Biology subjects

Xie, Z.

Publications and source records attributed to Xie, Z..

At least 109 records · Page 6Linked to original sources

bioRxiv2022

GenSLMs: Genome-scale language models reveal SARS-CoV-2 evolutionary dynamics.

We seek to transform how new and emergent variants of pandemiccausing viruses, specifically SARS-CoV-2, are identified and classified. By adapting large language models (LLMs) for genomic data, we build genome-scale language models (GenSLMs) which can learn the evolutionary landscape of SARS-CoV-2 genomes. By pretraining on over 110 million prokaryotic gene sequences and finetuning a SARS-CoV-2-specific model on 1.5 million genomes, we show that GenSLMs can accurately and rapidly identify variants of concern. Thus, to our knowledge, GenSLMs represents one of the first whole genome scale foundation models which can generalize to other prediction tasks. We demonstrate scaling of GenSLMs on GPU-based supercomputers and AI-hardware accelerators utilizing 1.63 Zettaflops in training runs with a sustained performance of 121 PFLOPS in mixed precision and peak of 850 PFLOPS. We present initial scientific insights from examining GenSLMs in tracking evolutionary dynamics of SARS-CoV-2, paving the path to realizing this on large biological data.

bioinformatics↗

bioRxiv2022

Improve the Protein Complex Prediction with Protein Language Models

AlphaFold-Multimer has greatly improved protein complex structure prediction, but its accuracy also depends on the quality of the multiple sequence alignment (MSA) formed by the interacting homologs (i.e., interologs) of the complex under prediction. Here we propose a novel method, denoted as ESMPair, that can identify interologs of a complex by making use of protein language models (PLMs). We show that ESMPair can generate better interologs than the default MSA generation method in AlphaFold-Multimer. Our method results in better complex structure prediction than AlphaFold-Multimer by a large margin (+10.7% in terms of the Top-5 best DockQ), especially when the predicted complex structures have low confidence. We further show that by combining several MSA generation methods, we may yield even better complex structure prediction accuracy than Alphafold-Multimer (+22% in terms of the Top-5 best DockQ). We systematically analyze the impact factors of our algorithm and find out the diversity of MSA of interologs significantly affects the prediction accuracy. Moreover, we show that ESMPair performs particularly well on complexes in eucaryotes.

bioinformatics↗

bioRxiv2022

ReproTox-KG: Toxicology Knowledge Graph for Structural Birth Defects

Birth defects are functional and structural abnormalities that impact 1 in 33 births in the United States. Birth defects have been attributed to genetic as well as other factors, but for most birth defects there are no known causes. Small molecule drugs, cosmetics, foods, and environmental pollutants may cause birth defects when the mother is exposed to them during pregnancy. These molecules may interfere with the process of normal fetal development. To characterize associations between small molecule compounds and their potential to induce specific birth abnormalities, we gathered knowledge from multiple sources to construct a reproductive toxicity Knowledge Graph (ReproTox-KG) with an initial focus on associations between birth defects, drugs, and genes. Specifically, to construct ReproTox-KG we gathered data from drug/birth-defect associations from co-mentions in published abstracts, gene/birth-defect associations from genetic studies, drug- and preclinical-compound-induced gene expression data, known drug targets, genetic burden scores for all human genes, and placental crossing scores for all small molecules in ReproTox-KG. Using the data stored within ReproTox-KG, we scored 30,000 preclinical small molecules for their potential to induce birth defects. Querying the ReproTox-KG, we identified over 500 birth-defect/gene/drug cliques that can be used to explain molecular mechanisms for drug-induced birth defects. The ReproTox-KG is provided as curated tables and via a web-based user interface that can enable users to explore the associations between birth defects, approved and preclinical drugs, and human genes.

bioinformatics↗

bioRxiv2022

Graph pangenome reveals functional, evolutionary, and phenotypic significance of human nonreference sequences

Thousands of DNA sequences in global populations are not present in the human reference genome, named nonreference sequence (NRS). Long-read sequencing (LRS) technologies enable better discovery of NRS with large length, particularly in repetitive regions. Here, we de novo assembled 539 genomes in five genetically divergent human populations sequenced by LRS technology and identified 5.1 million NRSs. These NRSs were merged into 45,284 nonredundant NRSs, of which 66.2% were novel. 78.5% of NRSs were repeat sequences, such as VNTR and STR. 38.7% of NRSs were common in the five populations, 35.6% were population specific, while 21.3% were ancestral and present in nonhuman primates. 144 NRS hotspots spanned 141 Mb of the human genome and many NRSs contained known functional domains or intersected with coding genes. Based on graph-based pangenome, we detected 565 transcript expression quantitative trait loci on NRSs, of which 467 were novel. We also detected 39 NRS candidates for adaptive selection within the human population related to the language system and diabetes. GWAS revealed 14 NRSs significantly associated with eight phenotypes, such as anaemia. Furthermore, we identified 154 NRSs in strong linkage disequilibrium with 258 phenotype-associated SNPs in the GWAS catalogue. Our work expands the landscape of human NRS and provides novel insights into functions of NRS to facilitate evolutionary and biomedical research.

genetics↗

bioRxiv2022

Potent, Gut-Restricted Inhibitors of Divalent Metal Transporter 1 (DMT1): Preclinical Efficacy Against Iron Overload and Safety Evaluation

Divalent metal transporter 1 (DMT1) cotransports ferrous iron and protons and is the primary mechanism for uptake of non-heme iron by enterocytes. Inhibitors are potentially useful as therapeutic agents to treat iron overload disorders such as hereditary hemochromatosis or {beta}-thalassemia intermedia, provided that inhibition can be restricted to the duodenum. We used a calcein quench assay to identify human DMT1 (hDMT1) inhibitors. Dimeric constructs were made to generate more potent compounds with low systemic exposure. Direct block of DMT1 was confirmed by voltage clamp measurements. The lead compound, XEN602, strongly inhibits dietary iron uptake in both rats and pigs, yet has negligible systemic exposure. Efficacy is maintained for >2 weeks in a rat subchronic dosing assay. Doses that lowered iron content in the spleen and liver by >50% had no effect on the tissue content of other divalent cations except for cobalt. XEN602 represents a powerful pharmacological tool for understanding the physiological function of DMT1. SIGNIFICANCE STATEMENTThis report introduces methodology to develop potent, gut-restricted inhibitors of DMT1 and identifies XEN602 as a suitable compound for in vivo studies. We also report novel animal models to quantify the inhibition of dietary uptake of iron in both rodents and pigs. This research shows that inhibition of DMT1 is a promising means to treat iron overload disorders.

pharmacology and toxicology↗

bioRxiv2022

Biochemical characterization of a Mycobacterium tuberculosis prolyl dipeptidyl peptidase that displays a remarkable thermostability

Dipeptidyl peptidases constitute a class of non-classical serine proteases that regulate an array of biological functions, making them pharmacologically attractive enzymes. With this work, we identified and characterized a dipeptidyl peptidase from Mycobacterium tuberculosis (MtDPP) displaying a strong preference for proline residues at the P1 substrate position and an unexpectedly high thermal stability. MtDPP was also characterized with alanine replacements of residues of its active site that yielded, for the most part, loss of catalysis. We show that MtDPP catalytic activity is inhibited by well-known human DPP4 inhibitors. Using MALDI-TOF mass spectrometry we also describe that in vitro, MtDPP mediates the truncation of the C-X-C motif chemokine ligand 10, indicating a plausible role in immune modulation for this mycobacterial enzyme.

biochemistry↗

bioRxiv2022

Restoration of PITPNA in Type 2 diabetic human islets reverses pancreatic beta-cell dysfunction

Defects in insulin processing and granule maturation are linked to pancreatic beta-cell failure during type 2 diabetes (T2D). Phosphatidylinositol transfer protein alpha (PITPNA) stimulates activity of phosphatidylinositol (PtdIns) 4-OH kinase to produce sufficient PtdIns-4-phosphate (PtdIns-4-P) in the trans-Golgi network to promote insulin granule maturation. PITPNA in beta-cells of T2D human subjects is markedly reduced suggesting its depletion accompanies beta-cell dysfunction. Conditional deletion of Pitpna in the beta-cells of Ins-Cre;Pitpnaflox/flox mice leads to hyperglycemia resulting from decreased glucose-stimulated insulin secretion (GSIS) and reduced pancreatic beta-cell mass. Furthermore, PITPNA silencing in human islets confirmed its role in PtdIns-4-P synthesis and led to impaired insulin granule maturation and docking, GSIS, and proinsulin processing with evidence of ER stress. Restoration of PITPNA in islets of T2D human subjects reversed these beta-cell defects and identify PITPNA as a critical target linked to beta-cell failure in T2D.

cell biology↗

bioRxiv2022

Single-Cell RNA-seq of Heart Reveals Intercellular Communication Drivers of Myocardial Fibrosis in Diabetic Mice

Diabetes-induced cardiomyopathy is characterized by myocardial fibrosis as a main pathology. In-depth study of cardiac heterogeneity and cell-to-cell interactions will help to reveal the pathogenesis of diabetic myocardial fibrosis and provide potential targets for the treatment of this disease. Here, we insighted into the intercellular communication drivers underlying myocardial fibrosis in mouse heart with high-fat-diet (HFD)/streptozotocin (STZ)-induced diabetes at single-cell resolution. Intercellular and protein-protein interaction networks of fibroblasts and macrophages, endothelial cells, as well as fibroblasts and epicardial cells reveal critical changes in ligand-receptor interactions such as Pdgf(s)-Pdgfra and Efemp1-Egfr, which promote the development of a profibrotic microenvironment during diabetes progression and confirm that specific inhibition of Pdgfra axis can significantly improve diabetic myocardial fibrosis. We further identified the phenotypically distinct Hrchi and Postnhi fibroblast subpopulations that are associated with pathological extracellular matrix remodeling, of which Hrchi fibroblasts are the most profibrogenic under diabetic conditions. Finally, we validated the role of Itgb1 hub gene mediated intercellular communication drivers of diabetic myocardial fibrosis in Hrchi fibroblasts, and confirmed the result by AAV9-mediated Itgb1 knockdown in the heart of diabetic mice. In summary, cardiac cell mapping provides novel insights into intercellular communication drivers underlying pathological extracellular matrix remodeling during diabetic myocardial fibrosis.

cell biology↗

bioRxiv2022

MrgprA3-expressing pruriceptors drive pruritogen-induced alloknesis through mechanosensitive Piezo2 channel

Although touch and itch are coded by distinct neuronal populations, light touch also provokes itch in the presence of exogenous pruritogens, resulting in a phenomenon called alloknesis. However, the cellular and molecular mechanisms underlying the initiation of pruritogen-induced mechanical itch sensitization are poorly understood. Here we show that intradermal injections of histamine or chloroquine (CQ) provoke alloknesis through activation of TRPV1- and MrgprA3- expressing prurioceptors, and functional ablation of these neurons reverses pruritogen-induced alloknesis. Moreover, genetic ablation of mechanosensitive Piezo2 channel function from MrgprA3-expressing prurioceptors also dampens pruritogen-induced alloknesis. Mechanistically, histamine and CQ sensitize Piezo2 channel function through activation of the PLC-PKC{delta} signaling. Collectively, our data uncovered a TRPV1+/MrgprA3+ prurioceptor-Piezo2 signaling axis in the initiation of pruritogen-induced mechanical itch sensitization in the skin.

neuroscience↗

bioRxiv2022

FunOMIC: Pipeline with built-in Fungal Taxonomic and Functional Databases for Human Mycobiome Profiling

While analysis of the bacterial microbiome has become routine, that of the fungal microbiome is still hampered by the lack of robust databases and bioinformatic pipelines. Here, we present FunOMIC, a pipeline with built-in taxonomic (1.6 million marker genes) and functional (3.4 million non-redundant fungal proteins) databases for the identification of fungi. Applied to more than 2,600 human metagenomic samples, the tool revealed fungal species associated with geography, body sites, and diseases. Correlation network analysis provided new insights into inter-kingdom interactions. With this pipeline and two of the most comprehensive fungal databases, we foresee a fast-growing resource for mycobiome studies.

bioinformatics↗

bioRxiv2022

Improving the RNA velocity approach using long-read single cell sequencing

The concept of RNA velocity has been recently developed that allowed to look at the otherwise static single-cell RNA sequencing data in a dynamic way, which permitted inferences about cell fates. However, the more precise parameters, such as the number of exons/introns, can also be determined using long-read methods. Comparing the numbers of exons and introns allows including more genes for downstream velocity analysis and resolves the precise cell fate. The recently developed concept of "RNA velocity" concerns with dynamic changes in mRNA expression and complements single-cell RNA sequencing (scRNA-seq) data, which are static snapshots of a certain cell state taken at a given time point1. RNA velocity measures the change in mRNA abundance by differentiating the newly transcribed unspliced pre-mRNAs from mature spliced mRNAs. The rapidly developing long-read sequencing technology lends itself for RNA velocity analysis of scRNA-seq data, which was previously performed primarily using second-generation sequencing.

bioinformatics↗

bioRxiv2022

Spatial chromatin accessibility sequencing resolves next-generation genome architecture

As the genome is organized into a three-dimensional structure in intracellular space, epigenomic information also has a complex spatial arrangement. However, most epigenetic studies describe locations of methylation marks, chromatin accessibility regions, and histone modifications in the horizontal dimension. Proper spatial epigenomic information has rarely been obtained. In this study, we designed spatial chromatin accessibility sequencing (SCA-seq) to resolve the genome conformation by capturing the epigenetic information in single-molecular resolution while simultaneously resolving the genome conformation. Using SCA-seq, we are able to examine the spatial interaction of chromatin accessibility (e.g. enhancer-promoter contacts), CpG island methylation, and spatial insulating functions of the CCCTC-binding factor. We demonstrate that SCA-seq paves the way to explore the mechanism of epigenetic interactions and extends our knowledge in 3D packaging of DNA in the nucleus.

genomics↗

bioRxiv2022

PvC3H29 interacts with and inhibits DNA binding of PvNAPs to finetune leaf senescence in switchgrass (Panicum virgatum)

Finetuning the progression of leaf senescence is important for plants fitness in nature, while staygreen with delayed leaf senescence has been considered as a valuable agronomic trait in crop genetic improvement. In this study, a switchgrass CCCH-type Zinc finger gene, PvC3H29, was characterized as a suppressor of leaf senescence that over-expressing or suppressing the gene led to delayed or accelerated leaf senescence, respectively. Transcriptomic analysis marked that chlorophyll catabolic pathway genes were involved in the PvC3H29-regulated leaf senescence. PvC3H29 was a nucleus-localized protein with no transcriptional activity. By Y2H screening, we identified its interacting proteins, including a pair of paralogous transcription factors, PvNAP1&2. Over-expressing the PvNAPs led to precocious leaf senescence at least partially by directly targeting and transactivating chlorophyll catabolic genes to promote chlorophyll degradation. PvC3H29, through protein-protein interaction, repressed the DNA-binding efficiency of PvNAPs and alleviated its transactivating effect on downstream genes, thereby functioned as a brake in the progression of leaf senescence. Moreover, over-expressing PvC3H29 resulted in up to 47% higher biomass yield and improved biomass feedstock quality, reiterating the importance of leaf senescence regulation in the genetic improvement of switchgrass and other feedstock crops. One-sentence summaryPvC3H29 interacts with transcription factors, PvNAP1&2, to inhibit their transactivation on chlorophyll catabolism and leaf senescence in switchgrass.

plant biology↗

bioRxiv2022

High coverage of single cell genomes by T7-assisted enzymatic methyl-sequencing

Conventional approaches to studying 5mC marks in single cells or samples with picogram input DNA amounts usually suffer from low genome coverage due to DNA degradation. Many methods have been developed to optimize the library construction efficiency for bisulfite-treated DNA. However, most of these approaches ignored the amplification bias of bisulfite-treated DNA, which leads to shallow genome coverage. In this study, we developed the T7-assisted enzymatic methyl-sequencing method (TEAM-seq), which adopts enzymatic conversion to minimize DNA degradation and T7 polymerase-assisted unbiased amplification. We demonstrate that TEAM-seq delivered, to the best of our knowledge, the highest reported coverage(70% for 100pg, 35% for 20pg) of single cell genomes in whole-genome 5mC sequencing.

genomics↗

bioRxiv2022

An evaluation of RNA-seq differential analysis methods

RNA-seq is a high-throughput sequencing technology widely used for gene transcript discovery and quantification under different biological or biomedical conditions. A fundamental research question in most RNA-seq experiments is the identification of differentially expressed genes among experimental conditions or sample groups. Numerous statistical methods for RNA-seq differential analysis have been proposed since the emergence of the RNA-seq assay. To evaluate popular differential analysis methods used in the open source R and Bioconductor packages, we conducted multiple simulation studies to compare the performance of eight RNA-seq differential analysis methods used in RNA-seq data analysis (edgeR, DESeq, DESeq2, baySeq, EBSeq, NOISeq, SAMSeq, Voom). The comparisons were across different scenarios with either equal or unequal library sizes, different distribution assumptions and sample sizes. We measured performance using false discovery rate (FDR) control, power, and stability. No significant differences were observed for FDR control, power, or stability across methods, whether with equal or unequal library sizes. For RNA-seq count data with negative binomial distribution, when sample size is 3 in each group, EBSeq performed better than the other methods as indicated by FDR control, power, and stability. When sample sizes increase to 6 or 12 in each group, DESeq2 performed slightly better than other methods. All methods have improved performance when sample size increases to 12 in each group except DESeq. For RNA-seq count data with log-normal distribution, both DESeq and DESeq2 methods performed better than other methods in terms of FDR control, power, and stability across all sample sizes. Real RNA-seq experimental data were also used to compare the total number of discoveries and stability of discoveries for each method. For RNA-seq data analysis, the EBSeq method is recommended for studies with sample size as small as 3 in each group, and the DESeq2 method is recommended for sample size of 6 or higher in each group when the data follow the negative binomial distribution. Both DESeq and DESeq2 methods are recommended when the data follow the log-normal distribution.

bioinformatics↗

bioRxiv2022

CRISPR-based targeted haplotype-resolved assemblies of a megabase region

Constructing high-quality haplotype-resolved genome assemblies has substantially improved the ability to detect and characterize genetic variants. A targeted approach providing readily access to the rich information from haplotype-resolved genome assemblies will be appealing to groups of basic researchers and medical scientists focused on specific genomic regions. Here, using the 4.5 megabase, notoriously difficult-to-assemble major histocompatibility complex (MHC) region as an example, we demonstrated an approach to construct haplotype-resolved de novo assemblies of targeted genomic regions with the CRISPR-based enrichment. Compared to the results from haplotype-resolved genome assemblies, our targeted approach achieved comparable completeness and accuracy with greatly reduced computing complexity, sequencing cost, as well as the amount of starting materials. Moreover, using the targeted assembled personal haplotypes as the reference both improves the quantification accuracy for sequencing data and enables allele-specific functional genomics analyses. Given its highly efficient use of resources, our approach can greatly facilitate population genetic studies of targeted regions, and may pave a new way to elucidate the molecular mechanisms in disease etiology.

genomics↗

bioRxiv2021

PCBP1 Deficient Pigs Hold the Potential to Inhibit CSFV Infection

Classical swine fever virus (CSFV), pathogen of classic swine fever, has caused severe economic losses worldwide. Poly (rC)-binding protein 1 (PCBP1), interacting with Npro of CSFV, plays a vital role in CSFV growth. Here, our research is the first report to generate PCBP1 knockout pigs via gene editing technology. The PCBP1 knockout pigs exhibited normal birth weight, reproductive-performance traits, and developed normally. Viral challenge results indicated that primary cells isolated from F0 and F1 generation pigs could significantly reduce CSFV infection. Additional mechanism exploration further confirmed that PCBP1 KO mediated antiviral effect is related with the activation of type I interferon. Beyond showing that gene editing strategy can be used to generate PCBP1 KO pigs, our study introduces a valuable animal model for further investigating infection mechanisms of CSFV that help to develop better antiviral solution. ImportanceAs a negative regulator in immune modulation, the effects of PCBP1 on viral replication have been found to be valuable. Here, this study was the first report to generate PCBP1 knockout pigs with normal pregnancy rate and viability. Primary cells isolated from F0 and F1 generation PCBP1 knockout pigs could significantly reduce CSFV infection. The PCBP1 knockout pigs could be used as a natural host models for investigating the effects of PCBP1-mediating critical interactions on viral replication and helping to develop better antiviral solution.

molecular biology↗

bioRxiv2021

Automated identification of chicken distress vocalisations using deep learning models

The annual global production of chickens exceeds 25 billion birds, and they are often housed in very large groups, numbering thousands. Distress calling triggered by various sources of stress has been suggested as an "iceberg indicator" of chicken welfare. However, to date, the identification of distress calls largely relies on manual annotations, which is very labour-intensive and time-consuming. Thus, a novel light-VGG11 was developed to automatically identify chicken distress calls using recordings (3,363 distress calls and 1,973 natural barn sounds) collected on intensive chicken farms. The light-VGG11 was modified from VGG11 with a significantly smaller size in parameters (9.3 million vs 128 million) and 55.88% faster detection speed while displaying comparable performance, i.e., precision (94.58%), recall (94.89%), F1-score (94.73%), and accuracy (95.07%), therefore more useful for model deployment in practice. To further improve the light-VGG11s performance, we investigated the impacts of different data augmentation techniques (i.e., time masking, frequency masking, mixed spectrograms of the same class, and Gaussian noise) and found that they could improve distress calls detection by up to 1.52%. In terms of precision livestock farming, our research opens new opportunities for developing technologies used to monitor the output of distress calls in large, commercial chicken flocks.

animal behavior and cognition↗