Biology subjects

Wang, L.

Publications and source records attributed to Wang, L..

At least 253 records · Page 14Linked to original sources

bioRxiv2026

Mechanism of substrate recognition and catalysis in a mammalian phosphatidylserine synthase 2

Mammalian phosphatidylserine synthase-1 and -2 synthesize phosphatidylserine (PS) by replacing the headgroup of either phosphatidylcholine (PC, PTDSS1) or phosphatidylethanolamine (PE, PTDSS2) with a serine. We determined structures of PTDSS2 from Equus caballus in complex with either PE or serine substrates to resolutions of 2.8-3.2 [A]. The structures define substrate binding sites and reveal that the phosphate group of PE is coordinated by two Ca2+. In addition, we found that PTDSS2 has significant phospholipase D (PLD) activity in the absence of serine, which was not reported previously, and that Ca2+ is required for the PLD activity. These discoveries enrich our knowledge in the mechanism of mammalian PTDSS.

biochemistry↗

bioRxiv2026

Structural basis of canonical TIR-NLR activation in plant innate immunity

In plants, intracellular NLRs (nucleotide-binding leucine-rich repeat receptors) detect pathogen effector proteins, form oligomeric resistosomes, and activate ETI (effector-triggered immunity). NLRs contain N-terminal signaling, central NB-ARC (nucleotide-binding) and C-terminal LRR (leucine-rich repeat) domains. NLRs with N-terminal TIR (Toll/interleukin-1 receptor) domains (TNLs) hydrolyze NAD+ (nicotinamide adenine dinucleotide) to generate signaling molecules. We determined cryo-EM structures of flax M, a canonical non C-JID (C-terminal jellyroll/Ig-like domain) TNL, in both monomeric autoinhibited conformation, and tetrameric resistosome after activation by its rust fungal effector AvrM-A. AvrM-A homodimers dissociate into monomers to bind directly to the LRR and NB-ARC domains in the M resistosome. The resistosome structure includes a non-hydrolyzable NAD+ analogue, revealing the substrate NAD+ recognition mechanism by the TIR domains. M cleaves NAD+ and generates the same signaling compounds as the related flax TNL, L6. Our findings explain the mechanism of TNL signaling, and provide a basis for rational engineering of disease-resistant crops.

plant biology↗

bioRxiv2026

Ripk1 Deficiency or K376R Mutation in Dendritic Cells Promotes Potent Antitumor Immunity

Dendritic cells (DCs) are critical for initiating adaptive immunity through antigen presentation and T cell priming. While immunogenic cell death (ICD) enhances anti-tumor immunity, how programmed cell death in DCs shapes tumor immune responses remains unclear. Receptor-interacting protein kinase 1 (RIPK1) regulates apoptosis, necroptosis, and inflammation. Here, we show that mice with disrupted RIPK1 ubiquitination on K376 (Ripk1fl/K376R Cd11c-Cre, Ripk1DC-KO/K376R) develop inflammation and autoimmunity, phenocopying DC-specific RIPK1-deficient(Ripk1fl/fl Cd11c-Cre, Ripk1DC-KO) mice. These phenotypes were rescued by the genetic ablation of Ripk3 or Mlkl, establishing that RIPK1 ubiquitination is essential for suppressing DC necroptosis to maintain immune homeostasis. Remarkably, both Ripk1DC-KO and Ripk1DC-KO/K376R mice exhibit robust resistance to tumor growth, characterized by significantly enhanced T cell infiltration and activation. Mechanistically, we show that Ripk1 deficiency or the loss of its K376 ubiquitination augments antigen-presenting capacity of DCs through both necroptosis-dependent and -independent pathways. Furthermore, adoptive transfer of T cells from Ripk1DC-KO mice further amplifies anti-tumor immunity in recipient hosts. These results identify RIPK1 as a critical immunological checkpoint in DCs that constrains their immunogenic potential. Targeting this pathway represents a promising therapeutic strategy to potentiate T cell-mediated cancer immunotherapy.

immunology↗

bioRxiv2026

A lightweight, ultrafast and general embedding framework for large-scale spatial omics data

Deciphering ultra-large-scale omics data with minimal resources while maintaining high computational efficiency is a longstanding challenge in biology. Here, we present Local Pooling (LP), a lightweight, ultrafast and general framework that leverage neighbor-indexing strategy and local pooling module to generate omics embedding and compatible with variety of downstream analyses. We developed its adaptation for spatial omics, SpaLP, and evaluated it on over 20 large-scale datasets spanning 9 technology platforms. SpaLP consistently outperformed baseline methods across multiple benchmarks, including niche identification, expression reconstruction, multiple slices integration, 3D organ construction, multi-omics integration and cross-platform generalization. Notably, SpaLP processed a 1.35-million-cell mouse embryo slice in just 47 seconds, achieving up to 300-fold increase in computational efficiency compared to Graph Neural Network (GNN)-based methods. Meanwhile, SpaLP increased the average adjusted rand index (ARI) by over 30% for niche identification in simulated and realistic settings. Furthermore, we applied SpaLP to integrate 8.4 million mouse brain cells within 4 minutes on a single GPU and constructed a 3D spatial atlas. Finally, we explored SpaLPs ability of cross-platform generalization and potential for developing an omics foundation model. As a novel and general framework, we believe that LP could help more researchers develop new model on large-scale data and overcome the research barriers caused by computing resources in more fields.

bioinformatics↗

bioRxiv2026

The genetic repertoire of the deep sea: from sequence to structure and function

The deep sea as the largest and maybe most hostile environment on Earth is still underexplored especially regarding its genetic repertoire. Yet, previous work has revealed significant habitat-specific deep-sea biodiversity. Here, we present an integrated deep-sea genetic dataset comprising 502 million nonredundant genes from 2,138 samples and 2.4 million predicted structures, and used it to link specific protein structures with genetic variants associated with life in the deep sea and to assess their biotechnology potential. Combining global sequence analysis with biophysical and biochemical measurements revealed unprecedented sequence diversity, yet substantial structural conservation of proteins. Especially proteins involved in replication, recombination, and repair were identified to be under rapid evolution and with specialized properties. Among these, a structurally divergent helicase exhibited advantages in controlling nanopore sequencing speed. Thus, our work positions the deep sea as a unique evolutionary engine that generates and hosts genetic diversity and bridges genetic knowledge with biotechnology.

bioinformatics↗

bioRxiv2026

Defining and rescuing pathomechanisms of myotubularin and autophagy disruption in a novel human cell model of Charcot-Marie-Tooth Type 4B3

Charcot-Marie-Tooth Type 4B3 (CMT4B3) is a genetic disorder leading to peripheral axon degeneration and clinical manifestations of distal weakness and gait impairment. CMT4B3 is caused by mutations in SBF1/MTMR5, a negative regulator of phosphoinositide signaling and autophagy. Although SBF1 mutations are ubiquitously expressed, how and why loss of SBF1/MTMR5 exerts deleterious effects predominantly in neurons of the peripheral nervous system (PNS) is unknown. To investigate the effects of mutant SBF1/MTMR5 on PNS neurons compared to non-neurons, we engineered a novel and unique model system of CMT4B3 using human induced pluripotent stem cells (iPSCs) differentiated into key components of the PNS: motor neurons (iMNs), sensory neurons (iSNs), or skeletal muscle (iMuscle). To model CMT4B3, we used iPSCs derived from a CMT43B patient, or genetically knocked down SBF1 in WT cells. Strikingly, CMT4B3 iMNs showed the highest degree of cell degeneration among all cell types, concordant with the clinical phenotype of patients. We also found that CMT4B3 iMNs and iSNs showed attenuated expression of MTMR5 and related paralogs MTMR2 and MTMR13. Knockdown of SBF1 most significantly augmented autophagy in iMNs than other cell types. Finally, we tested treatment with VPS34-IN1, a pharmacologic inhibitor of the Class III PI3-Kinase functioning in opposition to MTMR5 in regulating phosphoinositides, and found that VPS34-IN1 rescued cell death in CMT4B3 iMNs. Together, our results for the first time confirm PNS cell type-specific differences in myotubularin expression, autophagy, and vulnerabilities to SBF1 mutations, and identify a novel therapeutic strategy of high disease-modifying potential for CMT4B3.

neuroscience↗

bioRxiv2026

Joint Modeling of Transcriptomic and Morphological Phenotypes for Generative Molecular Design

BackgroundEffectively translating data on complex cellular responses from transcriptomic and morphological measurements into molecular design remains a significant computational challenge. Existing generative methods operate on single modalities and condition on post-treatment measurements without leveraging paired control-treatment dynamics to capture perturbation effects. ResultsWe present Pert2Mol, a framework for multi-modal phenotype-to-structure generation that integrates transcriptomic and morphological features from paired control-treatment experiments. Pert2Mol employs bidirectional cross-attention between control and treatment states to capture perturbation dynamics, conditioning a rectified flow transformer that generates molecular structures along straight-line trajectories. We introduce Student-Teacher Self-Representation (SERE) learning to stabilize training in high-dimensional multi-modal spaces. On the Ginkgo Data Platform (GDP) dataset, Pert2Mol achieves Frechet ChemNet Distance of 4.996 compared to 7.343 for diffusion baselines and 59.114 for transcriptomics-only methods, while maintaining perfect molecular validity and appropriate physicochemical property distributions. The model demonstrates 84.7% scaffold diversity and 12.4 times faster generation than diffusion approaches with deterministic sampling suitable for hypothesis-driven validation. ConclusionsPert2Mol establishes a new paradigm for linking high-content phenotypic screening data with computational hypothesis generation in drug discovery: joint multi-modal perturbation modeling enables more accurate and structurally diverse molecular design than single-modality or diffusion-based approaches. Code and pretrained models are available at https://github.com/wangmengbo/Pert2Mol.

bioinformatics↗

bioRxiv2026

PepMCP: A Graph-Based Membrane Contact Probability Predictor for Membrane-Lytic Antimicrobial Peptides

MotivationThe membrane-lytic mechanism of antimicrobial peptides (AMPs) is often overlooked during their in silico discovery process, largely due to the lack of a suitable metric for the membrane-binding propensity of peptides. Previously, we proposed a characteristic called membrane contact probability (MCP) and applied it to the identification of membrane proteins and membrane-lytic AMPs. However, previous MCP predictors were not trained on short peptides targeting bacterial membranes, which may result in unsatisfactory performance for peptide studies. ResultsIn this study, we present PepMCP, a peptide-tailored model for predicting MCP values of short peptides. We collected more than 500 membrane-lytic AMPs from the literature, conducted coarse-grained molecular dynamics (MD) simulations for these AMPs, and extracted their residue MCP labels from MD trajectories to train PepMCP. PepMCP employs the GraphSAGE framework to address this node regression task, encoding each peptide sequence as a graph with 4-hop edges. PepMCP achieved a Pearson correlation coefficient of 0. 883 and an RMSE of 0. 123 on th e node-level test set. It can recognize membrane-lytic AMPs with the predicted MCP values for each sequence, thereby facilitating mechanism-driven AMP discovery. Additionally, we provide a database, MemAMPdb, which includes the membrane-lytic AMPs, as well as the PepMCP web server for easy access. Availability and ImplementationThe code and data are available at https://github.com/ComputBiophys/PepMCP. Contactc.song@pku.edu.cn Supplementary InformationSupplementary data are available online.

bioinformatics↗

bioRxiv2026

Generating Joint Transcriptomic and Morphological Responses to Drug Perturbations via Rectified Flow

MotivationPredicting cellular responses to drug perturbations requires capturing complex dependencies between transcriptomic and morphological changes. Existing approaches model these modalities in isolation, missing critical molecular-phenotypic relationships that occur simultaneously during drug treatment. No current method jointly predicts gene expression profiles and cellular morphology from chemical perturbations. ResultsWe introduce PertFlow, a unified computational frame-work that simultaneously predicts treatment gene expression (bulk RNA-seq) and generates cellular morphology (Cell Painting images) from control cellular states, conditioned on drug metadata. Evaluated on paired RNA-seq and imaging data from 3 cell lines and 40 compounds (17,242 samples), PertFlow achieves Pearson correlation of 0.780{+/-}0.264 for transcriptomic prediction and FID of 24.06 for morphological generation. Board-certified pathologists rated generated images with median similarity scores of 7.11-7.89/10. The model successfully recovers known drug mechanisms including microtubule disruption, DNA damage, and MAPK pathway inhibition, with gene enrichment analysis confirming activation of expected biological pathways (EMT, p53, apoptosis). AvailabilityCode and pretrained models will be available at https://github.com/wangmengbo/PertFlow.

bioinformatics↗

bioRxiv2026

p300/CBP-Driven Acetylation Stabilizes PARP7 to Reinforce Negative Feedback on IFN-I Signaling

Immune suppression within the tumor microenvironment (TME) remains a major barrier to effective immunotherapy, yet the molecular mechanisms that constrain type I interferon (IFN-I) signaling in tumor cells are incompletely understood. PARP7 (also known as TIPARP), a mono-ADP-ribosyltransferase, has been reported to suppress IFN-I signaling by modifying TBK1. Here, we identify PARP7 as a previously unrecognized acetylation substrate of the histone acetyltransferases p300 and CBP. We show that p300/CBP-mediated acetylation of PARP7 at lysine 32 markedly enhances its stability, thereby reinforcing repression of the TBK1-IRF3-STAT1 axis and suppressing IFN-{beta}-induced expression of interferon-stimulated genes, including CXCL10. Beyond their established function as transcriptional co-activators that support tumor cell proliferation, our findings uncover a direct role for p300/CBP in shaping innate antitumor immunity. Pharmacological inhibition of p300/CBP catalytic activity by A-485 abolishes PARP7 acetylation, accelerates its proteasomal turnover, and restores IFN-I signaling. In immunocompetent tumor models, this is accompanied by enhanced CD8+ T-cell infiltration and effector function. Together, these findings define a p300/CBP-PARP7 regulatory axis, uncover a non-canonical role for p300/CBP in innate immune suppression, and nominate this pathway as a tractable target for restoring antitumor immunity.

cancer biology↗

bioRxiv2026

Genetic analysis and mapping of adult plant stripe rust resistance loci in CIMMYT wheat 'Kijil under Mexican and Chinese field environments

Stripe rust, caused by Puccinia striiformis f. sp. tritici, can cause severe yield losses in wheat (Triticum aestivum L.) during epidemics. Breeding resistant wheat varieties remains the most cost-effective approach to manage this disease; and the identification of new resistance loci is essential for maintaining genetic diversity. The CIMMYT-derived wheat line Kijil was highly resistant to stripe rust in both Mexican and Chinese environments. A population of 153 F recombinant inbred lines (RILs) was derived from a cross between Kijil and the susceptible parent Apav#1. The population was phenotyped for stripe rust resistance across seven environments in two countries and genotyped using a genotyping-by-sequencing (GBS) platform. Inclusive composite interval mapping (ICIM) was uesd to construct a genetic map and identify significant resistance quantitative trait loci (QTLs) using 5,468 polymorphic markers. Mapping revealed the known resistance loci Yr29, Yr30 and QYr.hzau-3AS, along with two novel loci, QYr.hzau-2BS and QYr.hzau-5DL, across both Chinese and Mexican rust environments. Among these, QYr.hzau-2BS accounted for 11.75% to 19.19% of the phenotypic variance. A corresponding KASP marker, KASP_2BS, was developed to facilitate maker-assisted selection. Based on the mapping interval, four candidate genes underlying this locus were predicted. Further analysis revealed that Yr29 showed significant additive effects with other stripe rust resistance genes/loci, and the combination of Yr29, Yr30, and QYr.hzau-2BS reduced disease severity by up to 67.8%. Our findings suggest that Kijil and RILs carrying Yr29, Yr30, and QYr.hzau-2BS can serve as valuable donors for breeding wheat varieties with improved stripe rust resistance. Author summaryStripe rust is an important disease that seriously threatens the yield and quality of wheat. It is crucial to explore new resistant resources and cultivate durable resistant varieties in the current breeding programme. In this study, we analysed the genetic basis of resistance to stripe rust in the wheat line "Kijil", which has broad-spectrum resistance to stripe rust. Through genetic mapping, we identified five quantitative trait loci for stripe rust, including two new resistance loci. A closely linked KASP marker, KASP_2BS, was developed for the QYr.hzau-2BS, which can be used for rapid and accurate screening of resistant plants in early breeding populations. Meanwhile, within this QTL region, we screened four candidate genes based on expression analysis. In addition, it was found that polymerization of QYr.hzau-2BS with known resistance genes, Yr29 and Yr30, significantly enhanced resistance and reduced disease severity to low levels and near immunity. In conclusion, this study provides new genetic resources, practical molecular markers and effective gene polymerization strategies for breeding wheat for stripe rust resistance. Kijil and the lines containing resistance loci have important breeding utilization value.

plant biology↗

bioRxiv2026

SAGA1 and SAGA2 localize the starch sheath to the pyrenoid in Chlamydomonas reinhardtii

Most algae enhance their CO2 assimilation by concentrating CO2 within the pyrenoid, a biomolecular condensate that contains the CO2-fixing enzyme Rubisco. Many pyrenoids are surrounded by a starch sheath that is thought to slow the escape of CO2 from the pyrenoid, but how the starch sheath is localized to the pyrenoid remains poorly understood. Here, in the leading model alga Chlamydomonas reinhardtii, we find that the protein SAGA2 is necessary for early pyrenoid starch sheath biogenesis and works redundantly with its homolog, SAGA1, to localize the starch sheath to the pyrenoid. SAGA2 and SAGA1 were enriched in different regions of the pyrenoid-starch sheath interface, suggesting that they play complementary roles. Both saga2 and saga1 mutants showed defects in starch sheath coverage early during pyrenoid formation that were improved at a later timepoint. Strikingly, a saga1;saga2 double mutant did not have a starch sheath around the pyrenoid and showed decreased overall starch content. SAGA1 and SAGA2 starch-binding domains bound to starch, the starch mimic molecule {beta}-cyclodextrin, and the starch precursor molecule maltoheptaose, suggesting a role for SAGA1 and SAGA2 in starch granule initiation. We propose a model where SAGA1 and SAGA2 each locally prime starch sheath initiation in a distinct region of the pyrenoid surface by enriching starch precursor molecules around the pyrenoid. These findings advance the understanding of algal starch sheath biogenesis and provide insights into the associations between biomolecular condensates and other cellular structures. Significance StatementEukaryotic algae enhance their carbon assimilation using an organelle called the pyrenoid, where concentrated CO2 is supplied to the CO2-fixing enzyme Rubisco. In many algae, a starch sheath surrounding the pyrenoid is thought to enhance CO2 fixation, but how starch is localized to pyrenoids is unknown. Here, we show that two proteins, SAGA1 and SAGA2, each bind to starch precursor molecules and redundantly localize starch to the pyrenoid in the alga Chlamydomonas reinhardtii. Our results suggest that SAGA1 and SAGA2 promote starch sheath initiation at the pyrenoid, rather than merely tethering starch, as previously thought. This work advances the understanding of the proteins and molecular mechanisms involved in pyrenoid starch sheath biogenesis and lays the foundations for their further study.

molecular biology↗

bioRxiv2026

Cancer-associated synonymous mutations reveal stress signal-dependent mRNA folding that selectively modulates protein function.

Recent technical advances have facilitated studies on changes in mRNA structures in response to signaling pathways. However, if mRNA structures can affect the function of the encoded protein remains poorly understood. In-cell RNA structural probing (SHAPE-MaP) demonstrates how two cancer-associated synonymous mutations (CASMs) at proline codon 34 (c.102 C>A and c.102 C>G) prevent DNA damage-induced TP53 mRNA folding, whereas the non-cancer-associated c.102C>U mutation does not. Transcript and chromatin immunoprecipitation (ChIP) analysis reveal that p53 expressed from CASM34 has reduced promoter binding and reduced induction of p53 downstream target genes PUMA and 14-3-3-{sigma}, but not p21CDKN1A. Transcriptome analysis reveals a CASM34-mediated global attenuation of DNA damage-responsive gene expression. Together, the results demonstrate that CASM34 interferes with signal-induced p53 mRNA folding during DNA damage, leading to selective modulation of protein activity. More broadly, our findings highlight a general concept by which cancer-associated synonymous mutations target signal-induced mRNA structures that influence the encoded protein. Significance statementRecent studies have revealed that synonymous mutations can target RNA metabolism and are associated with numerous diseases; however, how these mutations affect the function of the encoded protein remain unclear. Using in-cell RNA structural probing, we show that two cancer-associated synonymous mutations in TP53 disrupt DNA damage-induced folding of p53 mRNA. These mutations selectively impair p53 promoter binding and transcriptional activation of specific downstream target genes. Our results demonstrate that disruption of signal-induced mRNA folding by synonymous mutations can modulate the activity of the encoded protein. More broadly, this work identifies signal-responsive mRNA structures as functional targets of cancer-associated synonymous mutations.

molecular biology↗

bioRxiv2026

A small molecule PTER-selective inhibitor reduces food intake and body weight

PTER (phosphotriesterase-related) is an amidohydrolase that mediates catabolism of the anorexigenic taurine metabolite N-acetyltaurine. However, the structural basis of PTER ligand binding and catalysis remain unknown, limiting our ability to harness this pathway therapeutically. Here we solve crystal structures of a eukaryotic PTER in apo and product-bound forms. These structures uncover an unexpected pocket homology between PTER and histone deacetylase (HDAC) enzymes. We exploit this similarity to engineer a first-in-class substrate-competitive PTER inhibitor called PTERi with nanomolar potency and >100-fold selectivity for PTER over HDACs in vitro. Administration of PTERi to diet-induced obese mice reduces feeding, enhances GLP1-RA (glucagon like peptide 1 receptor agonist)-induced weight loss, and prevents weight regain after GLP1-RA discontinuation. The structure of PTER connects histone and metabolite deacetylation into a parallel conceptual framework and enables proof-of-concept data for pharmacological inhibition of PTER in obesity.

biochemistry↗

bioRxiv2026

DNA-templated Chiral Metamaterial Array as Information Bits

The growth of digital information demands physically secure information bits that combine intrinsic randomness with multi-layered optical readout. Digital metamaterials with physical unclonable function characteristics are promising, but nearly all examples operate at microwave or terahertz frequencies. Extending digital metamaterials to the visible range requires a physical property that allows binary (0/1) encoding via nanoscale geometry and is orthogonal to intensity-or color-based imaging. Chiral plasmonic metamaterials satisfy these criteria perfectly: their broken mirror symmetry yields chirality polarity in the visible spectrum whose sign and magnitude directly encode bit values ("0", spin-up; "1", spin-down), while remaining invisible under linear polarization. Here, we realize visible-range digital metamaterials by programming self-assembled DNA origami templates with asymmetrically placed gold nanorods to create discrete 3D chiral metamolecules. Using on-surface DNA origami assembly and the Silica Microsphere-Assisted Patterning by Liquid Elimination (SiMPLE) method, we fabricate large-scale bit arrays on optical active glass substrate in solution with [~]1.33 m lattice spacing, [~]86% site occupancy, and [~]12-month stability after silicification. These bottom-up fabricated digital metamaterial array exhibits two independent security layers: (1) a macroscopic spatial pattern only visible by dark-field microscopy, and (2) hidden binary information stored in the single-particle chiroptical response, read out by position-resolved circular dichroism spectroscopy. Quantitative analysis confirms reliable bit encoding and high optical randomness arising from slight structural variations. By leveraging the polarity of plasmonic chirality to translate molecular-scale handedness into robust visible-range digital signals, this work establishes a scalable nanophotonic platform for secure optical information storage, authentication, and encryption.

bioengineering↗

bioRxiv2026

Phage Portal Proteins Suppress Bacterial Stringent Response to Promote Infection

Bacteria restrict viral replication not only through dedicated defense systems but also by entering global physiological states that limit cellular resources. How phages counter such host-imposed physiological barriers remains poorly understood. The stringent response (SR), mediated by the alarmones ppGpp and pppGpp, induces a growth-restrictive state that can act as a barrier to bacteriophage infection. Here we show that elevated alarmone levels delay T7-mediated host lysis, whereas (p)ppGpp-deficient cells are hypersensitive to infection, establishing alarmone signaling as a physiological constraint on phage replication. A systematic functional screen identifies the essential capsid portal protein Gp8 as a viral factor genetically linked to SR-dependent host physiology. Gp8 directly binds the alarmone synthetases RelA and SpoT, selectively inhibiting their synthetase activities in vitro and suppressing alarmone accumulation in vivo. Phages carrying interaction-defective portal mutations exhibit impaired replication, delayed lysis, and sustained (p)ppGpp elevation during infection, defects that are alleviated in SR-deficient hosts. Structural and mutational analyses reveal electrostatically mediated interfaces required for targeting RelA and SpoT. Portal proteins from diverse coliphages share similar structural features, interact with stringent-response enzymes, and display SR-linked phenotypes, indicating a broadly conserved viral strategy. Together, these findings identify phage portal proteins as a previously unrecognized class of viral counter-defense factors that target a central bacterial stress signaling pathway, revealing that essential structural virion components can moonlight as regulators of host stress physiology.

microbiology↗

bioRxiv2026

AMPBAN: A Deep Learning Framework Integrating Protein Sequence and Structural Features for Antimicrobial Peptide Prediction

The escalating crisis of antimicrobial resistance poses a devastating and immediate threat to human life. Antimicrobial peptides (AMPs) are a promising antibiotic substitute to combat antimicrobial resistance. Compared with the traditional wet-lab screening approaches, computational models have largely improved the efficiency of predicting antimicrobial peptide. However, most computational models overlook or underutilize the evolution and structural information of peptides, which is crucial for understanding the peptide functions. Here, we proposed a sophisticated deep learning model to predict AMPs, Antimicrobial Peptide Bilinear Attention Network (AMPBAN), which incorporates peptide evolution features from ESM3 protein language model, structure features from ESMFold predicted with equivariant graph neural network (EGNN), and the joint information from sequence and structure learned via Bilinear Attention Network. AMPBAN consistently demonstrated superior accuracy and generalization compared to nine state-of-the-art AMP prediction models across multiple independent benchmarks. Furthermore, an ablation study confirms that our multimodal fusion strategy significantly refines the integration of sequence and structural signals, yielding superior predictive balance over single-modality models. This framework provides a robust tool for the accelerated discovery of novel AMPs and the advancement of next-generation antimicrobial drug development. The datasets, source code and models are available at https://github.com/baiwenhuim/ampban.

bioinformatics↗

bioRxiv2026

Iron Deficiency Drives Sarcopenia in the Elderly: HIF-1α-Mediated Fibro-Adipogenic Progenitor Differentiation Induces Fat Infiltration and Impairs Muscle Function

Abnormal iron metabolism is closely linked to sarcopenia; however, the specific iron metabolism features of fat-infiltrating sarcopenia remain poorly understood. Proteomic sequencing revealed that in skeletal muscle with severe fat infiltration, the expression of iron utilization-related proteins was significantly downregulated, whereas that of iron uptake and storage proteins was markedly upregulated, thereby presenting an abnormal iron deficiency (ID) phenotype. Nevertheless, the mechanism by which ID drives intramuscular fat infiltration has not been fully elucidated. Using clinical samples, aged murine ID models, and in vitro cell assays, this study is the first to demonstrate that ID stabilizes hypoxia-inducible factor-1 (HIF-1), promoting aberrant adipogenic differentiation of fibro-adipogenic progenitors (FAPs), disrupting the homeostatic balance between satellite cells (SCs) and FAPs, exacerbating skeletal muscle fat infiltration, and impairing muscle repair capacity. Notably, treatment with the HIF-1 inhibitor PX478 reversed these pathological alterations and improved muscle function. Collectively, our findings identify the ID-HIF-1-FAPs axis as a key driver of intramuscular fat infiltration, offering a novel therapeutic target for the clinical management of fat-infiltrating sarcopenia.

cell biology↗