Biology subjects

Xie, L.

Publications and source records attributed to Xie, L..

At least 145 records · Page 8Linked to original sources

bioRxiv2022

Dietary fibre controls blood pressure and cardiovascular risk by lowering large intestinal pH and activating the proton-sensing receptor GPR65

High blood pressure (BP) is the most common cause of death globally, due to increasing the risk of cardiovascular diseases. Dietary fiber regulates BP through gut microbial production of acidic metabolites known as short-chain fatty acids (SCFAs). The specific mechanisms of how SCFAs regulate BP are still emerging. In a phenome-wide association study, we identified that the proton-sensing G-protein-coupled receptor GPR65 gene is associated with hypertension and its associated end-organ damage phenotypes. We hypothesized that acidic metabolites produced from the gut microbiota may activate GPR65, thus conferring BP regulating effects. We found that dietary fiber levels determined the luminal and interstitial tissue pH in the large intestine through production of SCFAs by the gut microbiota. We identified that low pH produced by high fiber intake, acting via GPR65 signaling, increased cAMP production and phosphorylation of CREB, and restricted the production of hypertension-promoting inflammatory cytokines by CD8+ T cells. Gpr65-/- mice spontaneously developed higher BP, cardiac and renal hypertrophy and fibrosis. We showed that the benefits of a diet high in fiber, which prevented hypertension and associated end-organ damage, were decreased in Gpr65-/- mice. Finally, adoptive transfers revealed that GPR65 deficiency in CD8+ T cells causally explained this phenotype. In conclusion, we showed that pH sensing by GPR65 in CD8+ T cells mediates much of the cardiovascular benefits of dietary fiber. pH sensing represents a novel gene-by-environment interaction of gut microbiota-to-host biological effects and may form the basis for new therapeutic strategies for hypertension.

molecular biology↗

bioRxiv2022

Binding Site-enhanced Sequence Pretraining and Out-of-cluster Meta-learning Predict Genome-Wide Chemical-Protein Interactions for Dark Proteins

Discovering chemical-protein interactions for millions of chemicals across the entire human and pathogen genomes is instrumental for chemical genomics, protein function prediction, drug discovery, and other applications. However, more than 90% of gene families remain dark, i.e., their small molecular ligands are undiscovered due to experimental limitations and human biases. Existing computational approaches typically fail when the unlabeled dark protein of interest differs from those with known ligands or structures. To address this challenge, we developed a deep learning framework PortalCG. PortalCG consists of four novel components: (i) a 3-dimensional ligand binding site enhanced sequence pre-training strategy to represent the whole universe of protein sequences in recognition of evolutionary linkage of ligand binding sites across gene families, (ii) an end-to-end pretraining-fine-tuning strategy to simulate the folding process of protein-ligand interactions and reduce the impact of inaccuracy of predicted structures on function predictions under a sequence-structure-function paradigm, (iii) a new out-of-cluster meta-learning algorithm that extracts and accumulates information learned from predicting ligands of distinct gene families (meta-data) and applies the meta-data to a dark gene family, and (iv) stress model selection that uses different gene families in the test data from those in the training and development data sets to facilitate model deployment in a real-world scenario. In extensive and rigorous benchmark experiments, PortalCG considerably outperformed state-of-the-art techniques of machine learning and protein-ligand docking when applied to dark gene families, and demonstrated its generalization power for off-target predictions and compound screenings under out-of-distribution (OOD) scenarios. Furthermore, in an external validation for the multi-target compound screening, the performance of PortalCG surpassed the human design. Our results also suggested that a differentiable sequence-structure-function deep learning framework where protein structure information serve as an intermediate layer could be superior to conventional methodology where the use of predicted protein structures for predicting protein functions from sequences. We applied PortalCG to two case studies to exemplify its potential in drug discovery: designing selective dual-antagonists of Dopamine receptors for the treatment of Opioid Use Disorder, and illuminating the undruggable human genome for targeting diseases that do not have effective and safe therapeutics. Our results suggested that PortalCG is a viable solution to the OOD problem in exploring the understudied protein functional space. Author SummaryMany complex diseases such as Alzheimers disease, mental disorders, and substance use disorders do not have effective and safe therapeutics due to the polygenic nature of diseases and the lack of thoroughly validate drug targets and their ligands. Identifying small molecule ligands for all proteins encoded in the human genome will provide new opportunity for drug discovery of currently untreatable diseases. However, the small molecule ligand of more than 90% gene families is completely unknown. Existing protein-ligand docking and machine learning methods often fail when the protein of interest is dissimilar to those with known functions or structures. We develop a new deep learning framework PortalCG for efficiently and accurately predicting ligands of understudied proteins which are out of reach of existing methods. Our method achieves unprecedented accuracy over state-of-the-arts by incorporating ligand binding site information and sequence-to-structure-to-function paradigm into a novel deep meta-learning algorithms. In a case study, the performance of PortalCG surpassed the human design. The proposed computational framework will shed new light into how chemicals modulate biological system as demonstrated by applications to drug repurposing and designing polypharmacology. It will open a new door to developing effective and safe therapeutics for currently incurable diseases. PortalCG can be extended to other scientific inquiries such as predicting protein-protein interactions and protein-nucleic acid recognition.

systems biology↗

bioRxiv2022

Spatial organization of the 3D genome encodes gene co-expression programs in single cells

Deconstructing the mechanism by which the 3D genome encodes genetic information to generate diverse cell types during animal development is a major challenge in biology. The contrast between the elimination of chromatin loops and domains upon Cohesin loss and the lack of downstream gene expression changes at the cell population level instigates intense debates regarding the structure-function relationship between genome organization and gene regulation. Here, by analyzing single cells after acute Cohesin removal with sequencing and spatial genome imaging techniques, we discover that, instead of dictating population-wide gene expression levels, 3D genome topology mediated by Cohesin safeguards long-range gene co-expression correlations in single cells. Notably, Cohesin loss induces gene co-activation and chromatin co-opening between active domains in cis up to tens of megabase apart, far beyond the typical length scale of enhancer-promoter communication. In addition, Cohesin separates Mediator protein hubs, prevents active genes in cis from localizing into shared hubs and blocks intersegment transfer of diverse transcriptional regulators. Together, these results support that spatial organization of the 3D genome orchestrates dynamic long-range gene and chromatin co-regulation in single living cells.

genomics↗

bioRxiv2022

TPL2 kinase activity regulates microglial inflammatory responses and promotes neurodegeneration in tauopathy mice

TPL2 (MAP3K8) is a central signaling node in the inflammatory response of peripheral immune cells. We find that TPL2 kinase activity modulates microglial cytokine release and is required for microglia-mediated neuron death in vitro. In acute in vivo neuroinflammation settings, TPL2 kinase activity regulates microglia activation states and brain cytokine levels. In a tauopathy model of chronic neurodegeneration, loss of TPL2 kinase activity reduces neuroinflammation and rescues synapse loss, brain volume loss, and behavioral deficits. Single-cell RNAseq analysis indicates protection in the tauopathy model was associated with reductions in activated microglia subpopulations as well as infiltrating peripheral immune cells. Overall, using various models, we find that TPL2 kinase activity can promote multiple harmful consequences of microglial activation in the brain including cytokine release, iNOS induction, astrocyte activation, and immune cell infiltration. Consequently, inhibiting TPL2 kinase activity could represent a potential therapeutic strategy in neurodegenerative conditions.

neuroscience↗

bioRxiv2022

A Peronophythora litchii RXLR effector PlAvh202 enhances plant susceptibility by destabilizing host SAMS

Oomycete pathogens can secret hundreds of effectors into plant cells to interfere with plant immune system during infection. Here we identified a cytoplasmic RXLR effector from the most destructive pathogen of litchi, Peronophythora litchii, and named it as PlAvh202. PlAvh202 could suppress cell death triggered by INF1 and Avr3a/R3a in Nicotiana benthamiana, and was essential for P. litchii virulence. In addition, PlAvh202 also suppressed plant immune responses and promoted the susceptibility of N. benthamiana to Phytophthora capsici. Further research revealed that PlAvh202 could suppress ethylene (ET) production by targeting and destabilizing plant S-adenosyl-L-methionine synthetase (SAMS), a key enzyme in ET biosynthesis pathway, in a 26S proteasome-dependent manner. Transient expression of LcSAMS3 induced ET production and enhanced plant resistance, whereas inhibition of ET biosynthesis promoted P. litchii infection, supporting that LcSAMS and ET positively regulate litchi immunity towards P. litchii. Overall, these findings highlight that SAMS can be targeted by oomycete RXLR effector to manipulate ET-mediated plant immunity.

plant biology↗

bioRxiv2022

Identification of unique α4 chain structure and conserved anti-angiogenic activity of α3NC1 type IV collagen in zebrafish

Type IV collagen is an abundant component of basement membranes in all multicellular species and is essential for the extracellular scaffold supporting tissue architecture and function. Lower organisms typically have two type IV collagen genes, encoding 1 and 2 chains, in contrast with the six genes in humans, encoding 1 to 6 chains. The chains assemble into trimeric protomers, the building blocks of the type IV collagen network. The detailed evolutionary conservation of type IV collagen network remains to be studied. We report on the molecular evolution of type IV collagen genes. The zebrafish 4 non-collagenous (NC1) domain, in contrast with its human ortholog, contains an additional cysteine residue and lacks the M93 and K211 residues involved in sulfilimine bond formation between adjacent protomers. This may alter 4 chain interactions with other chains, as supported by temporal and anatomic expression patterns of collagen IV chains during zebrafish development. Despite the divergence between zebrafish and human 3 NC1 domain (endogenous angiogenesis inhibitor, Tumstatin), the zebrafish 3 NC1 domain exhibits conserved anti-angiogenic activity in human endothelial cells. Our work supports type IV collagen is largely conserved between zebrafish and humans, with a possible difference involving the 4 chain.

developmental biology↗

bioRxiv2022

Loss of aquaporin-4 results in glymphatic system dysfunction via brain-wide interstitial fluid stagnation

The glymphatic system is a fluid transport network of cerebrospinal fluid (CSF) entering the brain along arterial perivascular spaces, exchanging with interstitial fluid (ISF), ultimately establishing directional clearance of interstitial solutes. CSF transport is facilitated by the expression of aquaporin-4 (AQP4) water channels on the perivascular endfeet of astrocytes. Mice with genetic deletion of AQP4 (AQP4 KO) exhibit abnormalities in the brain structure and molecular water transport. Yet, no studies have systematically examined how these abnormalities in structure and water transport correlate with glymphatic function. Here we used high-resolution 3D magnetic resonance (MR) non-contrast cisternography, diffusion-weighted MR imaging (MR-DWI) along with intravoxel-incoherent motion (IVIM) DWI, while evaluating glymphatic function using a standard dynamic contrast-enhanced MR imaging to better understand how water transport and glymphatic function is disrupted after genetic deletion of AQP4. AQP4 KO mice had larger interstitial spaces and total brain volumes resulting in higher water content and reduced CSF space volumes, despite similar CSF production rates and vascular density compared to wildtype mice. The larger interstitial fluid volume likely resulted in increased slow but not fast MR diffusion scores and coincided with reduced glymphatic influx. This markedly altered brain fluid transport in AQP4 KO mice may result from a reduction in glymphatic clearance, leading to stagnation of ISF movement and enlargement of the interstitial space. Overall, diffusion MR is a useful tool to evaluate glymphatic function and may serve as valuable translational biomarker to study glymphatics in human disease.

neuroscience↗

bioRxiv2022

Fatty acid binding protein 1 and fatty acid synthetase overexpression have differential effects on collagen III and cross-linking in Zongdihua pig tissues

The purpose of this study was to determine whether FABP1 and FAS regulate expression of collagen and its crosslinking via lysyl oxidase in Zongdihua pigs. We wished to identify processes affecting meat quality using molecular genetics to provide a basis for breeding improvement of these animals. We measured expression levels of FABP1 and related genes using qRT-PCR in longissimus dorsi muscle and subcutaneous adipose tissues. Primary adipocytes from fat tissues were isolated and FABP1 and FAS were overexpressed from recombinant plasmids. Our sequence analysis of the cloned genes indicated that FABP1 gene encodes a hydrophobic protein of 128 amino acids and contained 12 predicted phosphorylation sites and no transmembrane region. FAS encodes 333 amino acid hydrophobic protein containing with 26 phosphorylation sites and 0 transmembrane regions. The basal levels of FABP1 and FAS in pig tissues expression were 3 -3.5-fold higher in subcutaneous fat compared with muscle (P < 0.01). Recombinant expression plasmids were successfully transfected into the cloned preadipocytes and (a) overexpression of FAS resulted in significantly increased expression of COL3A1 gene (P < 0.05) and significantly inhibited lysyl oxidase LOX expression (P < 0.01); (b) overexpression of FABP1 significantly increased COL3A1 expression (P < 0.01) and significantly inhibited LOX expression (P < 0.05) and significantly reduced lysyl oxidase activity (P < 0.01). Therefore, enhancing FABP1 expression increases collagen accumulation and this preliminarily suggests that FAS and FABP1 can serve as fat-related candidate genes providing a theoretical basis for the study of fat deposition in Zongdihua pigs.

animal behavior and cognition↗

bioRxiv2022

Deep multi-omic network fusion for marker discovery of Alzheimer's Disease

Multi-omic data spanning from genotype, gene expression to protein expression have been increasingly explored to interpret findings from genome wide association studies of Alzheimers disease (AD) and to gain more insight of the disease mechanism. However, each -omics data type is usually examined individually and the functional interactions between genetic variations, genes and proteins are only used after discovery to interpret the findings, but not beforehand. In this case, multi-omic findings are likely not functionally related and therefore give rise to challenges in interpretation. To address this problem, we propose a new interpretable deep neural network model MoFNet to jointly model the prior knowledge of functional interactions and multi-omic data set. It aims to identify a subnetwork of functional interactions predictive of AD evidenced by multi-omic measures. Particularly, prior functional interaction network was embedded into the architecture of MoFNet in a way that it resembles the information flow from DNA to gene and protein. The proposed model MoFNet significantly outperformed all other state-of-art classifiers when evaluated using multi-omic data from the ROS/MAP cohort. Instead of individual markers, MoFNet yielded multi-omic sub-networks related to innate immune system, clearance of misfolded proteins, and neurotransmitter release respectively. Around 50% of these findings were replicated in another independent cohort. Our identified gene/proteins are highly related to synaptic vesicle function. Altered regulation or expression of these genes/proteins could cause disruption in neuron-neuron or neuron-glia cross talk and further lead to neuronal and synapse loss in AD. Further investigation of these identified genes/proteins could possibly help decipher the mechanisms underlying synaptic dysfunction in AD, and ultimately inform therapeutic strategies to modify AD progression in the early stage.

systems biology↗

bioRxiv2022

Nonlytic cellular release of hepatitis A virus requires dual capsid recruitment of the ESCRT-associated Bro1 domain proteins HD-PTP and ALIX

Although picornaviruses are conventionally considered nonenveloped, members of multiple picornaviral genera are released nonlytically from infected cells in extracellular vesicles. The mechanisms underlying this process are poorly understood. Here, we describe interactions of the hepatitis A virus (HAV) capsid with components of host endosomal sorting complexes required for transport (ESCRT) that play an essential role in release. We show release of quasi-enveloped virus (eHAV) in exosome-like vesicles requires a conserved export signal located within the 8 kDa C-terminal VP1 pX extension that functions in a manner analogous to late domains of canonical enveloped viruses. Fusing pX to a self-assembling engineered protein nanocage (EPN-pX) resulted in its ESCRT-dependent release in extracellular vesicles. Mutational analysis identified a 24 amino acid peptide sequence located within the center of pX that was both necessary and sufficient for nanocage release. Deleting a YxxL motif within this sequence ablated eHAV release, resulting in virus accumulating intracellularly. The pX export signal is conserved in non-human hepatoviruses from a wide range of mammalian species, and functional in pX sequences from bat hepatoviruses when fused to the nanocage protein, suggesting these viruses are released as quasi-enveloped virions. Quantitative proteomics identified multiple ESCRT-related proteins associating with EPN-pX, including ALG2-interacting protein X (ALIX), and its paralog, tyrosine-protein phosphatase non-receptor type 23 (HD-PTP), a second Bro1 domain protein linked to sorting of ubiquitylated cargo into multivesicular endosomes. RNAi-mediated depletion of either Bro1 domain protein impeded eHAV release. Super-resolution fluorescence microscopy demonstrated colocalization of viral capsids with endogenous ALIX and HD-PTP. Co-immunoprecipitation assays using biotin-tagged peptides and recombinant proteins revealed pX interacts directly through the export signal with N-terminal Bro1 domains of both HD-PTP and ALIX. Our study identifies an exceptionally potent viral export signal mediating extracellular release of virus-sized protein assemblies and shows release requires non-redundant activities of both HD-PTP and ALIX. Authors SummaryMechanisms underlying nonlytic release of canonical nonenveloped viruses from infected cells are poorly understood. We show here that release of hepatitis A virus from cells in exosome-like vesicles requires nonredundant activities of two distinct Bro1-domain proteins associated with host cell machinery (ESCRT) for endosomal sorting, HD-PTP and ALIX. We demonstrate both Bro1 domain proteins are recruited to the viral capsid by the pX segment of the 1D capsid protein, and that they act in a non-redundant manner to mediate virus release. Fusing pX to a self-assembling nanocage protein resulted in ESCRT-dependent release mediated by a short pX peptide sequence conserved in hepatoviruses from bats to humans. Mutations within the pX sequence ablate release and result in noncytolytic virus accumulating intracellularly. Our study identifies an exceptionally potent viral export signal mediating extracellular release of virus-sized protein assemblies and shows nonlytic release of quasi-enveloped virus is an ancient evolutionary trait of hepatoviruses.

microbiology↗

bioRxiv2022

Prostaglandin D2 pathway in a transgenic rat model of Alzheimer's disease: therapeutic potential of timapiprant a DP2 antagonist

The cyclooxygenase pathway, a key mediator of inflammation, is implicated in Alzheimers disease (AD). A deeper investigation is required into the contributions of this pathway to the neuropathology of AD. Cyclooxygenases produce prostaglandins, which have multiple receptors and functions including inflammation, nociception, sleep, cardiovascular maintenance and reproduction. In the brain, prostaglandin D2 (PGD2) is the most abundant prostaglandin, increases the most under pathological conditions, and plays roles in sleep, stroke and inflammation. PGD2 signals through its DP1 and DP2 receptors and their activation can be protective or detrimental. We address the relationship between the PGD2 pathway and AD neuropathology with F344-AD transgenic (Tg-AD) rats that exhibit age-dependent and progressive pathology similar to AD patients. We analyzed the PGD2 pathway in the hippocampus of wild type (WT) rats and their Tg-AD littermates, at the age of 11 months, when Tg-AD rats exhibit plaques and perform significantly worse in hippocampal-dependent cognitive tasks than WT rats. Using mass spectrometry, we determined that PGD2 levels were at least 14.5-fold higher than PGE2, independently of genotype. Immunohistochemistry established that microglial DP1 receptors were more abundant and neuronal DP2 receptors were fewer in Tg-AD than in WT rats. RNA sequencing profiling of 33 genes involved in the PGD2 and PGE2 pathways revealed that mRNA levels were the highest for L-PGDS, the major PGD2 synthase in the brain. To evaluate the pathophysiological significance of our findings on the PGD2 pathway, we treated a subset of rats (WT and Tg-AD males) with timapiprant, a potent and highly selective oral DP2 antagonist being developed as a once-daily oral treatment in patients with allergic inflammation. We conclusively show that timapiprant significantly mitigated some of the AD pathology exhibited by the Tg-AD male rats. More comprehensive studies are necessary to support the therapeutic potential of timapiprant and that of other PGD2-related compounds in the treatment of AD.

neuroscience↗

bioRxiv2022

Microtubule-Connexin-43 regulation suppresses arrhythmias and fibrosis in Duchenne muscular dystrophy mice.

Dilated cardiomyopathy is the leading cause of death in Duchenne muscular dystrophy (DMD) patients due to advancements in skeletal muscle therapies yet limited presence of cardiac treatments. The phosphorylation status of gap junction protein Connexin-43 (Cx43) drives Cx43 remodeling and the development of arrhythmias and fibrosis. Based on evidence that Colchicine drug treatment improves Cx43 phosphorylation and remodeling, we compared the microtubule cytoskeleton in DMD mice (mdx) versus mdx mice genetically altered to be Cx43-phosphorylation-deficient (mdxS3A). Reciprocally, we analyzed the microtubule cytoskeleton in mdx mice genetically altered to be Cx43-phospho-mimicking (mdxS3E). We found a link between the phospho-status of Connexin-43 and regulation of microtubule organization, in which phospho-dead Cx43 (S3A) inhibits improvements seen with Colchicine treatment in mdx mice, and phospho-mimic S3E promotes microtubule reorganization in mdx mice. A reduction in arrhythmias and fibrosis suggests an unsuspecting Cx43-microtubule link for translational corrective activities for DMD cardiomyopathy.

cell biology↗

bioRxiv2022

Females outperform males in spatial learning despite increased amyloid plaques and microgliosis in a TgF344-AD rat model of Alzheimer's disease

Alzheimers disease (AD) is a progressive neurodegenerative disease and is the sixth leading cause of death in the US. AD is more prevalent in females than males. While estrogen provides neuroprotection in females, sex mediated differences in the development of AD pathology are not fully elucidated. Therefore, a comparison of the events that develop between sexes in the early-stage of AD pathology may reveal new potential targets for more effective therapeutic intervention. To address sex differences, we analyzed early stage 9-month male and female TgF344-AD (Tg-AD) rats, an AD model carrying the APPswe and Presenilin 1 (PS1{Delta}E9) mutations that develops progressive age-dependent AD pathology similar to humans. Using active place avoidance (aPAT) tests that assess hippocampal-dependent spatial learning and memory, we found significant deficits in Tg-AD females compared to wild type females, but no significant difference between the two male genotypes. Moreover, significant sex differences were observed in that Tg-AD females outperformed Tg-AD males in several measures of the aPAT test. Unexpectedly, Tg-AD females displayed higher levels of hippocampal amyloid plaques and amoeboid microglia than their Tg-AD male littermates. Furthermore, Tg-AD females experienced less hippocampal neuronal loss and had higher GluA2 subunit levels than Tg-AD males. Based on our findings, we propose that estrogen may protect females against cognitive impairment at early stages of AD by regulating GluA2 levels independently of amyloid plaque deposition and gliosis. Elucidating this potential protective mechanism of action of estrogen in AD could lead to new targets for early intervention.

neuroscience↗

bioRxiv2022

A snapshot of the root phenotyping landscape in 2021

Root phenotyping describes methods for measuring root properties, or traits. While root phenotyping can be challenging, it is advancing quickly. In order for the field to move forward, it is essential to understand the current state and challenges of root phenotyping, as well as the pressing needs of the root biology community. In this letter, we present and discuss the results of a survey that was created and disseminated by members of the Graduate Student and Postdoc Ambassador Program at the 11th symposium of the International Society of Root Research. This survey aimed to (1) provide an overview of the objectives, biological models and methodological approaches used in root phenotyping studies, and (2) identify the main limitations currently faced by plant scientists with regard to root phenotyping. Our survey highlighted that (1) monocotyledonous crops dominate the root phenotyping landscape, (2) root phenotyping is mainly used to quantify morphological and architectural root traits, (3) 2D root scanning/imaging is the most widely used root phenotyping technique, (4) time-consuming tasks are an important barrier to root phenotyping, (5) there is a need for standardised, high-throughput methods to sample and phenotype roots, particularly under field conditions, and to improve our understanding of trait-function relationships.

scientific communication and education↗

bioRxiv2022

FAME: Fragment-based Conditional Molecular Generation for Phenotypic Drug Discovery

De novo molecular design is a key challenge in drug discovery due to the complexity of chemical space. With the availability of molecular datasets and advances in machine learning, many deep generative models are proposed for generating novel molecules with desired properties. However, most of the existing models focus only on molecular distribution learning and target-based molecular design, thereby hindering their potentials in real-world applications. In drug discovery, phenotypic molecular design has advantages over target-based molecular design, especially in first-in-class drug discovery. In this work, we propose the first deep graph generative model (FAME) targeting phenotypic molecular design, in particular gene expression-based molecular design. FAME leverages a conditional variational autoencoder framework to learn the conditional distribution generating molecules from gene expression profiles. However, this distribution is difficult to learn due to the complexity of the molecular space and the noisy phenomenon in gene expression data. To tackle these issues, a gene expression denoising (GED) model that employs contrastive objective function is first proposed to reduce noise from gene expression data. FAME is then designed to treat molecules as the sequences of fragments and learn to generate these fragments in autoregressive manner. By leveraging this fragment-based generation strategy and the denoised gene expression profiles, FAME can generate novel molecules with a high validity rate and desired biological activity. The experimental results show that FAME outperforms existing methods including both SMILES-based and graph-based deep generative models for phenotypic molecular design. Furthermore, the effective mechanism for reducing noise in gene expression data proposed in our study can be applied to omics data modeling in general for facilitating phenotypic drug discovery.

bioinformatics↗

bioRxiv2021

Chemical-induced Gene Expression Ranking and its Application to Pancreatic Cancer Drug Repurposing

Chemical-induced gene expression profiles provide critical information on the mode of action, off-target effect, and cellar heterogeneity of chemical actions in a biological system, thus offer new opportunities for drug discovery, system pharmacology, and precision medicine. Despite their successful applications in drug repurposing, large-scale analysis that leverages these profiles is limited by sparseness and low throughput of the data. Several methods have been proposed to predict missing values in gene expression data. However, most of them focused on imputation and classification settings which have limited applications to real-world scenarios of drug discovery. Therefore, a new deep learning framework named chemical-induced gene expression ranking (CIGER) is proposed to target a more realistic but more challenging setting in which the model predicts the rankings of genes in the whole gene expression profiles induced by de novo chemicals. The experimental results show that CIGER significantly outperforms existing methods in both ranking and classification metrics for this prediction task. Furthermore, a new drug screening pipeline based on CIGER is proposed to select approved or investigational drugs for the potential treatments of pancreatic cancer. Our predictions have been validated by experiments, thereby showing the effectiveness of CIGER for phenotypic compound screening of precision drug discovery in practice.

bioinformatics↗

bioRxiv2021

FoQDE2-dependent milRNA promotes Fusarium oxysporum f. sp. cubense virulence by targeting a glycosyl hydrolase coding gene at transcriptional level

MicroRNAs (miRNAs) are small non-coding RNAs that regulate protein-coding gene expression primarily found in plants and animals. Fungi produce microRNA-like RNAs (milRNAs) that are structurally similar to miRNAs and functionally important in various biological processes. The fungus Fusarium oxysporum f. sp. cubense (Foc) is the causal agent of Panama disease that threatens global banana production. It remains uncharacterized about the biosynthesis and functions of milRNAs in Foc. In this study, we investigated the biological function of milRNAs contributing to Foc pathogenesis. Within 24 hours post infecting the host, the Argonaute coding gene FoQDE2, and two Dicer coding genes FoDCL1 and FoDCL2, all of which are involved in milRNA biosynthesis, were significantly induced. FoQDE2 deletion mutant exhibited decreased virulence and hypersensitivity to hydrogen peroxide (H2O2). These results indicate that milRNA biosynthesis is crucial for Foc pathogenesis. By small RNA sequencing, we identified 364 small RNA-producing loci in the Foc genome, 25 of which were significantly downregulated in the FoQDE2 deletion mutant, from which milR-87 was verified as a FoQDE2-depedent milRNA based on qRT-PCR analysis. Through deletion and overexpression of milR-87 in the wild-type Foc strain, functions of milR-87 were studied. The results showed that milR-87 is crucial for Foc virulence in infection process. We furthermore identified a glycosyl hydrolase-coding gene, FOIG_15013, as the direct target of milR-87. The FOIG_15013 deletion mutant displayed a dramatic increase in the growth, conidiation and virulence. Transient expression of FOIG_15013 in Nicotiana benthamiana leaves activates the host defense responses. Collectively, this study documents the involvement of milRNAs in the manifestation of the devastating fungal disease in banana, and demonstrates the importance of milRNAs in the pathogenesis and other biological processes. Further analyses of the biosynthesis and expression regulation of fungal milRNAs may offer a novel strategy to combat devastating fungal diseases. Author summaryThe fungus Fusarium oxysporum f. sp. cubense (Foc) is the causal agent of Panama disease that threatens global banana production. As a typical representative of F. oxysporum species complex, the pathogen has been widely concerned. However, pathogenesis of Foc is not fully elucidated. In particular, pathogenic regulatory mechanism of the microRNA like small RNAs (milRNAs) found in Foc is unknown. Here, we found that FoQDE2, one Argonaute coding gene, and two Dicer coding genes FoDCL1 and FoDCL2, which are involved in milRNA biosynthesis, are significantly induced during the early infection stage of Foc. The results suggested that the milRNAs biosynthesis mediated by these genes may play an active role in the infection process of Foc. Based on this assumption, we subsequently found a FoQDE2-dependent milRNA (milR-87) and identified its target gene. Functional analysis showed that FoQDE2, miR-87 and its target gene were involved in the pathogenicity of Foc in different degree. The studies help us gain insight into the pathogenesis with FoQDE2, milR-87, and its target gene as central axis in Foc. The identified pathogenicity-involved milRNA provides an active target for developing novel and efficient biocontrol agents against Panama disease.

microbiology↗

bioRxiv2021

Directional selection at gene expression level contributes to the speciation of Asian rice cultivars

Differences in expression levels play important roles in phenotypic variation across species, especially those closely related species with limited genomic differences. Therefore, studying gene evolution at expression level is important for illustrating phenotypic differentiation between species, such as the two Asian rice cultivars, Oryza sativa L. ssp. indica and Oryza sativa L. ssp. japonica. In this study, we evaluated the gene expression variation at inter-subspecies and intra-subspecies level using transcriptome data from seedlings of three indica and japonica rice and defined four groups of genes under different natural selections. We found a substantial of genes (about 79%) that are under stabilizing selection at the expression level in both subspecies, while about 16% of genes are under directional selection. Genes under directional selection have higher expression level and lower expression variation than those under stabilizing selection, which suggest a potential explanation to subspecies adaptation to different environments and interspecific phenotypic differences. Subsequent functional enrichment analysis of genes under directional selection shows that indica rice have experienced the adaptation to environmental stresses, and also show differences in biosynthesis and metabolism pathways. Our study provide an avenue of investigating indica-japonica differentiation through gene expression variation, which may guide to rice breeding and yield improvement.

plant biology↗