Biology subjects

Ning, K.

Publications and source records attributed to Ning, K..

At least 19 recordsLinked to original sources

bioRxiv2026

Generation of human iPSC-derived motor neurons with defined regional identities across the rostrocaudal axis

This study establishes a unified pipeline for generating human induced pluripotent stem cell (iPSC)-derived motor neurons (iMNs) with defined hindbrain, cervical, and lumbar identities to support anatomically resolved studies of motor neuron development and degeneration. Conventional differentiation methods can efficiently produce iMNs but provide limited control over rostrocaudal identity and frequently generate populations with heterogeneous HOX profiles. We addressed this limitation by optimizing developmental morphogen combinations and treatment windows, establishing three complementary protocols that reproducibly generated hindbrain/rhombomere-associated motor neurons (RiMNs), cervical motor neurons (CiMNs), and lumbar motor neurons (LiMNs) across iPSC lines from six non-diseased donors balanced by sex. Transcriptomic and protein analyses demonstrated graded rostrocaudal specification from RiMNs through CiMNs to LiMNs, together with region-appropriate motor neuron subtype features. These identities showed the expected relationships to human week 9 fetal rostrocaudal HOX programs. High-density multielectrode array analysis further revealed synchronized network firing, electrically active neurite networks, and enhanced axonal signal propagation in the regional iMN populations compared with conventionally differentiated iMNs. Spike-sorted recordings additionally revealed protocol-associated differences in single-neuron electrophysiological profiles. Together, this three-protocol pipeline enables the reproducible generation and comparative analysis of anatomically defined human motor neuron populations and provides region-matched cellular models for investigating motor neuron development, selective degeneration, and therapeutic responses.

neuroscience↗

bioRxiv2026

NK-RCPP: a receptor-conditioned computational workflow for prioritizing neurokinin-receptor peptide candidates

Prioritizing peptides for closely related G protein-coupled receptors requires comparing target-directed scores with evidence from other structural contexts. We applied the Neurokinin Receptor-Conditioned Peptide Prioritization workflow (NK-RCPP) to 96 generated candidates for neurokinin receptors NK1R, NK2R and NK3R. The retrospective analysis combined receptor-matched on-target REF15 scores with 4,518 paired structural-neighbour comparisons covering 82 candidates. Lower on-target scores than the matched controls were observed for 14/32 NK1R, 23/32 NK2R and 6/32 NK3R candidates. Adding the neighbour-score criterion retained 16 of the 33 evaluable on-target-favoured candidates, reducing that pool by 51.5%. At matched receptor-specific shortlist sizes, half of the candidate identities differed from on-target-only selection. NK2R-04 and NK2R-19 remained selected at the higher neighbour-fraction threshold of 0.75. Fourteen NK2R candidates lacked neighbour summaries, leaving their joint-rule status unresolved. Sequence-frequency and embedding analyses described the candidates within a 1,048-peptide comparison space, and an exploratory 10-peptide NK2R comparison showed differences in standardized scores with uncertain score-potency agreement. These findings show how paired structural-neighbour scoring and threshold choice shape candidate selection within a receptor family. The resulting shortlist provides a defined set for testing whether joint prioritization enriches for active, receptor-preferred peptides.

bioinformatics↗

bioRxiv2026

PaleoRigor: a bounded research agent with a reproducible evidence trail for paleomicrobiome data processing

Background: Ancient microbial evidence informs research on past health, diet, and environments, but its scarcity and susceptibility to contamination demand care in selecting and processing data. Language models can now turn a research request into an analysis plan, but an unconstrained planner may reference files that do not exist, apply operations a format forbids, or answer a question the data cannot settle. A planner used on such evidence therefore needs to be bounded, and its decisions recoverable afterwards. Results: We developed PaleoRigor around three commitments. Agent: a language model plans the analysis. Constraint: a validator checks every file reference, format and step dependency before anything runs, execution is restricted to predefined local modules, and a planning contract directs the model to refuse out-of-scope requests with a named reason code rather than answer them. Reproducibility: each run is exported with its plan, intermediate files, checksums and decisions. In a preregistered, frozen held-out evaluation, PaleoRigor passed 23 of 24 runs (95.8%; Wilson 95% confidence interval, 79.8-99.3%): 11 of 12 supported runs and all 12 prespecified boundary decisions. An ablation that removed this planning contract from the same model passed 19 of 24 (79.2%). Repeating the frozen design with a second model configuration yielded 24 of 24 successes for PaleoRigor and 19 of 24 for the ablated arm. Pooled post hoc across both configurations, PaleoRigor passed 47 of 48 runs (97.9%), compared with 38 of 48 (79.2%) for the ablated arms. All 96 run labels were recomputed from retained records and matched. Dataset-level tests matched read counts and compressed-file sizes for six public sequencing records and documented the removal of 114 duplicate sequence entries. A retraction-associated audit illustrated how expert review limited conclusions to file-level evidence. Conclusions: For paleomicrobiome data, pairing a language-model agent with explicit constraints and a reproducible record makes the agent's proposed analysis reviewable before its outputs support a biological claim. PaleoRigor delivers checked inputs, a recorded transformation history and a named reason for each request it refuses, and runs as a local desktop application. It prepares data for specialist analysis; ancient-DNA authentication and biological interpretation remain with the paleobiologist.

bioinformatics↗

bioRxiv2026

MetaClaw: an auditable AI agent for end-to-end, multi-directional metagenomic and multi-omics analysis

End-to-end omics analysis requires more than selecting tools: a usable agent must bind data correctly, execute long workflows without blocking, preserve provenance and recover the biological conclusions that motivate an analysis. Existing LLM-driven bioinformatics agents automate parts of this process, but their operational dependencies and conclusion-level validity are often unclear. Here we present MetaClaw, an auditable agent that maps a user request to a registered workflow, executes standardized upstream processing on FlowHub, and runs study-specific downstream analyses in network-isolated OpenClaw containers. A YAML registry and an explicit plan-submit-poll-finalise lifecycle record file bindings, parameters, scripts, environments and outputs in per-job bundles. Across the full cohorts of four published studies (769 metagenomic profiles), MetaClaw recovered 4/4 sorghum marker groups, 3/3 RRMS features, 4/5 canonical CRC markers among the top 20 classifier features and 5/5 permafrost marker groups. In 45 model-by-prompt runs, upstream completion was consistent whereas downstream validity depended on the backend and instruction detail; three decoy-tested endpoints showed no significant differences. In 48 ablation sessions, removing the registry, planning loop or manifest caused distinct losses, with registry removal increasing time, tool calls and token cost. MetaClaw therefore connects standardized upstream execution, local analytical flexibility and conclusion-level validation in a rerunnable framework for metagenomic and microbiome multi-omics analysis.

bioinformatics↗

bioRxiv2026

MGM2 as a Unified Foundation Model for Microbiome World Exploration

Microbiomes are information-rich biological systems, yet most computational analyses still reduce communities to cohort-specific abundance tables. Here we introduce MGM2, a multimodal foundation model pretrained on 1,821,291 MicrobeAtlas samples and 225,067 OTUs clustered at 99% sequence similarity. MGM2 couples NTv3-derived microbial sequence embeddings with abundance conditioning and community-semantic alignment to learn transferable sample- and token-level representations. Frozen MGM2 representations outperformed DeepPhylo by 0.06-0.21 macro-AUROC across five temporally held-out MGnify hierarchy levels, with the largest gains for rare and fine-grained labels. In fecal microbiota transplantation, MGM2-XLarge achieved a response ROC AUC of 0.79 and reduced post-transplant Bray-Curtis distance by 15% relative to the recipient baseline. The same representation supported ASV-level trend forecasting across 24 wastewater treatment plants. Sparse autoencoder analysis resolved MGM2-XLarge token states into a 4,096-feature dictionary spanning taxonomic identity, abundance state, ecological context and technical variation. MGM2 therefore provides a sequence-aware and interpretable representation layer for microbiome classification, paired-community prediction, forecasting and feature discovery. Highlights[bullet] MGM2 integrates sequence, abundance and community semantics through pretraining on 1.82 million microbiome samples. [bullet]Frozen MGM2 improved macro-AUROC over DeepPhylo by 0.06-0.21 across five temporally held-out MGnify levels. [bullet]MGM2-XLarge reached a response ROC AUC of 0.79 and reduced post-FMT Bray-Curtis distance by 15%. [bullet]A 4,096-feature sparse autoencoder atlas resolves taxonomic, abundance, ecological and technical signals.

bioinformatics↗

bioRxiv2026

AAV VP1 unique region (VP1u) determines GPR108 dependence for AAV transduction of human airway epithelium and its rescue by Doxorubicin

rAAV2.5T was identified through directed evolution of an AAV capsid library in polarized human airway epithelium (HAE) cultured at an air-liquid interface (ALI). The capsid gene of rAAV2.5T is a chimera of the N-terminal unique region of AAV2 VP1 (VP1u) and the VP2 and VP3 regions of AAV5 with a single A581T substitution at the variable region (VR) VIII of the capsids. GPR108, a G protein-coupled receptor, is known as an essential host factor for the transduction of rAAV2 but not of rAAV5. Both AAV2 and AAV5 VP1u colocalized well with GPR108 and, to a lesser extent, with the trans-Golgi network (TGN). GPR108 knockout (KO) abolished rAAV2.5T transduction in both HeLa cells and HAE-ALI cultures. Remarkably, short-term treatment with doxorubicin (DOX) at 2 {micro}M completely restored transduction, indicating that DOX can compensate for the loss of GPR108 function. DOX enhanced rAAV2.5T transduction by 50-100-fold in wild-type HAE-ALI cultures and by over 300-fold in the GPR108-deficient cultures. Mechanistic studies demonstrated that this enhancement resulted from altered intracellular trafficking that promoted efficient vector nuclear import, rather than increased vector internalization, proteasome inhibition, or activation of the DNA damage response. Importantly, we identified that the N-terminal 15 amino acids of AAV2 VP1u as the primary determinant of rAAV2.5T dependence on GPR108 for transduction. Collectively, these findings demonstrate that productive transduction of rAAV2.5T in polarized HAE cultures depends on GPR108-mediated intracellular trafficking that limits efficient nuclear entry, and that DOX can relieve this constraint by promoting efficient vector import. SignificanceAAV2.5T is an airway-tropic vector with considerable promise for pulmonary gene therapy. We found that host factor GPR108 is required for rAAV2.5T trafficking from the TGN to the nucleus and that this step constitutes a major bottleneck to productive transduction in polarized HAE. In contrast, KIAA0319L (AAVR) plays a key role in AAV intracellular trafficking from the endosome to the TGN but not in internalization into polarized HAE during apical transduction. Transient treatment with low-dose doxorubicin (DOX, 2 {micro}M) enhanced rAAV2.5T transduction in HAE by 50-100-fold through a significant increase in vector nuclear import. Notably, DOX can overcome the transduction deficit caused by GPR108 deficiency, but not that caused by AAVR deficiency. Mechanistically, the N-terminal 15 amino acids of the VP1u confer GPR108 dependence during rAAV2.5T apical transduction of polarized HAE. DOX bypasses this requirement by promoting efficient nuclear import without affecting vector internalization, inhibiting proteasomes, or inducing DNA damage response.

microbiology↗

bioRxiv2026

Protein Predictors of Worsening Cerebral Edema in Traumatic Brain Injury in Critically Ill Patients

Cerebral edema (CE) is a major determinant of poor outcomes after traumatic brain injury (TBI). CE affects approximately 60% of patients with mass lesions on head computed tomography (hCT) and confers a tenfold increase in mortality. Despite this burden, no validated biomarkers exist to identify patients at risk for early CE worsening or to elucidate the biological processes underlying divergent edema trajectories. We conducted high-throughput plasma proteomic profiling in a prospective moderate-severe TBI cohort with evidence of intracranial blood on the initial post-injury hCT scan (n=123 patients) to (1) identify a clinically feasible set of biomarkers plus a predictive model that could be used to assess CE risk at admission, and (2) conduct a mechanistic, hypothesis-generating analysis to identify the molecular networks involved in CE progression in the first hours post-injury. Plasma collected within 3 hours of injury was analyzed using the Olink Explore platform (5394 proteins), and patients underwent head CT imaging upon presentation and at 6 hours. Two complementary analytic pathways, reflecting the studys two goals, were pursued in parallel. First, logistic regression with false discovery rate (FDR) correction identified a conservative 12-protein set associated with baseline CE. When combined with targeted clinical variables (i.e., age, sex, admission GCS, pupil reactivity, admission glucose, admission alcohol), this panel supported high-performing supervised classifiers for predicting CE worsening by 6 hours post-baseline scan (best model: XGBoost, AUC = 0.78; recall = 0.83). Second, a broader 60-protein panel, selected via bootstrapped elastic-net regression, was used to interrogate the mechanistic architecture of CE worsening using random forest SHAP attribution and protein-protein interaction modeling. Proteomic signatures diverged sharply between patients whose CE did and did not worsen. The CE worsening group was characterized by a coherent neuronal-synaptic injury axis dominated by ELAVL4, CEND1, NEFL, NECAB2, GFAP, CHGB, RPH3A, HPCAL4, and HOMER1, proteins involved in neuronal structural integrity, vesicular trafficking, synaptic vesicle cycling, calcium signaling, and axonal degeneration. These reciprocal proteomic states suggest that early edema progression may be driven by coordinated disruption of neuronal and synaptic resilience programs. Together, these findings suggest (1) a hyperacute biomarker panel plus predictive model with potential for prospective validation and (2) mechanistic evidence for a distinct neuronal-synaptic injury network that associates with early edema worsening after TBI.

neuroscience↗

bioRxiv2026

Identification of Human Transferrin Receptor as an Entry Co-receptor for Parvovirus B19 Infection of Human Erythroid Progenitor Cells

Parvovirus B19 (B19V), a member of the genus Erythroparvovirus within the Parvoviridae family, infects human erythroid progenitor cells (EPCs) of bone marrow and fetal liver, and causes various hematological disorders. The minor capsid protein VP1 of B19V contains a unique N-terminal region (VP1u) that facilitates virus binding and internalization into EPCs via its receptor-binding domain (RBD). We previously identified tyrosine protein kinase receptor UFO (AXL) as a proteinaceous receptor for B19V infection of EPCs. In this study, we employed an ascorbate peroxidase 2 (APEX2)-based proximity labeling method to identify host proteins that are associated with B19V VP1u during entry. This analysis revealed human transferrin receptor 1 (hTfR) as a key host protein associated with VP1u. hTfR knockdown in UT7/Epo-S1 cells, a B19V-permissive human megakaryoblastoid leukemia cell line, showed significantly reduced B19V internalization and replication. Biolayer interferometry (BLI) assays confirmed a direct interaction between B19V VP1u and hTfR extracellular domain (ECD). Inhibition of VP1u interaction with hTfR ECD, either by a monoclonal antibody targeting the apical domain of the ECD or human ferritin, a natural ligand of hTfR that binds the apical domain, significantly reduced VP1u binding to hTfR, as well as B19V internalization and B19V replication in ex vivo-expanded EPCs. Furthermore, mutant RBD proteins that bear amino acid substitutions in the three helical domains nearly abolished RBD binding to hTfR and significantly reduced the ability to inhibit B19V infection of EPCs. Collectively, our findings establish hTfR as a B19V entry co-receptor that mediates B19V internalization into its natural host EPCs. SignificanceB19V causes severe hematological disorders, including transient aplastic crisis, chronic pure red cell aplasia, and hydrops fetalis, by selectively infecting erythroid progenitor cells (EPCs). Despite its clinical impact, no approved antivirals or vaccines exist, largely due to limited understanding of viral entry mechanisms. A unique feature of B19V is the externalization of the VP1 unique region (VP1u) from the viral capsid, which mediates receptor engagement. Our prior studies identified AXL as an attachment receptor for B19V. Here, we identify that human transferrin receptor 1 (hTfR) acts as a critical co-receptor that directly binds VP1u and promotes viral internalization. Inhibition of the VP1u-hTfR interaction by competitive binding of hTfR with either an anti-hTfR monoclonal antibody or human ferritin significantly reduces B19V internalization and replication in ex vivo-expanded EPCs, highlighting a link between VP1u binding to the apical domain of hTfR and viral internalization. RBD mutants that disrupt its interaction with hTfR barely inhibited B19V infection in EPCs. These findings support a receptor-switch model in which AXL mediates attachment and hTfR drives internalization. Defining these mechanisms provides a foundation for developing antiviral strategies targeting B19V entry into EPCs.

microbiology↗

bioRxiv2025

Discovery and optimization of antimicrobial peptides from extreme environments on global scale

The escalating crisis of global antimicrobial resistance (AMR) necessitates the discovery of novel antibiotics. Antimicrobial peptides (AMPs), particularly those from under-explored extreme environments, represent a promising therapeutic class. Here, we introduce SEGMA (Structure-aware Extremophile Genome Mining for Antimicrobial peptides), a computational framework that integrates structure information to systematically mine AMPs from extremophile genomes on a global scale. By analyzing 60,461 extremophile metagenome-assembled genomes (MAGs) from diverse habitats, SEGMA identified 3,298 novel AMPs (termed "extremocins"), which exhibit unique amino acid profiles and physicochemical properties. Leveraging a beam search-guided optimization strategy, we further enhanced selected extremocins to achieve broad-spectrum antimicrobial activity. Experimental validation confirmed potent in vitro efficacy against clinically relevant pathogens. This study underscores the value of structure-aware mining and extremophile microbiomes in expanding the antibiotic arsenal against AMR. HighlightsO_LISEGMA, a structure-aware deep learning framework, mines 3,298 novel antimicrobial peptides (extremocins) from 60,461 extremophile genomes on global scale. C_LIO_LIExtremocins exhibit unique sequence features, and expand known antibiotic space, few of which shows homology to existing AMP databases. C_LIO_LIA beam search-guided optimization strategy enhanced selected extremocins to achieve broad-spectrum activity against clinically relevant pathogens. C_LIO_LIExperimental validation confirmed that candidate extremocins exhibit potent in vitro and in vivo antimicrobial activity, highlighting their therapeutic potential. C_LI

bioinformatics↗

bioRxiv2025

MicroProphet: A Digital Twin Framework for Predicting Microbial Community Dynamics with Personalized Precision

The ability to accurately predict the dynamic evolution of microbial communities is critical for advancing personalized medicine, precision intervention, and ecological system management. However, the irregular sampling, high missingness, and complex temporal behaviors that characterize longitudinal microbiome datasets present substantial challenges to reliable forecasting. Here we propose MicroProphet, a personalized digital twin framework capable of accurately forecasting microbial abundance trajectories from incomplete longitudinal observations without the need for data interpolation. By leveraging a time-aware Transformer architecture, MicroProphet reconstructs individualized microbial trajectories using as little as the initial 30% of time points, capturing critical transitional states through its attention mechanism. We demonstrate its robust cross-ecosystem generalizability across synthetic communities, human gut microbiomes, infant gut development, and corpse decomposition. In clinical contexts, MicroProphet enables early identification of disease-related microbial shifts and supports intervention timing optimization, exemplified in inflammatory bowel disease and antibiotic perturbation responses. By transforming incomplete and sparse data into actionable forecasts, MicroProphet establishes a foundation for real-time microbial monitoring, therapeutic decision support, and precision ecological management, paving the way for broader applications of digital twin systems in biology and personalized healthcare. HighlightsO_LIMicroProphet establishes a personalized digital twin framework for forecasting biological dynamics from incomplete longitudinal observations, enabling precision health monitoring and intervention planning. C_LIO_LIBy leveraging a transformer-based architecture, MicroProphet accurately reconstructs microbial community trajectories using as little as 30% of initial time points, without the need for data interpolation. C_LIO_LIThe framework demonstrates robust cross-ecosystem generalizability across clinical, early-life development, and forensic contexts. C_LIO_LIMicroProphet empowers real-time tracking of microbial shifts, early detection of disease-associated transitions, and timing optimization for microbiome-targeted therapeutic strategies. C_LI

microbiology↗

bioRxiv2025

Deciphering Biosynthetic Gene Clusters with a Context-aware Protein Language Model

Microbial secondary metabolites, synthesized by biosynthetic gene clusters (BGCs), offer vast potential for biotechnological applications. Among BGC profiling techniques, computational detection methods face challenges, including time-consuming alignment and reliance on predefined profiles. To address these, we present BGC-Finder, an end-to-end pipeline utilizing protein language models for BGC detection and annotation from microbial genomes and metagenomes. This approach achieves remarkable increase in profiling speed of up to 100-fold, and employs genomic context-aware modeling to facilitate interpretable genetic essentiality assessment and large-scale BGC clustering. BGC-Finder outperformed traditional methods, successfully detecting 9.49% more biosynthetic-core genes and 27.70% more cytochrome P450s in 742 experimentally-validated BGCs. Notably, it retrieved 31 remote biosynthetic homologs from 210 polar marine metagenomes and identified 4,585 BGCs with 6,388 core genes from 256 fungal genomes. These findings highlight BGC-Finders capability to illuminate "microbial biosynthesis dark matter" (sequence-unrelated, function-similar biosynthetic enzymes) and expedite natural product discovery. HighlightsO_LIBGC-Finder is an accurate and ultrafast pipeline leveraging protein language models (pLMs) to predict and annotate biosynthetic gene clusters (BGCs) from microbial genomes and metagenomes. C_LIO_LIThe genomic context-aware model enables interpretable analysis: attention-driven identification of essential biosynthetic genes and embedding-guided BGC clustering. C_LIO_LIBGC-Finder sensitively retrieves remote homologous BGCs from both bacteria and fungi genomes, uncovering hidden microbial biosynthesis dark matter. C_LIO_LIWe discovered a non-ribosomal peptide synthetase (NRPS) family, which involved into function-specific BGCs in two evolutionarily distant fungi. C_LI

bioinformatics↗

bioRxiv2025

Towards a Generative Paradigm for Large-scale Microbiome Analysis by Generative Language Model

Microbiome analysis has traditionally relied on taxonomic abundance tables, which, while effective, often constrain the exploration of deeper contextual relationships. In this study, we present MGM 2.0, a novel framework that applies advanced natural language processing (NLP) techniques to microbiome research. By reimagining microbiome samples as sentences and microbial species as words, MGM 2.0 enabled the extraction of nuanced patterns and relationships. The model demonstrated robust predictive performance in identifying exogenous species colonization (AUROC = 0.86). Additionally, through prompt-guided microbiome data generation, MGM 2.0 produced realistic microbial profiles conditioned on disease labels. The framework further revolutionized donor selection in fecal microbiota transplantation (FMT) by framing it as a sequence-to-sequence prediction task, enabling the prediction of post-transplantation community compositions and the identification of super donors for personalized treatments (average increase in C2R = 0.52). This innovative integration of NLP and microbiome science provides a versatile toolkit for predictive modeling, data generation, and personalized medicine. HighlightsO_LIIntroduced MGM 2.0, a generative language model utilizing sentence-like representation for microbiome analysis and generation. C_LIO_LIDemonstrated that sentence-like representation preserves sample distinctions, enabling accurate microbial sample classification tasks, such as colonization prediction. C_LIO_LIGenerated realistic, disease-specific microbiome profiles using a prompt-guided approach, validated by a novel "Microbiome Turing Test." C_LIO_LIApplied MGM 2.0 to fecal microbiota transplantation (FMT) donor selection, accurately predicting post-transplant community compositions and identifying potential "super donors" for personalized treatment strategies. C_LI

bioinformatics↗

bioRxiv2025

Microbial general model: Leveraging large language model for contextualized microbiome analysis

Microbial communities significantly impact medicine, biotechnology, and agriculture. Advanced sequencing technologies have generated extensive microbiome data, enabling the discovery of substantial evolutionary and ecological patterns. However, traditional supervised learning methods struggle to capture universal patterns in microbial community data, largely due to the large data heterogeneity and profound batch effects among samples, rendering it difficult to classify samples as well as detect biomarkers from millions of samples, not to say the intricate but important dynamic patterns from a variety of contextualized sceneries. In this study, we introduce the Microbial General Model (MGM), the first microbiome community foundation model pre-trained on a dataset of 263,302 microbiome samples using language modeling techniques. MGM demonstrated significant improvements in microbial community classification compared to traditional machine learning methods. Additionally, MGM has enabled contextualized classification, effectively overcomes cross-regional limitations, showing enhanced performance on intercontinental datasets through transfer learning. Furthermore, fine-tuning MGM on a longitudinal infant dataset revealed distinct keystone genera during development, with Bacteroides and Bifidobacterium exhibiting higher attention weights in vaginal deliveries, and Haemophilus in cesarean deliveries. Finally, through in silico modeling, the model also uncovered novel microbial dynamic patterns in a Crohns disease cohort following antibiotic treatment. In conclusion, by leveraging self-attention and autoregressive pre-training, MGM serves as a versatile model for various downstream microbiome tasks and holds significant potential for achieving contextualized aims. Key pointsO_LIThe Microbial General Model (MGM) is a foundation model with millions of parameters pre-trained on sub-million microbial community data. C_LIO_LIMGM outperforms traditional methods in various microbiome classification and prediction tasks, such as microbial community classification. C_LIO_LIMGM effectively captures the spatial and temporal dynamics of microbial communities. C_LIO_LIMGM could detect the effects of perturbation on microbial community through in silico experiments. C_LI

bioinformatics↗

bioRxiv2024

Intra-tumor microbiome-based tumor survival indices predict immune interaction and drug sensitivity on pan-cancer scale

Growing research evidence indicates a substantial influence of the intra-tumor microbiome on tumor outcome. However, there is currently no consistent criterion for identifying the association of microbes with tumor progression and response to treatment across various types of cancer. In this study, we concentrate on the intra-tumor microbiome and develop the Tumor Microbiome Survival Index (TMSI), a measure indicative of cancer patient survival risk. Our indices revealed notable distinctions between two stratified risk groups for each of the 10 cancer types and could precisely predict patients overall survival. For each type of cancer, our findings unveiled two distinct gene expression profiles and shed light on the varying patterns of immune and stromal cell enrichment between the two risk groups. Additionally, we noted that the high-TMSI group exhibited substantially elevated IC50 values for a number of drugs, indicating that individuals in the low-TMSI group might experience superior therapeutic effects from chemotherapy. These findings illuminate the complex dynamics between the tumor microbiome, the patients immune reaction, and medical outcomes, thus shedding light on microbiome-based personalized therapeutic interventions. HighlightsO_LIThe TMSI model generates a survival risk score based on the intra-tumor microbiome. C_LIO_LIThe TMSI model predicts survival rates accurately across multiple cancer types. C_LIO_LIMolecular-level analysis highlights distinct immune interactions with tumor microbes. C_LIO_LIHigh-TMSI groups showed altered drug sensitivity and implying potential differences in treatment response. C_LI

bioinformatics↗

bioRxiv2024

Identification of SLC35A1 as an essential host factor for the transduction of multi-serotype recombinant adeno-associated virus (AAV) vectors

We conducted a genome-wide CRISPR/Cas9 screen in suspension 293-F cells transduced with rAAV5. The highly selected genes revealed after two rounds of screens included the previously reported KIAA039L, TM9SF2, and RNF121, along with a cluster of genes involved in glycan biogenesis, Golgi apparatus localization and endoplasmic reticulum penetration. In this report, we focused on solute carrier family 35 member A1 (SLC35A1), a Golgi apparatus-localized cytidine 5-monophosphate-sialic acid (CMP-SIA) transporter. We confirmed that SLC35A1 knockout (KO) significantly decreased rAAV5 transduction to a level lower than that observed in KIAA0319L or TM9SF2 KO cells. Although SLC35A1 KO drastically reduced the expression of 2,6-linked SIA on the cell surface, the expression of 2,3-linked SIA, as well as the cell binding and internalization of rAAV5, were only moderately affected. Moreover, SLC35A1 KO significantly diminished the transduction of AAV multi-serotypes, including rAAV2 and rAAV3 which do not utilize SIAs for primary attachment. Notably, the SLC35A1 KO markedly increased transduction of rAAV9 and rAAV11, which primarily attach to cells via binding to galactose. Further analyses revealed that SLC35A1 KO significantly decreased vector nuclear import. More importantly, although the C-terminal cytoplasmic tail deletion ({Delta}C Tail) mutant of SLC35A1 did not drastically decrease SIA expression, it significantly decreased rAAV transduction, as well as vector nuclear import, suggesting the C-tail is critical in these processes. Furthermore, the T128A mutant significantly decreased SIA expression, but still supported rAAV transduction and nuclear import. These findings highlight the involvement of the CMP-SIA transporter in the intracellular trafficking of rAAV vectors post-internalization. IMPORTANCErAAV is an essential tool for gene delivery in the treatment of genetic disorders, yet the mechanisms of rAAV transduction remain partially understood. GPR108 is vital for the transduction of most rAAV vectors, but not for rAAV5. We aimed to identify host factors that impact AAV5 transduction akin to GPR108. Using a genome-wide CRISPR/Cas9 screen in 293-F cells, we identified SLC35A1, a Golgi apparatus-localized CMP-sialic acid transporter that transports CMP-sialic acid from cytoplasm into the Golgi apparatus for sialylation, is essential to rAAV transduction. Further studies across various AAV serotypes showed SLC35A1 significantly affects vector nuclear import post-internalization. These results underscore the crucial role of SLC35A1 in intracellular trafficking beyond the initial cell attachment of rAAV.

microbiology↗

bioRxiv2024

SMeta, a binning tool using single-cell sequences to aid reconstructing metageome species accurately

Because of the large volume and complex structure of metagenomic data, traditional binning methods are often hard to classify microbial metagenomes effectively. To deal with these challenges, introducing longer and more accurate single-cell sequencing data is a possible solution. Inspired by the existing MetaBAT2 tool, this study develops a new vector-based binning algorithm, SMeta, which uses both metagenomic and single-cell sequencing data. SMeta is specifically designed for eukaryotic microbial metagenomes, with the long reads characteristic of single-cell data. By introducing the segment tree data structure, the algorithm aligns long single-cell sequences with short metagenomic sequences quickly. This approach allows for the use of reference genomes from genomic databases to replace single-cell data, which makes more precise identification and reconstruction possible for small genome fragments, which are typically overlooked by traditional methods. Also, it might provide a higher purity sequence set for subsequent assemblies.

bioinformatics↗

bioRxiv2024

Non-destructive in situ monitoring of structural changes of 3D tumor spheroids during the formation, migration, and fusion process

For traditional laboratory microscopy observation, the multi-dimensional, real-time, in situ observation of three-dimensional (3D) tumor spheroids has always been the pain point in cell spheroid observation. In this study, we designed a side-view observation petri dish/device that reflects light, enabling in situ observation of the 3D morphology of cell spheroids using conventional inverted laboratory microscopes. We used a 3D-printed handle and frame to support a first surface mirror, positioning the device within a cell culture petri dish to image cell spheroid samples. The imaging conditions, such as the distance between the mirror and the 3D spheroids, the light source, and the impact of the culture medium, were systematically studied to validate the in-situ side-view observation. The results proved that placing the surface mirror adjacent to the spheroids enables non-destructive in situ real-time tracking of tumor spheroid formation, migration, and fusion dynamics. The correlation between spheroid thickness and dark core appearance under light microscopy and the therapeutic effects of chemotherapy doxorubicin and Natural Killer cells on spheroidss 3D structure was investigated.

cancer biology↗

bioRxiv2024

Regulation of brain development by the Minibrain/Rala signaling network

The human dual specificity tyrosine phosphorylation regulated kinase 1A (DYRK1A) is implicated in the pathology of Down syndrome, microcephaly, and cancer, however the exact mechanism through which it functions is unknown. Here, we have studied the role of the Drosophila ortholog of DYRK1A, Minibrain (Mnb), in brain development. The neuroblasts (neural stem cells) that eventually give rise to differentiated neurons in the adult brain are formed from a specialized tissue in the larval optic lobe called the neuroepithelium, in a tightly regulated process. Molecular marker analysis of mnb mutants revealed alterations in the neuroepithelium and neuroblast regions of developing larval brains. Using affinity purification-mass spectrometry (AP-MS), we identified the novel Mnb binding partners Ral interacting protein (Rlip) and RALBP1 associated Eps domain containing (Reps). Rlip and Reps physically and genetically interact with Mnb, and the three proteins may form a ternary complex. Mnb phosphorylates Reps, and human DYRK1A binds to the Reps orthologs REPS1 and REPS2. Furthermore, Mnb engages the small GTPase Ras-like protein A (Rala) to regulate brain and wing development. This work uncovers a previously unrecognized early role of Mnb in the neuroepithelium and defines the functions of the Mnb/Reps/Rlip/Rala signaling network in brain development. Significance statementThe kinase Minibrain(Mnb)/DYRK1A regulates the development of the brain and other tissues across many organisms. Here we show the critical importance of Mnb within the developing neuroepithelium. Advancing our understanding of Mnb function, we identified novel protein interactors of Mnb, Reps and Rlip, which function together with Mnb to regulate growth in Drosophila melanogaster. We also identify and characterize a role for the small GTPase Rala in Mnb-regulated growth and nervous system development. This work reveals an early role of Mnb in brain development and identifies a new Mnb/Reps/Rlip/Rala signaling axis.

developmental biology↗