Biology subjects

Zhao, H.

Publications and source records attributed to Zhao, H..

At least 523 records · Page 29Linked to original sources

bioRxiv2020

To improve the predictions of binding residues with DNA, RNA, carbohydrate, and peptide via multiple-task deep neural networks

MotivationThe interactions of proteins with DNA, RNA, peptide, and carbohydrate play key roles in various biological processes. The studies of uncharacterized protein-molecules interactions could be aided by accurate predictions of residues that bind with partner molecules. However, the existing methods for predicting binding residues on proteins remain of relatively low accuracies due to the limited number of complex structures in databases. As different types of molecules partially share chemical mechanisms, the predictions for each molecular type should benefit from the binding information with other molecules types. ResultsIn this study, we employed a multiple task deep learning strategy to develop a new sequence-based method for simultaneously predicting binding residues/sites with multiple important molecule types named MTDsite. By combining four training sets for DNA, RNA, peptide, and carbohydrate-binding proteins, our method yielded accurate and robust predictions with AUC values of 0.852, 0836, 0.758, and 0.776 on their respective independent test sets, which are 0.52 to 6.6% better than other state-of-the-art methods. More importantly, this study provides a new strategy to improve predictions by combining multiple similar tasks. Availabilityhttp://biomed.nscc-gz.cn/server/MTDsite/ Contactyangyd25@mail.sysu.edu.cn

bioinformatics↗

bioRxiv2020

DNA Methylation Repels Binding of HIF Transcription Factors to Maintain Tumour Immunotolerance

AO_SCPLOWBSTRACTC_SCPLOWO_ST_ABSBackgroundC_ST_ABSHypoxia is pervasive in cancer and other diseases. Cells sense and adapt to hypoxia by activating hypoxia-inducible transcription factors (HIFs), but it is still an outstanding question why cell types differ in their transcriptional response to hypoxia. ResultsHere, we report that HIFs fail to bind CpG dinucleotides that are methylated in their consensus binding sequence, both in in vitro biochemical binding assays and in vivo studies of differentially methylated isogenic cell lines. Based on in silico structural modelling, we show that 5-methylcytosine indeed causes steric hindrance in the HIF binding pocket. A model wherein cell-type-specific methylation landscapes, as laid-down by the differential expression and binding of other transcription factors under normoxia control cell-type-specific hypoxia responses is observed. We also discover ectopic HIF binding sites in repeat regions which are normally methylated. Genetic and pharmacological DNA demethylation, but also cancer-associated DNA hypomethylation, expose these binding sites, inducing HIF-dependent expression of cryptic transcripts. In line with such cryptic transcripts being more prone to cause double-stranded RNA and viral mimicry, we observe low DNA methylation and high cryptic transcript expression in tumours with high immune checkpoint expression, but not in tumours with low immune checkpoint expression, where they would compromise tumour immunotolerance. In a low-immunogenic tumour model, DNA demethylation upregulates cryptic transcript expression in a HIF-dependent manner, causing immune activation and reducing tumour growth. ConclusionsOur data elucidate the mechanism underlying cell-type specific responses to hypoxia, and suggest DNA methylation and hypoxia to underlie tumour immunotolerance.

genomics↗

bioRxiv2020

Imputing Single-cell RNA-seq data by combining Graph Convolution and Autoencoder Neural Networks

Single-cell RNA sequencing technology promotes the profiling of single-cell transcriptomes at an unprecedented throughput and resolution. However, in scRNA-seq studies, only a low amount of sequenced mRNA in each cell leads to missing detection for a portion of mRNA molecules, i.e. the dropout problem. The dropout event hinders various downstream analysis, such as clustering analysis, differential expression analysis, and inference of gene-to-gene relationships. Therefore, it is necessary to develop robust and effective imputation methods for the increasing scRNA-seq data. In this study, we have developed an imputation method (GraphSCI) to impute the dropout events in scRNA-seq data based on the graph convolution networks. The method takes advantage of low-dimensional representations of similar cells and gene-gene interactions to impute the dropouts. Extensive experiments demonstrated that GraphSCI outperforms other state-of-the-art methods for imputation on both simulated and real scRNA-seq data. Meanwhile, GraphSCI is able to accurately infer gene-to-gene relationships by utilizing the imputed matrix that are concealed by dropout events in raw data.

bioinformatics↗

bioRxiv2020

Rab10 Regulates Macropinocytosis in Phagocytes

Genetic variation in LRRK2 associates with susceptibility to Parkinsons disease, Crohns disease, and mycobacteria infection, with high expression of LRRK2, and the LRRK2 kinase substrate Rab10, in phagocytic cells in the immune system. In mouse and human primary monocyte-derived macrophages, dendritic cells, and microglia-like cells, we find that Rab10 specifically regulates a specialized form of endocytosis known as macropinocytosis, without affecting phagocytosis or clathrin-mediated endocytosis. LRRK2 phosphorylates cytoplasmic PI(3,4,5)P3-positive GTP-Rab10 early macropinosomes, before EEA1 and Rab5 recruitment occurs. Macropinosome cargo in macrophages includes CCR5, CD11b, and MHCII, with LRRK2-phosphorylation of Rab10 potently blocking EHBP1L1-mediated recycling tubules and cargo turnover. EHBP1L1 over-expression competitively inhibits LRRK2-phosphorylation of Rab10, mimicking the effects of LRRK2 kinase inhibition in promoting cargo recycling. Both Rab10 knockdown and LRRK2 kinase inhibition potently suppresses the maturation of macropinosome-derived CCR5-loaded signaling endosomes important for CCL5-induced AKT-activation and chemotaxis. These data support a novel axis in the endolysosomal system whereby LRRK2-mediated Rab10 phosphorylation stalls vesicle fast-recycling to promote PI3K-AKT signal transduction.

cell biology↗

bioRxiv2020

Evolutionary context-integrated deep sequence modeling for protein engineering

Protein engineering seeks to design proteins with improved or novel functions. Compared to rational design and directed evolution approaches, machine learning-guided approaches traverse the fitness landscape more effectively and hold the promise for accelerating engineering and reducing the experimental cost and effort. A critical challenge here is whether we are capable of predicting the function or fitness of unseen protein variants. By learning from the sequence and large-scale screening data of characterized variants, machine learning models predict functional fitness of sequences and prioritize new variants that are very likely to demonstrate enhanced functional properties, thereby guiding and accelerating rational design and directed evolution. While existing generative models and language models have been developed to predict the effects of mutation and assist protein engineering, the accuracy of these models is limited due to their unsupervised nature of the general sequence contexts they captured that is not specific to the protein being engineered. In this work, we propose ECNet, a deep-learning algorithm to exploit evolutionary contexts to predict functional fitness for protein engineering. Our method integrated local evolutionary context from homologous sequences that explicitly model residue-residue epistasis for the protein of interest, as well as the global evolutionary context that encodes rich semantic and structural features from the enormous protein sequence universe. This biologically motivated sequence modeling approach enables accurate mapping from sequence to function and provides generalization from low-order mutants to higher-orders. Through extensive benchmark experiments, we showed that our method outperforms existing methods on [~]50 deep mutagenesis scanning and random mutagenesis datasets, demonstrating its potential of guiding and expediting protein engineering.

bioinformatics↗

bioRxiv2020

mTORC2 mediate FLCN-induced HIF2α nuclear import and proliferation of clear cell renal cell carcinoma

Clear cell renal cell carcinoma (ccRCC), as the most important type of renal carcinoma, has a high incidence and easy metastasis. Folliculin (FLCN) was identified as a tumor suppressor gene. Its deletions and mutations are associated with a potential risk of kidney cancer. At present, the specific molecular mechanism of FLCN-induced proliferation, invasion and migration in clear cell renal cell carcinoma remains elusive. In this study, we demonstrated that FLCN controled cell proliferation, invasion and migration through PI3K/mTORC2 pathway. FLCN combined with HIF2 in various normal and cancerous renal cells, and mTORC2 mediate FLCN effectively alleviated the deterioration of renal cancer cells by degrading HIF2. Silencing of FLCN showed promotion of HIF2 protein expression, which in turn led to an increase in downstream target genes Cyclin D1 and MMP9. Moreover, when interfering with siFLCN, HIF2 degradation rate was delayed, and the time of entry into the nucleus was advanced. Taken together, our study illustrated that mTORC2 promoted the specific molecular mechanism of HIF2 by down-regulated FLCN, and might be a new therapeutic target against renal cancer progression.

cell biology↗

bioRxiv2020

Inactivation of LCN2/EGR1 Promotes Oligodendrocyte Progenitor Cell Differentiation and Remyelination after White Matter Injury

The insufficient remyelination due to the impaired oligodendrocyte precursor cell differentiation and maturation is highly associated with irreversible white matter injury and neurological deficits. Consequently, inhibitory components and microenvironment for remyelination might serve as potential therapeutic targets for treating white matter injury after acute central nervous system injury and neurodegeneration diseases. Lipocalin-2 was recently reported to corelate with white matter in both atypical, acute white matter injured disease subarachnoid hemorrhage and typical, chronic white matter injured disease multiple sclerosis. To elucidate the role and underlying mechanism of Lipocalin-2 in oligodendrocyte precursor cell differentiation and remyelination, we used genetic inhibition and a constitutive conditional knockout model with subarachnoid hemorrhage or multiple sclerosis. We found that the genetic inhibition of the increase in Lipocalin-2 promoted oligodendrocyte precursor cell differentiation, remyelination, and functional recovery after subarachnoid hemorrhage or multiple sclerosis. Unexpectedly, the inhibition of Lipocalin-2 did not reduce glial activation and inflammation. Lipocalin-2 was shown to activate Early Growth Response Protein 1 in oligodendrocyte precursor cells, which is partly regulated by its receptor SLC22A17. In the conditional knockout of Early Growth Response Protein 1 in oligodendrocyte precursor cells, we discovered enhanced oligodendrocyte precursor cell differentiation in developing and injured white matter; consistently, the specific inactivation of Early Growth Response Protein 1 promoted remyelination and neurological recovery after subarachnoid hemorrhage or multiple sclerosis. Thus, we propose that following white matter injury in humans, the increase in Lipocalin-2 activates Early Growth Response Protein 1 and consequently impair oligodendrocyte precursor cell differentiation and myelin repair. Our results suggest that therapies specifically inactivating Lipocalin-2/ Early Growth Response Protein 1 signal in oligodendroglial lineage cells could represent a novel strategy to enhance differentiation and remyelination in white matter injury patients.

neuroscience↗

bioRxiv2019

Replication timing maintains the global epigenetic state in human cells.

DNA is replicated in a defined temporal order termed the replication timing (RT) program. RT is spatially segregated in the nucleus with early/late replication corresponding to Hi-C A/B chromatin compartments, respectively. Early replication is also associated with active histone modifications and transcriptional permissiveness. However, the mechanistic interplay between RT, chromatin state, and genome compartmentalization is largely unknown. Here we report that RT is central to epigenome maintenance and compartmentalization in both human embryonic stem cells (hESCs) and cancer cell line HCT116. Knockout (KO) of the conserved RT control factor RIF1, rather than causing discrete RT switches as previously suspected, lead to dramatically increased cell to cell heterogeneity of RT genome wide, despite RIF1s enrichment in late replicating chromatin. RIF1 KO hESCs have a nearly random RT program, unlike all prior RIF1 KO cells, including HCT116, which show localized alterations. Regions that retain RT, which are prevalent in HCT116 but rare in hESCs, consist of large H3K9me3 domains revealing two independent mechanisms of RT regulation that are used to different extents in different cell types. RIF1 KO results in a striking genome wide downregulation of H3K27ac peaks and enrichment of H3K9me3 at large domains that remain late replicating, while H3K27me3 and H3K4me3 are re-distributed genome wide in a cell type specific manner. These histone modification changes coincided with global reorganization of genome compartments, transcription changes and a genome wide strengthening of TAD structures. Inducible degradation of RIF1 revealed that disruption of RT is upstream of genome compartmentalization changes. Our findings demonstrate that disruption of RT leads to widespread epigenetic mis-regulation, supporting previously speculative models in which the timing of chromatin assembly at the replication fork plays a key role in maintaining the global epigenetic state, which in turn drives genome architecture.

molecular biology↗

bioRxiv2019

Cep57 and Cep57l1 cooperate to recruit the Cep63-Cep152 complex for centriole biogenesis

Cep152 and Cep63 act as the cradle and recruit Plk4 to initiate the centriole biogenesis in a mother centriole dependent manner. However, how the Cep152-Cep63 complex is targeted to the proximal end of mother centrioles is unclear. In this study, we show that Cep57 and its paralog, Cep57l1, colocalize with Cep63 and Cep152 at the proximal end of mother centrioles in both cycling cells and differentiated multiciliated cells. Both Cep57 and Cep57l1 associate with the Cep152-Cep63 complex by directly binding to the centrosomal targeting region of Cep63. Co-depletion of Cep57 and Cep57l1 blocks the loading of Cep63-Cep152 to the centriole and subsequently prevents centriole duplication. We propose that Cep57 and Cep57l1 act together to ensure the recruitment of the Cep63-Cep152 complex to the mother centrioles for procentriole formation. Summary statementCep57 and its paralog, Cep57l1, cooperatively act as a molecular scaffold to recruit the Cep63-Cep152 cradle to the proximal end of centrioles in the early stages of centriole duplication.

cell biology↗

bioRxiv2019

Deficiency of Norepinephrine Transmission In Brain Contributed To Seizure-Induced Respiratory Arrest In the DBA/1 Mouse SUDEP Model

ObjectiveSudden unexpected death in epilepsy (SUDEP) is the leading cause of mortality in patients in patients with intractable epilepsy. However, the pathogenesis of SUDEP seems to be poorly understood. Our previous findings showed that the incidence of seizure-induced respiratory arrest (S-IRA) was markedly reduced by atomoxetine in a murine SUDEP model. Because the central NE -1 receptor (NE-1R) plays a vital role in regulating respiratory function, we hypothesized that the suppression of S-IRA by atomoxetine was mediated by NE/NE-1R interactions that can be reversed by NE-1R antagonism. MethodsWe examined whether atomoxetine-mediated suppression of S-IRA evoked by either acoustic stimulation or pentylenetetrazole (PTZ) in DBA/1 mice can be reversed by intraperitoneal (IP) and intracerebroventricular (ICV) administration of prazosin, a selective antagonist of NE-1R. The content and activity of tyrosine hydroxylase (TH), a rate-limiting enzyme for NE synthesis, in the lower brainstem was measured by ELISA. Electroencephalograms (EEG) were obtained by using the PTZ-evoked SUDEP model. ResultsAtomoxetine-mediated suppression of S-IRA evoked by either acoustic stimulation or PTZ was significantly reversed by low doses of IP and ICV prazosin. Neither repetitive acoustic stimulation nor S-IRA reduced TH levels in lower brainstem. However, the enzyme activity of TH levels in lower brainstem was significantly increased by mechanical ventilation with DBA/1 mice ,which makes dead DBA/1 mice suffered from S-IRA and SUDEP recover. EEG data showed that although the protective effect of atomoxetine was reversed by prazosin, neither drug affected EEG activity. SignificanceThese data suggest that deficient synthesis of NE and norepinephrinergic neurotransmission contributes to S-IRA and that the NE-1R is a potential therapeutic target for the prevention of SUDEP.

neuroscience↗

bioRxiv2019

The Hippo pathway origin and its oncogenic alteration in evolution

The Hippo pathway is a central regulator of organ size and a key tumor suppressor via coordinating cell proliferation and death. Initially discovered in Drosophila, the Hippo pathway has been implicated as an evolutionarily conserved pathway in mammals; however, how this pathway was evolved to be functional from its origin is still largely unknown. In this study, we traced the Hippo pathway in premetazoan species, characterized the intrinsic functions of its ancestor components, and unveiled the evolutionary history of this key signaling pathway from its unicellular origin. In addition, we elucidated the paralogous gene history for the mammalian Hippo pathway components and characterized their cancer-derived somatic mutations from an evolutionary perspective. Taken together, our findings not only traced the conserved function of the Hippo pathway to its unicellular ancestor components, but also provided novel evolutionary insights into the Hippo pathway organization and oncogenic alteration.

evolutionary biology↗

bioRxiv2019

microRNA seedless sites attenuate strong-seed-site-mediated target repression

MicroRNAs (miRNAs) regulate protein-coding gene expression primarily through cognitive binding sites in the 3 untranslated regions (3' UTRs). Seed sites are sequences in messenger RNAs (mRNAs) that form perfect Watson-Crick base-paring with a miRNAs seed region, which can effectively reduce mRNA abundance and/or repress protein translation. Some seedless sites, which do no form perfect seed-pairing with a miRNA, can also lead to target repression, often with lower efficacy. Here we report the surprising finding that when seedless sites and seed sites are co-present in the same 3UTR, seedless sites attenuate strong-seed-site-mediated target suppression, independent of 3' UTR length. This attenuation effect is detectable in >70% of transcriptomic datasets examined, in which specific miRNAs are experimentally increased or decreased. The attenuation effect is confirmed by 3UTR reporter assays and mediated through base-pairing between miRNA and seedless sites. Furthermore, this seedless-site-based attenuation effect could affect seed sites of the same miRNA or another miRNA, thus partially explaining the variability in target suppression and miRNA-mediated gene upregulation. Our findings reveal an unexpected principle of miRNA-mediated gene regulation, and could impact the understanding of many miRNA-regulated biological processes.

bioinformatics↗

bioRxiv2019

Biosynthetic engineering of the antifungal, anti-MRSA auroramycin

Using an established CRISPR-Cas mediated genome editing technique for streptomycetes, we explored the combinatorial biosynthesis potential of the auroramycin biosynthetic gene cluster in Streptomyces roseoporous. Auroramycin is a potent anti-MRSA polyene macrolactam. In addition, it also displays antifungal activities, which is unique among structurally similar polyene macrolactams, such as incednine and silvalactam. In this work, we employed different engineering strategies to target glycosylation and acylation biosynthetic machineries within its recently elucidated biosynthetic pathway. Six auroramycin analogs with variations in C-, N-methylation, hydroxylation and extender units incorporation were produced and characterized. By comparing the bioactivity profiles of these analogs, we determined that unique disaccharide motif of auroramycin is essential for its antimicrobial bioactivity. We further demonstrated that C-methylation of the 3, 5-epi-lemonose unit, which is unique among structurally similar polyene macrolactams, is key to its antifungal activity.

bioengineering↗

bioRxiv2019

Alpha-synuclein antisense oligonucleotides as a disease-modifying therapy for Parkinson’s disease

Parkinsons disease (PD) is a prevalent neurodegenerative disease with no approved disease-modifying therapies. Multiplications, mutations, and single nucleotide polymorphisms in the SNCA gene, encoding alpha-synuclein protein (aSyn), either cause or increase risk for PD. Intracellular accumulations of aSyn are pathological hallmarks of PD. Taken together, reduction of aSyn production may provide a disease-modifying therapy for PD. We show that antisense oligonucleotides (ASOs) reduce production of aSyn in rodent pre-formed fibril (PFF) models of PD. Reduced aSyn production leads to prevention and removal of established aSyn pathology and prevents dopaminergic cell dysfunction. In addition, we address the translational potential of the approach through characterization of human SNCA targeting ASOs that efficiently suppress the human SNCA transcript in vivo. We demonstrate broad activity and distribution of the human SNCA ASOs throughout the non-human primate brain and a corresponding decrease in aSyn cerebral spinal fluid (CSF) levels. Taken together, these data suggest that by inhibiting production of aSyn it may be possible to reverse established pathology and thus supports the development of SNCA ASOs as a potentially disease modifying therapy for PD and related synucleinopathies.\n\nSummaryAntisense oligonucleotides designed against SNCA, which are progressing to the clinic, have the potential to be a disease modifying therapeutic for Parkinsons disease patients.

neuroscience↗

bioRxiv2019

Establishment of the trans-resveratrol-induced mouse pulmonary fibrosis model and its comparison with the bleomycin-induced model

Pulmonary fibrosis (PF) is a progressive, fatal disease, and its pathogenic mechanism has not yet been identified. The bleomycin (BLM)-induced animal model has several shortcomings that cannot be overcome. We accidentally found that the continuous oral administration of trans-resveratrol (TR) at levels above the prescribed dose can induce PF within a certain period of time, and this model can be successfully replicated in mice, as indicated by the hydroxyproline (HYP) test and pathological analysis. The TR model requires no anesthesia or surgery and thus causes less damage to mice than the BLM model. The progress of fibrosis in the TR model was slow but progressed steadily compared with the BLM model, and the TR model was more comparable to human PF. Although the pathogenesis is not understood, the TR-induced PF model indicated that there may be a close relationship between PF and tumors and antitumor drugs, which requires further exploration.

pharmacology and toxicology↗

bioRxiv2019

Imputing missing RNA-seq data from DNA methylation by using transfer learning based-deep neural network

BackgroundGene expression plays a key intermediate role in linking molecular features at DNA level and phenotype. However, due to various limitations in experiments, the RNA-seq data is missing in many samples while there exists high-quality of DNA methylation data. As DNA methylation is an important epigenetic modification to regulate gene expression, it can be used to predict RNA-seq data. For this purpose, many methods have been developed. A common limitation of these methods is that they mainly focus on single cancer dataset, and do not fully utilize information from large pan-cancer dataset. ResultsHere, we have developed a novel method to impute missing gene expression data from DNA methylation data through transfer learning-based neural network, namely TDimpute. In the method, the pan-cancer dataset from The Cancer Genome Atlas (TCGA) was utilized for training a general model, which was then fine-tuned on the specific cancer dataset. By testing on 16 cancer datasets, we found that our method significantly outperforms other state-of-the-art methods in imputation accuracy with 7%-11% increase under different missing rates. The imputed gene expression was further proved to be useful for downstream analyses, including the identification of both methylation-driving and prognosis-related genes, clustering analysis, and survival analysis on the TCGA dataset. More importantly, our method was indicated to be useful for general purpose by the independent test on the Wilms tumor dataset from the Therapeutically Applicable Research To Generate Effective Treatments (TARGET) project. ConclusionsTDimpute is an effective method for RNA-seq imputation with limited training samples.

bioinformatics↗

bioRxiv2019

SCRIBE: a new approach to dropout imputation and batch effects correction for single-cell RNA-seq data

Single-cell RNA sequencing technologies are widely used in recent years as a powerful tool allowing the observation of gene expression at the resolution of single cells. Two of the major challenges in scRNA-seq data analysis are dropout events and batch effects. The inflation of zero(dropout rate) varies substantially across single cells. Evidence has shown that technical noise, including batch effects, explains a notable proportion of this cell-to-cell variation. To capture biological variation, it is necessary to quantify and remove technical variation. Here, we introduce SCRIBE (Single-Cell Recovery Imputation with Batch Effects), a principled framework that imputes dropout events and corrects batch effects simultaneously. We demonstrate, through real examples, that SCRIBE outperforms existing scRNA-seq data analysis tools in recovering cell-specific gene expression patterns, removing batch effects and retaining biological variation across cells. Our software is freely available online at https://github.com/YiliangTracyZhang/SCRIBE.

bioinformatics↗

bioRxiv2019

Genomic Characterization of Posttraumatic Stress Disorder in a Large US Military Veteran Sample

Individuals vary in their liability to develop Posttraumatic Stress Disorder (PTSD), the symptoms of which are highly heterogeneous, following exposure to life-threatening trauma. Understanding genetic factors that contribute to the biology of PTSD is critical for refining diagnosis and developing new treatments. Using genetic data from more than 250,000 participants in the Million Veteran Program, genomewide association analyses were conducted using a validated electronic health record-based algorithmically-defined PTSD diagnosis phenotype (48,221 cases and 217,223 controls), and PTSD quantitative symptom phenotypes (212,007 individuals). We identified several genome-wide significant loci in the case-control analyses, and numerous such loci in the quantitative trait analyses, including some (e.g., MAD1L1; TCF4; CRHR1) that were associated with multiple symptom sub-domains and total symptom score, and others that were more specific to certain symptom sub-domains (e.g., CAMKV to re-experiencing; SOX6 to hyperarousal). Genetic correlations between all pairs of symptom sub-domains and their total were very high (rg 0.93 - 0.98) supporting validity of the PTSD diagnostic construct. We also demonstrate strong shared heritability with a range of traits, show that heritability persists when conditioned on other major psychiatric disorders, present independent replication results, provide support for one of the implicated genes in postmortem brain of individuals with PTSD, and use this information to identify potential drug repositioning candidates. These results point to the utility of genetics to inform and validate the biological coherence of the PTSD syndrome despite considerable heterogeneity at the symptom level, and to provide new directions for treatment development.

genetics↗