Biology subjects

Li, Z.-R.

Publications and source records attributed to Li, Z.-R..

7 recordsLinked to original sources

bioRxiv2025

Predicting the Evolutionary and Functional Landscapes of Viruses with a Unified Nucleotide-Protein Language Model: LucaVirus

Predicting viral evolution and function remains a central challenge in biology, hindered by high sequence divergence and limited knowledge compared to cellular organisms. Here, we introduce LucaVirus, a multi-modal foundation model for viruses, trained on 25.4 billion nucleotide and amino acid tokens covering nearly all known viruses. LucaVirus learns biologically meaningful representations capturing relationships between sequences, protein/gene homology, and evolutionary divergence. Using these embeddings, we developed downstream models that address key virology tasks: identifying hidden viruses in genomic "dark matter", annotating enzymatic activities of uncharacterized proteins, predicting viral evolvability, and identifying antibody candidates for emerging viruses. LucaVirus achieves state-of-the-art results in three tasks and matches leading models in the fourth with one-third the parameters. Together, these findings demonstrate the power of a unified foundation model to comprehensively decode the viral world and establish LucaVirus as an efficient and versatile platform for AI-driven virology, from virus discovery to functional and therapeutic predictions.

bioinformatics↗

bioRxiv2025

Composition of Carotid Plaques Differs Between Chinese and United States Patients: A Histology Study

BackgroundThe clinical manifestations of cerebrovascular disease are known to differ between the Chinese and United States (U.S.) populations as do the plaque features on imaging. ObjectivesThe aim of this study was to investigate and compare the histological features of excised carotid plaques from Chinese and U.S. patients. MethodsCarotid endarterectomy specimens collected from two prospective studies were included. The entire plaque was serially sectioned (10 {micro}m thickness) at 0.5-1 mm intervals. Hematoxylin and eosin staining and Mallorys trichrome staining were performed. The morphology and components of the plaques were measured and compared between the two groups. ResultsA total of 1,152 histological sections from 75 Chinese patients and 1,843 sections from 111 U.S. patients were analyzed. The Chinese group had significantly smaller minimum lumen diameters (median: 1.1 vs. 1.3 mm, p=0.046) and a larger percent wall volume (median: 74% vs. 70%, p=0.018) than the U.S. group. After adjusting for confounding factors, carotid plaques in the Chinese population were more likely to have more lipid pools ({beta}=10.0%, 95%CI: 4.9 to 15.9%), more recent intraplaque hemorrhage (IPH; {beta}=8.4%, 95%CI: 4.5 to 12.7%), and less late IPH ({beta}=-8.2%, 95%CI: -11.3 to -5.4), and fewer fibrous cap disruptions (45% vs. 67%, p=0.061). Chinese plaques were more homogeneous and had a higher percentage of plaques with features of xanthomas than did U.S. plaques (20% vs 2.7%, p<0.001). ConclusionsThe histology of Chinese plaques differs significantly from that of U.S. plaques, suggesting substantial differences in the pathophysiology of atherosclerotic cerebrovascular disease between Chinese and North American populations, which could enhance the gap in racial pathology comparison, indicating a need for a different management approach. Central illustration O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=154 SRC="FIGDIR/small/658993v1_ufig1.gif" ALT="Figure 1"> View larger version (83K): org.highwire.dtl.DTLVardef@14f16corg.highwire.dtl.DTLVardef@f60d66org.highwire.dtl.DTLVardef@1e9a822org.highwire.dtl.DTLVardef@11d7a31_HPS_FORMAT_FIGEXP M_FIG C_FIG

pathology↗

bioRxiv2024

Illuminating microbial phosphorus cycling in deep-sea cold seep sediments using protein language models

Phosphorus is essential for life and critically influences marine productivity. Despite geochemical evidence of active phosphorus cycling in deep-sea cold seep ecosystems, the microbial processes involved remain poorly understood. Traditional sequence-based searches often fail to detect proteins with remote homology. To address this, we developed a protein language model (PLM) named LucaPCycle that integrates sequence and structural information. This PLM-based approach identified 4,606 new phosphorus-cycling protein families based on the non-redundant gene and genome catalogs from global cold seeps, substantially enhancing our understanding of their diversity, ecology, and function. Among previously unannotated sequences, we discovered three novel alkaline phosphatase families (ALP1, ALP2 and ALP3) that feature unique domain organizations and preserved enzymatic capabilities. These results highlight previously overlooked ecological importance of phosphorus cycling within cold seeps, corroborated by data from porewater geochemistry, metatranscriptomics, and metabolomics. We identified a previously unrecognized diversity of archaea contributing to organic phosphorus mineralization and inorganic phosphorus solubilization through various mechanisms. This includes ecologically significant groups such as Asgardarchaeota, anaerobic methanotrophic archaea (ANME), and Thermoproteota. Additionally, viruses can enhance their hosts (e.g., ANME) phosphorus utilization through the PhoR-PhoB regulatory system and PhnCDE transporter, indirectly influencing methane dynamics. Overall, our PLM-based functional predictions are capable of accessing previously hidden sequence spaces for microbial phosphorus cycling, and can be applied to other various ecosystems.

microbiology↗

bioRxiv2024

Murine modeling of menstruation identifies immune correlates of protection during Chlamydia muridarum challenge.

The menstrual cycle influences the risk of acquiring sexually transmitted infections (STIs), including Chlamydia trachomatis (C. trachomatis), although the underlying immune contributions are poorly defined. A mouse model simulating the immune-mediated process of menstruation could provide valuable insights into tissue-specific determinants of protection against chlamydial infection within the cervicovaginal and uterine mucosae comprising the female reproductive tract (FRT). Here, we used the pseudopregnancy approach in naive C57Bl/6 mice and performed vaginal challenge with Chlamydia muridarum (C. muridarum) at decidualization, endometrial tissue remodeling, or uterine repair. This strategy identified that the time frame comprising uterine repair correlated with robust infection and greater bacterial burden as compared with mice on hormonal contraception, while challenges during endometrial remodeling were least likely to result in a productive infection. By comparing the infection site at early time points following chlamydial challenge, we found that a greater abundance of innate effector populations and proinflammatory signaling, including IFN{psi} correlated with protection. FRT immune profiling in uninfected mice over pseudopregnancy or in pig-tailed macaques over the menstrual cycle identified NK cell infiltration into the cervicovaginal tissues and lumen over the course of endometrial remodeling. Notably, NK cell depletion over this time frame reversed protection, with mice now productively infected with C. muridarum following challenge. This study shows that the pseudopregnancy murine menstruation model recapitulates immune changes in the FRT as a result of endometrial remodeling and identifies NK cell localization at the FRT as essential for immune protection against primary C. muridarum infection. Author SummaryAlthough the vast majority of women and adolescent girls of reproductive age experience menstruation, we have little insight into how this tissue remodeling process alters mucosal immune defenses against infection by genitourinary pathogens. In this study, we used a murine model of menstruation to investigate how endometrial shedding and repair alters the immune landscape in the female reproductive tract (FRT) to influence chlamydial infections. Using this approach, we identified that endometrial remodeling regulates a substantial pro-inflammatory immune response, including NK cell recruitment into the cervicovaginal tissues, and we further confirmed this phenomenon is occurring in a naturally menstruating species. The localization of NK cells in the FRT at the time of challenge was determined to be responsible for rapid immune protection that reduced C. muridarum burden, as experimental depletion of these cells over this timeframe now led to productive infections. Taken together, this study identifies that murine models of menstruation can be a valuable tool for investigating how the menstrual cycle modulates immune homeostasis and for identifying ways to strengthen mucosal immune defenses against genitourinary pathogens in women.

immunology↗

bioRxiv2023

Artificial intelligence redefines RNA virus discovery

Current metagenomic tools can fail to identify highly divergent RNA viruses. We developed a deep learning algorithm, termed LucaProt, to discover highly divergent RNA-dependent RNA polymerase (RdRP) sequences in 10,487 metatranscriptomes generated from diverse global ecosystems. LucaProt integrates both sequence and predicted structural information, enabling the accurate detection of RdRP sequences. Using this approach we identified 161,979 potential RNA virus species and 180 RNA virus supergroups, including many previously poorly studied groups, as well as RNA virus genomes of exceptional length (up to 47,250 nucleotides) and genomic complexity. A subset of these novel RNA viruses were confirmed by RT-PCR and RNA/DNA sequencing. Newly discovered RNA viruses were present in diverse environments, including air, hot springs and hydrothermal vents, and both virus diversity and abundance varied substantially among ecosystems. This study advances virus discovery, highlights the scale of the virosphere, and provides computational tools to better document the global RNA virome. In briefA deep learning algorithm (LucaProt) that integrates both sequence and predicted structural information was employed to identify highly divergent RNA viral "dark matter" in 10,487 metatranscriptomes from diverse global ecosystems. A total of 161,979 potential RNA virus species and 180 RNA virus supergroups was unveiled using this AI approach, including many understudied groups.

microbiology↗

bioRxiv2020

Identification of potential biomarkers or therapeutic targets of mesenchymal stem cells in multiple myeloma by bioinformatics analysis.

ObjectivesMesenchymal stem cells (MSCs) play important roles in multiple myeloma (MM) pathogenesis. Previous studies have discovered a group of MM-associated potential biomarkers in MSCs derived from bone marrow (BM-MSCs). However, no study of the bioinformatics analysis was conducted to explore the key genes and pathways of MSCs derived from adipose (AD-MSCs) in MM. The aim of this study was to screen potential biomarkers or therapeutic targets of AD-MSCs and BM-MSCs in MM. MethodsThe gene expression profiles of AD-MSCs (GSE133346) and BM-MSCs (GSE36474) were downloaded from Gene Expression Omnibus (GEO) database. Gene Oncology (GO) enrichment, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and protein-protein interaction (PPI) network of differentially expressed genes (DEGs) were performed. ResultsA total of 456 common downregulated DEGs in two datasets were identified and the remaining DEGs in GSE133346 were further identified as specific DEGs of AD-MSCs. Furthermore, a PPI network of common downregulated DEGs was constructed and seven hub genes were identified. Importantly, cell cycle was the most significantly enrichment pathway both in AD-MSCs and BM-MSCs from MM patients. ConclusionWe identified key genes and pathways closely related with MM progression, which may act as potential biomarkers or therapeutic targets of MM.

bioinformatics↗

bioRxiv2019

Macrocyclic colibactin induces DNA double-strand breaks via copper-mediated oxidative cleavage

Colibactin is an as-yet-uncharacterized human gut bacterial genotoxin, whose biosynthesis is linked to clb genomic island that distributes widespread in pathogenic and commensal human enterobacteria. Colibactin-producing gut microbes promote colon tumor formation and enhance progression of colorectal cancer (CRC) via DNA double-strand breaks (DSBs)-induced cellular senescence and death; however, the chemical basis contributing to the pathogenesis at the molecular level remains elusive. Here we report the discovery and the mechanism of action of colibactin-645 as the highly sought final colibactin metabolite with a novel molecular scaffold. Colibactin-645 recapitulates its previously assumed genotoxicity and cytotoxicity, exhibiting a strong DNA DSBs activity in vitro and in human cell cultures via a unique copper-mediated oxidative mechanism. We also present a complete model for colibactin biosynthesis, revealing an unprecedented dual function of the aminomalonate-utilizing polyketide synthases. This work thus provides the first molecular basis for colibactins genotoxic activity and facilitates further mechanistic study of colibactin-related CRC incidence and prevention.

microbiology↗