Biology subjects

Wang, M.

Publications and source records attributed to Wang, M..

At least 199 records · Page 11Linked to original sources

bioRxiv2025

BioLab: End-to-End Autonomous Life Sciences Research with Multi-Agents System Integrating Biological Foundation Models

Scientific discovery in the life sciences remains hindered by fragmented workflows, narrow-scope computational models, and inefficient links between in silico prediction and wet-lab validation. We present BioLab, a multi-agent system that integrates domain-specialized foundation models to automate end-to-end biological research. BioLab comprises eight collaborating agents, including a Planner, Reasoner, and Critic, orchestrated through a Memory Agent that enables iterative refinement via retrieval-augmented generation and a suite of 219 computational xBio-Tools spanning five biological scales (DNA, RNA, protein, cell, and chemical). These tools are built on the xTrimo Universe, a collection of 104 models derived from six foundation models (xTrimoChem, Protein, RNA, DNA, Cell, and Text), the majority of which achieve state-of-the-art (91-100% SOTA ratios) on domain benchmarks. Across standard reasoning tasks (PubMedQA, MMLU-Pro/Biology, GPQA-diamond), BioLab consistently outperformed leading large language models, including GPT-4o, Gemini-2.5, and DeepSeek-R1. Beyond benchmarks, BioLab autonomously executed a fully computational pipeline for de novo macrophage-targeting antibody design, progressing from target mining to multi-objective antibody optimization, where molecular dynamics simulations revealed structural mechanisms underlying enhanced affinity of optimized variants. Closing the computational-experimental loop, BioLab designed optimized antibodies (Pem-MOO-1, Pem-MOO-2) that achieved IC50 values of 0.01-0.016 nM, markedly surpassing the parental Pembrolizumab (0.027 nM) for PD-1. Functional assays confirmed enhanced pathway blockade and improved multi-parameter performance profiles. Together, these results establish BioLab as a generalizable framework for AI-native scientific discovery, demonstrating how multi-agent systems coupled with foundation models can autonomously generate, execute, and experimentally validate novel biological hypotheses.

bioinformatics↗

bioRxiv2025

Peptide-based ligand antagonists block a Vibrio cholerae adhesin

Vibrio cholerae, the causative agent of cholera, uses surface proteins such as the repeats-in-toxin (RTX) adhesin FrhA to colonize hosts and initiate infection. Blocking bacterial adhesion represents a promising therapeutic strategy to treat infections without promoting drug resistance. FrhA contains a peptide-binding domain (PBD) that is key for hemagglutination, human epithelial cell binding, and V. cholerae biofilm formation. Previous studies identified a lead pentapeptide ligand with the sequence Ala-Gly-Tyr-Thr-Asp (AGYTD) that blocks V. cholerae colonization of the mouse small intestine at high micromolar concentrations. A structure-guided approach has now identified a minimal D-amino acid-containing tripeptide motif with higher affinity for the FrhA-PBD and predicted metabolic stability. Our results contribute to the development of anti-adhesion strategies to combat infections.

microbiology↗

bioRxiv2025

Identification of stripe rust adult plant resistance genes in the hard winter wheat cultivar Bakers Ann

Stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is among the most destructive wheat diseases. Identifying resistance genes is crucial for the development of resistant cultivars. "Bakers Ann", a hard winter wheat cultivar developed by Oklahoma State University, has shown durable adult plant resistance to stripe rust. To dissect the genetic basis underlying stripe rust resistance in Bakers Ann, 125 doubled haploid lines, derived from the cross OK12D22004-016 x Bakers Ann, were evaluated at the adult plant stage in the greenhouse and in field environments in Oklahoma, Kansas, and Washington. This population was genotyped using genotyping-by-sequencing, which produced 7,268 single-nucleotide polymorphisms for genetic mapping. Quantitative trait loci (QTL) analysis identified six loci, four from Bakers Ann on chromosomes 2DL, 4BS, 4BL, and 7BL, and two from OK12D22004-016 on chromosomes 2AS and 2AL. Although OK12D22004-016 is susceptible in the US Great Plains, it was found to carry QYr.osu-2AS, which was linked to Yr17 on the 2NvS translocation and explained up to 30% of the phenotypic variation, but was effective in a single location in Washington. Two major QTL were identified in Bakers Ann, QYr.osu-2DL on chromosome 2DL that explained up to 57% of the phenotypic variation and mapped close to Yr54, and QYr.osu-4BL on chromosome 4BL that explained up to 15% of the phenotypic variation and mapped close to Yr62. Resistance in Bakers Ann resulted from additive effects of the four QTL. Two kompetitive allele-specific PCR markers were developed for QYr.osu-2DL to facilitate marker-assisted selection for stripe rust resistance.

genetics↗

bioRxiv2025

A comprehensive RNA virome from molluscan transcriptomes reveals extensive diversity and modular genome evolution

Mollusca is the second-largest animal phylum and an important marine food resource for humans. While DNA viruses that threaten molluscan aquaculture have received much attention, molluscan RNA viromes remain poorly explored. Here, based on 223 molluscan metatranscriptomes covering eight animal classes, we identified 80 species-level RNA viruses spanning three viral phyla and nine viral families. Phylogenetic result combined structural modeling found a major increase in the number of Pisuviricota-related lineages. Extensive modular evolution in viral genomes was observed including gene rearrangements and co-evolution of capsid and RdRP genes. Host prediction linked 78% of the RNA viruses to a diverse range of eukaryotes, including vertebrates, invertebrates, and algae. These findings expand the known diversity of RNA viruses in molluscs and shed light on their phylogenetic relationships, highlighting molluscs as valuable models for RNA virus evolution. Graphical abstractRNA virus diversity within 223 molluscan transcriptome samples were analyzed based on RdRP identification. Viruses belonging to the phylum Pisoviricota were the most abundant; co-evolution between RdRP and capsid proteins of molluscan Picornavirales was revealed, and a complex virus-host interaction network was constructed. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=153 SRC="FIGDIR/small/673676v1_ufig1.gif" ALT="Figure 1"> View larger version (44K): org.highwire.dtl.DTLVardef@1ce8fc0org.highwire.dtl.DTLVardef@3f9068org.highwire.dtl.DTLVardef@121cc2forg.highwire.dtl.DTLVardef@1e431a0_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIComprehensive metatranscriptomic analysis found extensive RNA viral diversity in molluscs. C_LIO_LIPisuviricota dominate the molluscan RNA virome, substantially broadening known viral lineages. C_LIO_LIModular genome evolution involving gene rearrangements and conserved structural features characterize molluscan RNA viruses. C_LIO_LIMolluscs potentially serve as major ecological reservoirs for diverse RNA viruses infecting various eukaryotic hosts. C_LI

bioinformatics↗

bioRxiv2025

Deciphering Cell Fate and Clonal Dynamics via Integrative Single-Cell Lineage Modeling

Through natural or synthetic lineage barcodes, single-cell technologies now enable the joint measurement of molecular states and clonal identities, providing an unprecedented opportunity to study cell fate and dynamics. Yet, most computational methods for inferring cell development and differentiation rely exclusively on transcriptional similarity, overlooking the lineage information encoded by lineage barcodes. This limitation is exemplified by T cells, where subtle transcriptional differences mark divergent fates with distinct biological activity. Single-cell RNA and matched TCR sequencing is now ubiquitous in the analysis of clinical samples, where the TCR sequence provides an endogenous clonal barcode and could reveal clonal T cell responses. We present Clonotrace, a computational framework that jointly models gene expression and clonotype information to infer cell state transitions and fate biases with higher fidelity. While motivated by challenges in analyzing T cell populations, especially in the tumor microenvironment and immunotherapy settings, Clonotrace is broadly applicable to any lineage-barcoded single-cell dataset. Across diverse systems including T cells, hematopoietic differentiation, and cancer therapy resistance models, Clonotrace reveals differentiation hierarchies, distinguishes unipotent from multipotent states, and identifies candidate fate-determining genes driving lineage commitment.

bioinformatics↗

bioRxiv2025

Microbial Named Entity Recognition and Normalisation for AI-assisted Literature Review and Meta-Analysis

MotivationManual curation of biomedical literature is slow and error-prone and while large language models (LLMs) trained on general texts have shown to be useful for text summarisation, these methods lack the domain-specific expertise required to perform this task accurately. Here we describe the creation of the first microbiome-specific text corpus, use this to train deep learning algorithms for named-entity recognition (NER) and normalisation (NEN), and demonstrate their use to meta-analyse microbiome literature. MethodsWe developed an automated pipeline to annotate all mentions of bacteria, archaea, and fungi in 1,410 full-text microbiome articles. We manually annotated (gold-standard) a separate test set of 288 documents. We trained different transformer-based language models for microbiome recognition and normalisation to taxonomic identifiers and evaluate their performance using the precision, recall, F1-score, and accuracy on the test set. The best models were used to automatically annotate all available Open Access, full-text microbiome articles (n=6,927) and identify taxa that are significantly overrepresented across 14 domains. ResultsThe training and validation set contained a total of 90,150 annotations (both long form and abbreviations). Using the gold-standard test set, with an inter-annotator agreement rate of 99.52% for NER and 88.31% for NEN, the trained models were evaluated and our fine-tuned BioBERT model achieved an F1-score of 96% for NER surpassing a rule- and dictionary-based annotation pipeline (94%). For NEN the accuracy obtained by the deep learning models greatly surpassed that of the pipeline (91% vs 69%). Evaluated across the entire available literature, our models annotate an entire full-text document in only 7 seconds. ConclusionOur algorithms have near perfect precision and greatly speed up the process of annotating microbes in full-text articles. We demonstrated the capabilities of these methods by analysing the entire available literature and describe the taxa associated with each of the domains in our meta-analysis, and exemplify how these methods can be integrated into literature review workflows improving both the speed and accuracy of results. AvailabilityAll codes and data for automatic annotation, model training, and generation of taxonomic trees visualising the data will be made available following peer review with instructions on how to deploy the model on new texts from https://github.com/omicsNLP/microbELP.

bioinformatics↗

bioRxiv2025

Partitioning of nighttime transpiration and stem water refilling using VPD and dendrometer data: insights into baselining and nighttime sap flux interpretation

The baseline for thermal dissipation probes is a reference for simultaneous cessation of both nighttime transpiration and refilling of trees internal water stores. Finding this moment is challenging as the baseline is dynamically changing, and the complete sap flux cessation may not even occur. We proposed a novel framework using VPD and dendrometer measurements to partition transpiration and refilling from nighttime sap flux density, and introduced a night-by-night baselining scheme that does not rely on identifying the zero-flow moment. One conventional baselining method using the maximum probe voltage (or temperature) difference substantially underestimated sap flux density by 204.56% for nighttime, 28.30% for daytime, and 45.20% for full-day flux in P. sylvestris, and by 344.77%, 45.68%, and 54.76%, respectively, in A. glutinosa. Baseline calibrations using commonly available variables are thus provided. Evaluation through gas exchange and evapotranspiration data confirmed the accuracy of partitioning between nighttime transpiration and refilling, demonstrating that nighttime transpiration can drive measurable nighttime evapotranspiration responses. Nightly variations in water allocation within stem tissue add complexity to stem swelling. Seasonally, sap flux density showed an initial increase in both transpiration and refilling, followed by a decline, with transpiration dominating until mid-summer before refilling gradually took precedence. HighlightsO_LIPartitioned nighttime transpiration & stem refilling using VPD & dendrometers. C_LIO_LITracked seasonal shifts in transpiration & refilling ratios to nighttime sap flux. C_LIO_LIIndependently evaluated model with eddy covariance & leaf gas exchange measurements. C_LIO_LIProposed a new high-frequency baselining strategy for thermal dissipation probes. C_LIO_LIRevealed & corrected errors in sap flux density baselining methods. C_LI

plant biology↗

bioRxiv2025

An Objective Detection Method for Tinnitus using Memristors

Approximately 10-20% of the world population experiences tinnitus, which is increased with age and comorbid with frustration, annoyance, anxiety, depression, cognitive dysfunction, insomnia and stress. The underlying mechanisms of tinnitus remain largely unknown and there are no quantifiable biomarkers, making this prevalent disorder a global burden. There are no FDA-approved therapeutics and the current assessment mainly relies on self-reports (e.g., questionnaires), which underscore the need for diagnostic tools for tinnitus. Memristors, nanoscale electronic components mimicking synaptic plasticity, can monitor and control neuronal activity, hold transformative potential for detecting and characterizing tinnitus objectively. Here, we harvested the blood from human and mice with tinnitus, and engineered memristors via magnetron sputtering and made the functional layer with serum. Its core mechanism lies in regulating the reversible formation and rupture of conductive filaments which is formed by utilizing abnormally elevated glutamate concentrations in serum assisted Ag+ ions. This serum-based memristor achieves unprecedented discrimination between tinnitus and non-tinnitus individuals. Furthermore, this platform provides mechanistic insights into tinnitus pathogenesis, with potential implications in other neurological disorders.

pathology↗

bioRxiv2025

Engineering Antibody-Drug Conjugates targeting an Adhesion GPCR, CD97

Adhesion G protein-coupled receptors (aGPCRs) are key cell-adhesion molecules involved in many cellular functions and contribute to human diseases, including cancer. aGPCRs are characterized by large extracellular regions that could serve as readily accessible antigens. However, the potential of aGPCRs as targets for biologic therapeutics has not been extensively explored. CD97, also known as ADGRE5, is an aGPCR that is upregulated in various cancer types, including acute myeloid leukemia (AML) and glioblastoma (GBM), and their respective cancer stem cells. Here, we developed antibody-drug conjugates (ADCs) targeting CD97 and assessed their efficacy against AML and GBM cells. We generated a panel of synthetic human antibodies targeting distinct epitopes of CD97, from which we identified an antibody that was efficiently internalized. This antibody binds to all isoforms of human CD97 but not to its close homolog, EMR2. Structure determination by single-particle cryo-electron microscopy revealed that this antibody targets the CD97 GPCR autoproteolysis-inducing (GAIN) domain, whose presence is conserved in aGPCRs, through an unconventional binding mode where it extensively utilizes the light chain framework for antigen recognition. Screening of conjugation methods and payloads resulted in a stable ADC that effectively killed AML and GBM cell lines, as well as patient-derived GBM stem cells, with minimal cytotoxicity against peripheral blood mononuclear cells from healthy donors. Our study demonstrates the therapeutic potential of targeting CD97, as well as the aGPCR GAIN domain in general, and broadens our mechanistic understanding of antibody-antigen interactions.

biochemistry↗

bioRxiv2025

Precise Extraction of Neural Motifs Reveals Multiscale, Parallel Encoding Schemes in Auditory Cortex

Neural population activity is often stereotyped into recurring activity patterns, i.e., neural motifs, which can be seen as the fundamental building blocks in sensory processing and cognition. In this work we study the codes carried by such neural motifs in primary auditory cortex A1, and analyze how they build on and complement traditional views of single-unit coding. In particular, we study, using two-photon calcium imaging (CI), how activity in A1 differentially represents both sensory stimuli and task and behavioral variables in each of the parallel single neuron and population motif scales. While CI enables the study of neural motifs by capturing the activity of large neural populations, identifying motifs in CI is hampered by the temporal imprecision of current motif-detection algorithms when applied to CI data. We thus developed a new algorithm for motif detection, which enabled us to identify widespread stimulus-encoding motifs as well as a small number of motifs jointly encoding stimulus and choice in Layer 2/3 of A1. These motifs consist of small groups of neurons that are neither clustered nor regularly ordered in space. Interestingly, active neurons within task-encoding motifs exhibit mixed encoding properties inconsistent with the motifs they participate in. Together, these results demonstrate how single unit activity and neural motifs in A1 L2/3 provide different levels of coding granularity containing different information in parallel within the greater neural population. Generally, our results indicate that downstream populations, by selecting which scale of a population drives them, can be selective in the information collected for later cognitive processing.

neuroscience↗

bioRxiv2025

Viruses associated with Nitrososphaerota archaea: diversity, evolution and horizontal gene transfers

Nitrososphaeria, an archaeal class containing all ammonium-oxidizing archaea (AOA) and their closely related clades (e.g. Ca. Caldarchaeales), is a widely distributed archaeal lineage across diverse ecosystems. Due to the difficulty of isolation and cultivation, the viral diversity and evolution associated with non-marine AOA lineages and Ca. Caldarchaeales remain largely uncharacterized. Here, based on globally sampled genomes and metagenomes, we assembled a genomic catalogue associated with AOA, including 283 species-level viruses and 415 species-level mobile genetic elements (MGEs). Notably, 259 previously unrecognized species-level viruses or MGEs associated with Ca. Caldarchaeales were identified. Across the two major Nitrososphaeria lineages, we uncovered an expansion of archaeal viral families: 11 families associated with AOA, including seven new head-tailed lineages, one of which comprises a mega archaeal virus with a genome size exceeding 220 kb; ten families associated with Ca. Caldarchaeales, including four novel head-tailed groups and two previously unknown rod-shaped family (Demeterviridae and Gaiaviridae). Phylogenetic inference showed that rod-shaped viruses associated with Ca. Caldarchaeales and methanogens are phylogenetically intertwined and likely share a common ancestor with Hoswirudivirus, a Sulfolobales-associated rod-shaped virus characterized by two copies of the coat gene. Through structural modelling, a rare fusion of the coat protein in Demeterviridae was observed, in which a single open reading frame within a gene island adjacent to a glycosyl transferase appears capable of forming a complete capsomer. This provides new insights into the origin and evolution of Adnaviria viral realm. Functional genomics analyses further revealed that these novel rod-shaped viruses encode virus-specific rare quinone biosynthesis genes and arsenic response transcriptional regulator genes, providing potential clues to their mechanisms of adaptation to hyperthermal environments. Collectively, our work uncovers the hidden genomic, evolution and functional diversity of Nitrososphaeria viromes for the first time and closes a long-standing gap in archaeal virome research. HighlightsO_LIFirst genomic catalogue of viruses and MGEs linked to Nitrososphaeria. C_LIO_LIDiscovery of diverse novel archaeal viral families. C_LIO_LINew insights into rod-shaped virus evolution and archaeal virus-host adaptation. C_LI

bioinformatics↗

bioRxiv2025

CREG1 restricts ALV-J replication via the mitochondrial dysfunction-driven activation of innate immunity and apoptosis

Despite the implementation of purification strategies to partially limit J subgroup avian leukosis virus (ALV-J) infection, the involvement of host factors in the underlying infection mechanism remains largely undefined. Here, we identify cellular repressor of E1A-stimulated genes 1 (CREG1) as a key regulator of mitochondrial function and a critical immune-related gene involved in ALV-J infection. The objective of this study was to explore the effects and underlying mechanisms of CREG1 in the context of ALV-J infection. Overexpression of CREG1 upregulates the expression of type I interferon (I-IFN) and certain interferon-stimulated genes (ISGs), thereby suppressing viral replication. Mechanistically, overexpression of CREG1 induces mitochondrial dysfunction, characterized by a decrease in mitochondrial membrane potential ({Delta}{psi}m), reduced adenosine triphosphate (ATP) production and respiratory chain activity, enhanced mitophagy, and increased release of mitochondrial DNA (mtDNA), which in turn triggers the activation of innate immune responses. Mitochondrial dysfunction further leads to the cytosolic release of cytochrome c and an increase in reactive oxygen species (ROS) levels, thereby triggering a robust apoptotic response. Moreover, the regulation of mitochondrial function by CREG1 depends on its interaction with the mitochondrial chaperone protein heat shock protein 1 (HSPD1), and their co-expression synergistically amplifies the antiviral response. In this study, we identify CREG1 as a potent antiviral gene and underscore the pivotal roles of mitochondria-mediated innate immunity and apoptosis during ALV-J infection. Author summaryDespite efforts to limit J subgroup avian leukosis virus (ALV-J) infection, the role of host factors in the infection process is still not fully understood. In this study, we highlight the cellular repressor of E1A-stimulated genes 1 (CREG1) as a crucial regulator of mitochondrial function and immune responses during ALV-J infection. We show that CREG1 overexpression enhances type I interferon (I-IFN) and interferon-stimulated genes (ISGs) expression, inhibiting viral replication. This is achieved through mitochondrial dysfunction, including reduced mitochondrial membrane potential, lower ATP production, and increased mitophagy and mitochondrial DNA release, which activate innate immunity. CREG1-induced mitochondrial dysfunction also triggers apoptosis by releasing cytochrome c and increasing reactive oxygen species (ROS). Furthermore, CREG1 interacts with heat shock protein 1 (HSPD1) to amplify its antiviral effects. Our findings establish CREG1 as a key antiviral gene and emphasize the importance of mitochondrial-mediated immune responses and apoptosis in ALV-J infection.

immunology↗

bioRxiv2025

IsoPairFinder: A tool for biochemical pathway discovery using stable isotope tracing metabolomics

SummaryThe functional annotation of microbial genes lags far behind genome sequencing, leaving critical gaps in our knowledge of metabolic pathways. While integrating genetic manipulation with stable isotope tracing (SIT) metabolomics holds promise for pathway discovery, existing tools lack specialized capabilities for gene perturbation experiments. To address this need, we developed IsoPairFinder, a computational tool that identifies pathway intermediates by analyzing paired unlabeled (12C) and isotope-labeled (13C) metabolomics data from gene-edited microbes. By prioritizing substrate-specific feature pairs, IsoPairFinder efficiently prioritizes biologically relevant intermediates. Implemented as an open-source R package and integrated into the GNPS2 ecosystem, IsoPairFinder provides an accessible platform for the research community to accelerate novel pathway discovery and validation. Availability and implementationIsoPairFinder is freely available as an open-source R package (https://github.com/DoddLab/IsoPairFinder) and as a web-based workflow in the GNPS2 ecosystem (https://gnps2.org/workflowinput?workflowname=isopairfinder_nextflow_workflow). Comprehensive tutorials are provided at https://doddlab.github.io/IsoPairFinder_Tutorials. These implementations support both command-line and graphical analysis workflows. Contactddodd2@stanford.edu Supplementary informationSupplementary information is available at Bioinformatics online.

bioinformatics↗

bioRxiv2025

Epilepsy-Associated SCN2A-L1342P Mutation Drives Network Hyperexcitability and Widespread Transcriptomic Changes in Human Cortical Organoids

ObjectiveSCN2A pathogenic mutations, such as the recurrent heterozygous Nav1.2-L1342P, are monogenic causes of epilepsy. In this human-induced pluripotent stem cell model system, we aim to investigate the molecular and cellular mechanisms underlying the SCN2A-L1342P-associated pathology. MethodsUsing a human male iPSC reference line (KOLF) carrying the Nav1.2-L1342P mutation, we generated 3D cortical organoids for functional studies. Patch clamp and multi-electrode array (MEA) recordings, immunocytochemistry, and RNA sequencing were used to characterize the disease phenotypes. ResultsNav1.2-L1342P organoid neurons displayed increased intrinsic excitability, and amplified excitatory post-synaptic currents, which are consistent with an increase in excitatory synapse formation revealed by PSD95/SYN1 immunostaining. Moreover, elevated network firing activity, as demonstrated by MEA, indicates a pronounced network hyperexcitability. Transcriptomic profiling of organoids carrying the Nav1.2-L1342P mutation further revealed significant alterations in synaptic, glutamatergic, and developmental pathways. SignificanceOur findings demonstrate that the Nav1.2-L1342P mutation drives a multifaceted disease phenotype, including network hyperexcitability and disruption of pathways related to neuronal and synaptic functions. These results advance our understanding of SCN2A-related Developmental and Epileptic Encephalopathy (DEE), laying a foundation for personalized interventions. Key PointsO_LIKey Point 1: Human neurons carrying epilepsy-causing SCN2A-L1342P display intrinsic hyperexcitability in cortical organoids. C_LIO_LIKey Point 2: Synaptic transmission is enhanced in organoids carrying the SCN2A-L1324P mutation, consistent with the presence of elevated excitatory synapses, which leads to increased network hyperexcitability. C_LIO_LIKey Point 3: Transcriptomics analysis reveals that the SCN2A-L1342P mutation causes significant differential changes in forebrain developmental genes, synaptic and glutamatergic signaling pathways, and enhances cellular senescence and apoptosis. C_LI

neuroscience↗

bioRxiv2025

Helicobacter pylori genome aggregation database reveals complex evolutionary forces shaping its genomic landscape and clinical impact in East Asia

We assembled HPgnomAD, a comprehensive global genome aggregation database of Helicobacter pylori (H. pylori) featuring 7,544 high-quality genomes from 278 populations, along with the first species-wide haplotype reference panel for genotype imputation. The panel provides high accuracy across diverse H. pylori populations, including those from low-coverage data, and is openly accessible with integrated analysis tools (https://www.hpgnomad.top/). Variant discovery revealed 1.82 million SNPs and 0.65 million InDels, with African strains exhibiting the most tremendous diversity and East Asian strains showing unexpectedly high novel variation. Fine-scale analysis of the hpEAsia lineage revealed six new sublineages shaped by altitude-, latitude-related divergence and region-specific admixture. Phylogenomic dating revealed two divergence waves in East Asia, paralleling Upper Paleolithic settlement and Neolithic human expansions, with highland-lowland separation at [~]18.8 kya, followed by the formation of complex geography-related population substructures. Genome-wide scans revealed adaptive loci related to metal acquisition, nitrogen metabolism, surface adhesion, and membrane transport, including altitude-associated highly differentiated variants linked to antibiotic resistance and gastric disease. By integrating genomic, evolutionary, environmental, and clinical data, HPgnomAD offers a framework for understanding H. pylori evolution, host-pathogen adaptation, and precision medicine worldwide.

evolutionary biology↗

bioRxiv2025

Concurrent inhibition of ICMT and RAF/MEK suppresses RAC1P29S-driven MAPKi resistance in BRAFV600E melanoma by regulating TAZ activity

The RAC1 GTPase hotspot mutation P29S (RAC1P29S) is among the top driver oncogenes of cutaneous melanoma, which is known to develop resistance to MAPK pathway inhibitors including those targeting BRAF and MEK. Isoprenylcysteine carboxyl-methyltransferase (ICMT) is the enzyme catalyzing the last step of post-translational prenylation of RAC1, which is among its substrates. We demonstrate that RAC1P29S/C189S, which lacks C-terminal prenylation site, has lost the ability to induce resistance toward vemurafenib or trametinib in BRAFV600E melanoma cells. Furthermore, the combination of vemurafenib with cysmethynil, a proof-of-concept ICMT inhibitor, showed efficacy in combating RAC1P29S-driven resistance of BRAFV600E melanoma cells in both in vitro and in vivo settings. Concurrent treatment with cysmethynil and the MAPK pathway inhibitors efficiently inhibited proliferation and tumor formation of RAC1P29S cells that are resistant to MAPK pathway inhibitor alone. Mechanistically, we found that the combined treatment impaired the nuclear translocation of TAZ, whose transcriptional activity is shown to account for MAPKi resistance in RAC1P29S melanoma. We further validated the role of TAZ in RAC1P29S-driven resistance by demonstrating that introducing a constitutively-active TAZ mutant enhanced the MAPKi resistance in native cells, phenocopying the effect of RAC1P29S. The novel application of MAPKi and cysmethynil combination in RAC1P29S-driven MAPKi-resistant melanoma cells extends the potential utility of ICMT inhibitors, and also provides a new mechanism for targeting ICMT in cancer.

cancer biology↗

bioRxiv2025

DeepHalo: Deep Learning-Powered Exploration of Halogenated Metabolites Uncovering Antibacterial Depsipeptides

In the omics era, confident high-throughput analytical tools are crucial for the efficient identification of metabolites. Here, we present DeepHalo, a deep learning-integrated and hierarchically optimized workflow designed for high-throughput exploration of halogenated metabolites from high-resolution mass spectrometry-based metabolomics. DeepHalo leverages deep learning models combined with a comprehensive scoring to enhance the reliability of halogen predictions. It integrates PyOpenMS for fast isotope pattern detection and incorporates a halogen-based dereplication algorithm with GNPS molecular networking to efficiently exploit and annotate halogenates from complex biological matrices. To validate its performance, DeepHalo was applied to explore halogenated metabolites from 1,296 microbial culture crude, leading to the discovery of six families of structurally diverse halogenated molecules. This included a new class of cyclic depsipeptides, aglomycins A-E, featuring rare 3-chloroanthranilic acid and/or epoxyvaline blocks. Additionally, a plausible biosynthetic pathway of aglomycins was proposed through bioinformatics analyses and targeted gene knockout experiments. Bioassays revealed that aglomycin A exhibits synergistic antibacterial activity with linezolid against vancomycin-resistant Enterococcus faecium (VRE) both in vitro and in vivo. We envision that DeepHalo (freely available at https://github.com/xieyying/DeepHalo) will become a powerful tool for accelerating the discovery of halogenated "dark matter".

biochemistry↗

bioRxiv2025

Protein Function Prediction via Contig-Aware Multi-Level Feature Integration

Proteins play a central role in biological processes, and accurately predicting their functions is crucial for biomedical research. While computational methods have advanced significantly, most approaches rely solely on sequence or structure, neglecting critical inter-protein relationships, such as the topological arrangement of coding sequences (CDSs) within contigs. To address this gap, we propose CAML, a novel deep learning model that integrates intra-protein features including sequence and predicted structure with inter-protein features capturing functional linkages among CDSs in contigs. Specifically, CAML employs a Graph Isomorphism Network (GIN) to extract structural features from predicted protein contact graphs and ESM-2 for sequence embeddings. Additionally, it leverages kmer frequencies and a Bidirectional Long Short-Term Memory (BiLSTM) network to model functional relationships among colocalized CDSs within contigs, capturing operon-like associations. Extensive experiments demonstrate that CAML outperforms the state-of-the-art methods in accuracy, precision, recall and F1-score, achieving improvements of 11.24%, 12.43%, 13.59%, and 13.30%, respectively over the second-best model. Ablation studies further confirm the critical contribution of CAMLs multi-level biological feature integration in enhancing functional annotation accuracy. Our study demonstrates the importance of integrating structural, sequential, and CDSs topological features for accurate protein function prediction, providing a robust computational framework for genomics research.

bioinformatics↗