the underlying cellular mechanisms remain unknown. Using tissue-specific NEU1 knockout mouse models, we studied the contribution of platelet and macrophage NEU1 to ITP pathogenesis and evaluated whether pharmacological inhibition of NEU1 could preserve platelet counts in a murine ITP model. Constitutive and macrophage-specific, but not platelet-specific, NEU1 knockout mice showed a protection against reduction of platelet counts in the passive ITP model suggesting that macrophage, but not platelet, NEU1 promotes platelet clearance. Genetic deletion or pharmacological blockade of macrophage NEU1 also reduced platelet phagocytosis by cultured macrophages in vitro. The selective NEU1 inhibitor CG33301 protected mice against anti-CD41a antibody-induced thrombocytopenia and showed a higher efficacy compared to pan neuraminidase inhibitor oseltamivir phosphate. Our results demonstrate that the macrophage pool of NEU1 plays a central role in platelet clearance by splenocytes during ITP by activating their phagocytosis and suggest that selective NEU1 inhibition may be a promising therapeutic strategy for this disease.

biochemistry↗

bioRxiv2026

VGLL3 Links Pericyte Hypercontractility to Perivascular Fibrosis of the Cerebral Microcirculation, a Novel Vasculopathy Leading to Distinct Long-Term Cerebral Autoregulation Dysfunction After Subarachnoid Hemorrhage

BACKGROUND: Cerebral ischemia following subarachnoid hemorrhage (SAH) has traditionally been considered transient because functional alterations of the cerebral microcirculation are thought to be self-limiting. However, we identified a previously unrecognized vasculopathy, perivascular fibrosis of the cerebral microcirculation (PFCM), characterized by excessive type I collagen deposition after SAH. This study investigated the mechanisms underlying PFCM and its subsequent effects on cerebral hemodynamics. METHODS: In vivo SAH was modeled in mice by autologous blood injection, whereas oxygenated hemoglobin (OxyHb) exposure was used to mimic SAH in vitro. Pericyte-deficient mice (Pdgfr{beta}+/-) and pericyte-specific vestigial-like family member 3 (VGLL3) conditional knockout mice (Vgll3{Delta}PC) were generated. Pericyte contractility was measured by nanoindentation and traction force microscopy. Molecular mechanisms were examined using Western blotting, immunofluorescence, CUT&Tag, RNA-seq, transmission electron microscopy, and molecular docking. PFCM, impaired dilation of the cerebral microcirculation, and cerebral autoregulation were assessed by two-photon imaging, transcranial Doppler with continuous blood pressure monitoring, super-resolution ultrasound imaging, and photoacoustic imaging. RESULTS: After SAH, mice developed long-term cerebral autoregulation dysfunction marked by impaired dilation of the cerebral microcirculation, with the abnormality being most evident within the relatively lower blood pressure range. The marked reduction in PFCM in Pdgfr{beta}+/- mice indicated that pericytes were the principal cellular contributors. Mechanistically, OxyHb-induced cytoskeletal remodeling in vitro increased pericyte contractility and promoted nuclear translocation of SAH-upregulated VGLL3. This was followed by increased genomic occupancy, Col1a1 transcriptional activation, and type I collagen deposition. Pericyte-specific VGLL3 knockout abolished PFCM and, consequently, significantly alleviated long-term cerebral autoregulation dysfunction. CONCLUSIONS: Our findings identify PFCM mediated by pericytic VGLL3 as a novel vasculopathy leading to long-term cerebral autoregulation dysfunction after SAH.

neuroscience↗

bioRxiv2026

A complementary learning system for continual episodic memory in large language models

Humans retain memories of individual experiences for a lifetime, an ability attributed to a complementary learning system in which a fast process encodes episodes and a slow process integrates them into semantic knowledge. In classical Hebbian models such as Hopfield networks, memory traces are superposed in shared weights. This makes learning naturally continual but causes strong interference among correlated memories, a failure that reappears as catastrophic forgetting in deep networks. Here we use a large language model as a model system for continual episodic memory, with its pretrained weights supplying the semantic context in which new episodes are embedded. Fast learning is implemented by a hippocampus-like module that assigns each episode to a dedicated, extremely sparse low-rank adapter; competitive gating then selects among these separated traces during recall. Across streams of up to 1,000 factual and autobiographical episodes, each adapter requires only 2-3 parameters per token while preserving excellent recall. An internal retrieval-augmented generation mechanism reconstructs the selected episode in context and supports high-accuracy question answering over stored memories. Finally, slow cortical consolidation is modeled by fine-tuning the base weights through batch replay, enabling reconstruction and direct question answering without episodic adapters. Together, fast storage and slow consolidation implement both components of a complementary learning system within a single language model, yielding a neural-network model that stores, recalls, and consolidates naturalistic episodic memories, thereby capturing key functional features of human memory.

neuroscience↗

bioRxiv2026

A leakage-controlled benchmark shows apparent codon-language-model advantages in synonymous-variant prediction are evaluation artifacts

Synonymous codon choices shape mRNA stability, translation, and folding, and codon language models (cLMs) are increasingly reported to read this biology from sequence. However, when a true signal is thin relative to a confounding one, standard evaluation protocols can manufacture the reported gain rather than measure it--and we show this is what has happened for cLMs on synonymous-variant prediction. Under random splits, the codon advantage is large: tokenization gaps of +2.3-14.3 percentage points (pp) and pretrained codon leads of +2.9 pp over the strongest protein model (ESM-1b) and up to +4.9 pp over ESM-2. We find these numbers are properties of the measurement, not the models. A memorization baseline outscores every neural model; the advantage collapses under gene-held-out evaluation; the sole surviving residual dissolves into six defensible probe defaults; and the synonym-randomization drop that appeared to confirm true signal is itself variance under pooled analysis (0.3 pp, p = 0.49). No advantage survives leakage-controlled evaluation with pooled statistics. We release CodonBench, a leakage-controlled benchmark with an emergent audit cascade, and characterize how artifacts accumulate at every pipeline step. A thin signal (I({sigma}; Y |A) {approx} 0.04 bits) may exist but is not reliably detectable at current sample sizes; we specify what detecting it would require.

bioinformatics↗

bioRxiv2026

Distal conformational steering by N-terminal pyroglutamylation enables subtype-selective GPCR activation across Aplysia PRXamide and human Neuromedin U signaling

Post-translational modifications (PTMs) diversify neuropeptide function, yet how minimal modifications encode receptor specificity without direct contact remains a fundamental challenge in chemical biology. Pyroglutamylation (pQ), a prevalent N-terminal PTM in bioactive peptides, introduces a rigid cyclic constraint, yet its mechanistic role in receptor signaling is unclear. Here, using two newly identified endogenous Aplysia PRXamide receptors (ApPRXa R1 and ApPRXa-R2) as a model system, we find that the same ligand, MMG2-pDPb (pQPPLPRYamide), produces opposite functional outcomes: pQ suppresses ApPRXa-R1 activation while enhancing ApPRXa-R2 activation. In vitro and in silico analyses demonstrate that the N-terminal pQ/Q remains solvent-exposed and does not directly contact receptor residues. Instead, pQ reshapes the ligand conformational ensemble and redistributes interaction networks across shared receptor contact sites. Strikingly, this molecular logic extends to mammalian Neuromedin U (NmU) receptors, as canine NmU (pQFLFRPRNamide) similarly biases subtype preference of human NmU receptors. Both static and dynamic analyses further reveal that receptor pocket mechanics determine the direction of this modulation: a loose and permissive pocket better accommodates the rigid pQ-constrained ligand, whereas a more compact pocket favors the non-pyroglutamylated ligand. These findings define a "PTM distal steering" mechanism that bridges "lock-key" and "induced-fit" paradigms and establish a general principle by which a minimal, non-contacting modification encodes receptor preference through ligand conformational biasing and pocket-dependent permissiveness, providing a chemical framework for optimizing stable, conformation-biased neuropeptide analogs.

biochemistry↗

bioRxiv2026

Extracellular vesicles derived from cells overexpressing HGSNAT rescue defects in Mucopolysaccharidosis IIIC neurons

Mucopolysaccharidosis III type C (MPS IIIC) is a rare neurological lysosomal storage disorder caused by genetic deficiency of the lysosomal membrane enzyme, heparan--glucosaminide N-acetyltransferase (HGSNAT). To assess the feasibility of therapeutic strategies based on cross-correction of neurons by HGSNAT secreted from transplanted cells overexpressing the enzyme, we generated induced cortical neurons (iCN) from induced pluripotent stem cells (iPSCs) derived from MPS IIIC patients. The neurons were treated with extracellular vesicles (EV) purified from the culture medium conditioned by human endothelial cells transduced with a lentiviral vector encoding EGFP-tagged HGSNAT (LV-HGSNAT-EGFP). The isolated EV showed supraphysiologic HGSNAT activity levels and efficiently delivered the enzyme to the lysosomes of MPS IIIC iCN reducing lysosomal size and restoring normal synaptic protein levels. EV-mediated delivery of HGSNAT to neurons was further confirmed by the analysis of MPS IIIC iCN either co-cultured with iPSC-derived MPS IIIC microglia (iMGL) transduced with LV-HGSNAT-EGFP or treated with the iMGL conditioned medium. MPS IIIC iCN co-cultured with iMGL overexpressing HGSNAT achieved a complete phenotypic rescue, including normalization of lysosomal size, and the levels of heparan sulfate, GM2-ganglioside, synaptic proteins and brain-derived neurotropic factor. Treatment of MPS IIIC iCNs with conditioned medium led to a partial defects correction. Our findings reveal the translational potential of EV-mediated enzyme delivery in MPS IIIC patients treated with LV-mediated haematopoietic progenitor stem cell gene therapy.

biochemistry↗

bioRxiv2026

Systematic Development of a Compact Genome-Editing Tool Leveraging the TAM-Independent DNA Nuclease TasR in Bacillus subtilis

Bacillus subtilis is a core microbial chassis in biomanufacturing, and establishing efficient gene editing technologies is key to engineering this strain. In conventional CRISPR gene editing technologies, the large size of DNA nucleases leads to difficulties in plasmid construction, low transformation efficiency, and cumbersome multi-round editing operations; therefore, developing miniature gene editing tools can effectively address these issues. Although our group previously established a miniature gene editing tool based on IscB in B. subtilis SCK6, IscB relies on the 5'-CAGGAA-3' TAM recognition sequence, and 83.36% of the genes in the SCK6 genome harbor no or only one TAM sequence, indicating a bottleneck of restricted editing for IscB in this strain. The novel miniature DNA nuclease TasR does not require a TAM sequence and can thus compensate for the limitation of IscB; however, the applicability of TasR in B. subtilis remains unknown. Therefore, this study first constructed a single plasmid, pBsuTasR, capable of expressing TasR and its guide RNA (tigRNA), which enabled gene deletion of regular-sized fragments in SCK6 with editing efficiencies of 21.7%- 78.3%. Subsequently, the capacity of TasR to delete a long DNA fragment (169.9 kb) was evaluated, and it was found that under the guidance of a single tigRNA, the deletion efficiency was 21.73%, whereas after optimizing to two tigRNAs, the efficiency increased to 39.13%. Furthermore, the gene integration capability of pBsuTasR was further tested, and TasR was able to integrate the aprN gene into the amyE locus at an efficiency of 13.3% under the guidance of a single tigRNA, and after increasing to two tigRNAs, the integration efficiency increased to 91.3%. In terms of iterative genome editing, this study developed the pBsu-SRP (Scissors-Rock-Paper) iterative editing system, which automatically cures the editing plasmid from the previous round while performing a new round of gene editing, with sequential gene deletion efficiencies of 4.34%-26.08%, and using this system, the editing cycle can be shortened from 4N days by the conventional method to 3N+1 days. Subsequently, the pBsu-SRP system was successfully used to achieve the integration of two and three copies of the mCherry fluorescent reporter gene in SCK6, and it was found that the fluorescence intensity increased with the copy number. Finally, this study also explored the escape of SCK6 from TasR cleavage and found that mutations in the tigRNA sequence are the cause of the escape. In summary, this study constructed a novel miniature genome editing system in B. subtilis using the TAM-independent nuclease TasR as the core component. This system can not only provide an efficient technical tool for genetic manipulation of industrial microorganisms, but also offer new instrumental support for the iterative engineering and functional optimization of chassis cells in biomanufacturing.

molecular biology↗

bioRxiv2026

IgGM2: An All-Atom Foundation Model for Adaptive Immune Receptor Design

Accurate immune receptor design requires modeling the coupled variation of aminoacid sequence, full-atom conformation, and target-binding geometry across antibodies, nanobodies, and T-cell receptors (TCRs). Existing methods often address only part of this problem, either by separating structure generation from sequence design, relying on fixed-backbone inverse folding, or focusing on a single receptor class. We introduce IgGM2, a unified all-atom generative framework for immune receptor structure prediction and CDR sequence-structure co-design. IgGM2 follows a structure-to-design strategy: it first learns how immune receptors are positioned around fixed target structures, and then transfers this target-conditioned structural prior to CDR design. Unlike modular design pipelines, IgGM2 jointly generates CDR residue identities and full-atom receptor structures, allowing frame-work geometry to adapt to designed CDRs without separate inverse folding or external sidechain packing. Unlike continuous residue encodings based on virtualatom geometry, IgGM2 keeps sequence prediction explicit while using atom14 placeholders only for full-atom representation. On structure prediction benchmarks, IgGM2 better captures receptor-target spatial relationships than AlphaFold3 on FoldBench and achieves strong performance on TCR-pMHC modeling. On sequence design benchmarks, IgGM2 achieves competitive amino-acid recovery and improves Rosetta-based interface preference metrics, suggesting more favorable generated binding interfaces. These results support IgGM2 as a unified all-atom framework for adaptive immune receptor structure prediction and design.

bioinformatics↗

bioRxiv2026

A modular generalist-specialist AI framework for ROI selection across spatial profiling workflow

Selection of regions of interest (ROIs) is often a crucial step in spatial molecular profiling and many pathology tasks, with substantial implications for research reproducibility and biological interpretability. To provide a reproducible and adaptive framework for AI-guided ROI selection, we developed a modular generalist-specialist solution across spatial profiling platforms. In a cohort comprising 55 tumor types from 160 tissue donors profiled using NanoString Digital Spatial Profiling and multiplex immunofluorescence, we first established a protein-profiling reference atlas capturing compartment-specific immune, checkpoint, stromal, and proliferation patterns. We then developed an AI Specialist Task-Oriented Model for ROI Selection (ASTROS) and tested comprehensive benchmarks considering specialist-only (ASTROS), generalist-only (PLIP/GFM), and hybrid generalist-specialist strategies, showing that the latter provides a balanced tradeoff across slide-level signal preservation, pathologist-reference concordance, within-slide placement consistency, and large-slide computational efficiency. We further demonstrated the feasibility of virtual staining for ROI preview and modular ROI placement for other spatial omics technologies, Visium and Visium HD workflows. Together, these results support our proposed framework to enable ROI selection responding to unmet needs for reducing inter-rater variability, reproducibility, and versatility in spatial profiling experiments.

pathology↗

bioRxiv2026

Prenatal e-cigarette exposure induces region- and cell type-specific metabolic and developmental reprogramming

Prenatal e-cigarette exposure (PeCE) is increasingly prevalent and has been associated with adverse neurodevelopmental outcomes, yet how maternal vaping perturbs early brain development remains poorly understood. We integrated spatial transcriptomics, snRNA-seq and lipidomics to define neonatal rat brain responses to PeCE in rats at regional and cellular resolution. PeCE induced pronounced spatial heterogeneity in vulnerability, with the striatum exhibiting the strongest developmental transcriptional disruption. PeCE disrupts lipid metabolic homeostasis and induced molecular signatures suggestive of enhance Ca2+ signaling, dopaminergic responsiveness and region-specific synaptic stress programs. In the striatum, PeCE induced transcriptional programs consistent with altered lipid utilization, silent synapse-like molecular features, suppressed dendritic spine development and delayed D1-medium spiny neuron maturation. PeCE-sensitive genes were enriched in human autism spectrum disorders and neurodevelopment risk loci. Together, these findings identify disrupted metabolic and developmental reprogramming as a central feature of neonatal brain vulnerability to maternal vaping and provide a mechanistic framework linking PeCE to neurodevelopmental disease risk.

neuroscience↗

bioRxiv2026

Systematic benchmarking of zero-shot utility and robustness in single-cell transcriptomic foundation models

Single-cell foundation models (scFMs) have emerged as powerful representation learning approaches for single-cell transcriptomics. However, the utility and robustness of their pretrained representations across diverse analytical tasks and data conditions remain insufficiently characterized, particularly in zero-shot settings without task-specific fine-tuning. Here, we systematically analyze zero-shot performance of single-cell transcriptomic representations across 20 methods, 6 downstream tasks and 1,607 datasets comprising nearly 21.8 million cells. We evaluate model behavior along three complementary dimensions: utility on original datasets, robustness to controlled changes in dataset structure, and exploratory associations between dataset characteristics and performance variation. Our results show that scFM performance is strongly task dependent, with no single method consistently outperforming others across cell- and gene-level analyses. Notably, high utility on original datasets did not necessarily translate into robustness under structural perturbations, and several top-ranking methods were sensitive to changes in cell number, gene number, class composition, class imbalance, and batch complexity. Conventional statistical and task-specific methods remained competitive in several settings, while greater computational cost did not consistently correspond to better performance. Driver analyses further identified task-specific associations between performance and dataset characteristics, including cell-type complexity, train-test class overlap, batch number, and regulatory target-set size. Together, these findings show that the zero-shot utility and robustness of pretrained scFM representations depend jointly on analytical task and dataset structure. Our study provides a practical basis for context-aware representation selection and underscores the importance of evaluating structural robustness alongside utility when developing and applying scFMs.

bioinformatics↗

bioRxiv2026

Trichome acylsugar defenses are equally effective against leaf-mining and free-living insects

Trichomes are epidermal projections, crucial components of plant defense systems. Whether these epidermal defenses are effective against leafminers, which infest the internal mesophyll layers, remains unclear. Using the devasting invasive tomato pest, Phthorimaea absoluta (tomato leafminer), and the trichome-rich wild tomato, Solanum pennellii, we demonstrate that trichome produced acylsugars effectively inhibit the growth of endophagous P. absoluta larvae feeding inside leaves. High spatial resolution leaf- and frass-metabolite analysis by MALDI-MSI and Stimulated Raman Scattering microscopy revealed that trichome-synthesized acylsugars are translocated into the mesophyll tissues when larvae infest leaves. We identified an ATP-binding cassette (ABC) transporter, SpABCB5, as the mediator of this translocation and as an essential component of defense against leafminers. These data reveal that epidermal trichomes not only defend against free-living (ectophagous) herbivores through direct contact or rapid deposition of chemical defenses onto leaf surfaces, but also defend against endophagous leafminers by rapidly depositing chemical defenses into internal tissues. By delivering mines against leaf miners, trichomes offer new avenues for breeding crops with resistance to leafminers.

plant biology↗

bioRxiv2026

Global cell-state and gene-program representations reveal conserved and context-specific perturbation responses of cells

Understanding how cell states change in response to genetic perturbations is critical for gene-function and therapeutics discovery. However, state-of-the-art deep-learning models trained on large single-cell omics datasets still struggle to accurately predict cellular responses to perturbations, highlighting the need for a better understanding of the cell-state space and how cells move through this space. Here, we present a contrastive learning model that integrates diverse scRNA-seq datasets into a global, interpretable cell-state manifold. We further develop a framework to integrate this global cell-state manifold with genome-scale perturbation data to identify gene-expression programs that define principal axes of cell-state transitions and functional embeddings of genes that define major perturbation classes. Applying this framework across Perturb-seq datasets on different cell types reveals conserved cellular responses to perturbations, as well as cell-type-specific rewiring of stress responses. Moreover, we perform a genome-scale Perturb-seq screen in human embryonic stem cells, validating and extending these findings and uncovering a class of mesenchymal transitions induced by diverse perturbations to cellular stress-response pathways.

systems biology↗

bioRxiv2026

eSIG-Net: Accurate prediction of single-mutation induced perturbations on protein interactions using a language model

Most proteins exert their functions in complex with other interactors. Single mutations can exhibit a profound impact on perturbing protein interactions, leading to human disease. However, predicting the effect of single mutations on protein interactions remains a major computational challenge. Deep learning, particularly protein language models or transformers, has become an effective tool in bioinformatics for protein structure prediction. However, the functional divergence of mutations makes it difficult to predict their interaction perturbation profiles. To address this fundamental challenge, we present eSIG-Net (edgetic mutation Sequence-based Interaction Grammar Network), a novel sequence-based "Interaction Language Model" for predicting protein interaction alterations caused by single mutations. eSIG-Net combines various protein sequence embeddings, introduces a mutation-encoding module with syntax and evolutionary insights, and employs contrastive learning to evaluate mutation-induced interaction changes. eSIG-Net significantly outperforms current state-of-the-art sequence-based and structure-based prediction methods at predicting mutational impact on protein interactions. We highlight examples where eSIG-Net nominates causal variants with high confidence and elucidates their functional role under relevant biological contexts. Together, eSIG-Net is a first-in-kind "interaction language model" that can accurately predict interaction-specific rewiring by single mutations with only sequence information, and exhibits generalizability across biological contexts.

bioinformatics↗

bioRxiv2026

Effects and Mechanisms of Aerobic Exercise on Myocardial AGEs/RAGE-p38 MAPK-NF-κB Pathway in SHR Rats

ObjectiveTo investigate the effects of aerobic swimming exercise on blood pressure, cardiac function, and the myocardial AGEs/RAGE-p38 MAPK-NF-{kappa}B signaling pathway in spontaneously hypertensive rats (SHR). MethodsTwenty-four male Wistar-Kyoto (WKY) rats were randomly assigned to three control subgroups (C-0, C-4, C-8; n=8 each). Fifty-six SHRs were allocated into seven subgroups (S-0, S-4, S-8, SE-4, LE-4, SE-8, and LE-8; n=8each) to receive different swimming intervention protocols. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured via the tail-cuff method. Cardiac parameters, including ejection fraction (EF), fractional shortening (FS), left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), and stroke volume (SV), were assessed using echocardiography. Myocardial morphological alterations were observed through hematoxylin-eosin (HE) staining, and myocardial hydroxyproline content was quantified using the alkaline hydrolysis method. Furthermore, the myocardial expressions of advanced glycation end products (AGEs), receptor for AGEs (RAGE), phosphorylated p38 mitogen-activated protein kinase (p-p38 MAPK), and nuclear factor-kappa B (NF-{kappa}) were detected by immunohistochemistry (IHC). ResultsAerobic exercise significantly reduced blood pressure in SHRs, partially improved cardiac function and myocardial architecture, and decreased hydroxyproline content (except in the SE-4 group). Furthermore, the exercise intervention downregulated the expressions of AGEs, RAGE (except in the SE-4 group), p-p38 MAPK, and NF-{kappa}, with efficacy varying according to exercise duration and intervention cycles. ConclusionAerobic exercise alleviates the progression of hypertension and mitigates the risk of cardiac dysfunction in SHRs by inhibiting myocardial glycation. This cardioprotective effect may be mediated by the suppression of p38 MAPK activation and the subsequent reduction of NF-{kappa} nuclear translocation.

physiology↗