Biology subjects

Bai, Y.

Publications and source records attributed to Bai, Y..

At least 91 records · Page 5Linked to original sources

bioRxiv2024

An Investigation into Mechanomyography for Signal Extraction and Classification of Human Lower Limb Activity

To mitigate the difficulties associated with the extraction of Mechanomyography (MMG) signals from raw Accelerometer (ACC) data and the subsequent classification of human lower limb activities based on MMG signals, the Feature Mode Decomposition (FMD) algorithm has been utilized for the isolation of the MMG signal. Simultaneously, surface Electromyography (sEMG) signals were recorded to perform correlation analyses, thereby validating the effectiveness of the extracted Mechanomyography (MMG) signals. The results demonstrate that the envelope entropy derived from the FMD was the lowest among the observed values, and the composite signal obtained via FMD displayed the highest correlation with the sEMG signal. This indicates that FMD is capable of efficiently isolating the MMG signal while maintaining the maximal quantity of muscle contraction data. To address the challenge of classifying human lower limb activities, a comprehensive feature extraction procedure was implemented, resulting in the derivation of 448 unique features from multi-channel mechanomyography (MMG) signals. Subsequently, Kernel Principal Component Analysis (KPCA) was employed to diminish the feature sets dimensionality. This was succeeded by the deployment of a Temporal Convolutional Network integrated with an Attention mechanism (TCN-Attention) to train the classification model. Additionally, an enhanced Northern Goshawk Optimization Algorithm was leveraged for optimization purposes. The findings indicate that FMD exhibited the minimum envelope entropy value of 8.13, concurrently attaining the maximum correlation coefficient of 0.87 between MMG and sEMG signals. Significantly, the SCNGO-TCN-Attention model demonstrated superior classification accuracy, attaining an exceptional accuracy rate of 98.44%.

bioengineering↗

bioRxiv2024

Glycophagy is an ancient bilaterian pathway supporting metabolic adaptation through STBD1 structural evolution

The selective autophagy of glycogen (glycophagy) has recently emerged as being crucial to glucose homeostasis in vertebrates, yet its origins remain elusive. Here, we provide evidence that starch-binding domain-containing protein 1 (STBD1), the key glycophagy receptor in vertebrates, is functionally conserved in the Pacific oyster, revealing its conserved position within ancient autophagy networks. We show that STBD1 in oysters - as seen in other invertebrate groups - possesses an N-terminal carbohydrate binding module family 20 (CBM20) domain, representing the ancestral state for this protein, while a shuffling of CBM20 to the C- terminus occurred during early chordate evolution. Structural modelling and functional studies reveal that the N-terminal CBM20 organization of STBD1 enhances glycogen binding. Functional experiments demonstrate that an STBD1-glycogen complex, anchored by GABARAPL2, facilitates an increased glycogen flux into autophagosomes for lysosomal degradation. We conclude that glycophagy is deeply conserved in bilaterians and that STBD1 structural evolution underlies potentially adaptive variation in metabolic strategies across distinct animal clades.

evolutionary biology↗

bioRxiv2024

Switch of TIR signaling by a Ca2+ sensor activates ADR1 recognition of pRib-AMP-EDS1-PAD4 for stomatal immunity

Plants swiftly close stomata upon detecting pathogen entry, a crucial defense termed stomatal immunity. The process is initiated by cell-surface pattern recognition receptors (PRRs) that perceive pathogen-associated molecular patterns (PAMPs) and evoke a series of early cellular responses including calcium ions (Ca2+) influx, and is conducted by the intracellular nucleotide-binding leucine-rich-repeat receptors (NLRs) ADR1s within an EDS1-PAD4-ADR1 module. However, the underlying mechanisms linking PRR signaling to the NLRs ADR1s remain unclear. Here, we show that the Nicotiana benthamiana Toll/interleukin-1 receptor (TIR)-only protein Stomatal TIR1 (STIR1) produces the immune molecule pRib-AMP, induces formation of EDS1-PAD4-ADR1 complexes, and mediates stomatal immunity. The Inhibitor of Stomatal Immunity C2-domain protein 1 (ISIC1) interacts with and constrains STIR1 function at basal condition, whereas upon pathogen infection, ISIC1 senses Ca2+ signals and de-represses STIR1 signaling. Cryo-electron microscopy structure of pathogen infection-elicited Arabidopsis AtEDS1-AtPAD4-AtADR1-L2 complex reveals the pRib-AMP binding to AtEDS1-AtPAD4 receptor and the AtADR1-L2 recognition of pRib-AMP-AtPAD4-AtEDS1 for stomatal immunity. Collectively, this study uncovers a repression/de-repression mechanism linking PRR signaling to NLRs by a Ca2+ sensor/TIR-only node, and elucidates an NLR recognition mechanism of the pRib-AMP-EDS1-PAD4 complex in governing innate immunity. SynopsisAt basal condition, the Ca2+ sensor ISIC1 interacts with and inhibits the TIR-only protein STIR1; upon pathogen infection, ISIC1 perceives Ca2+ signal and releases STIR1 to produce pRib-AMP; the EDS1-PAD4 receptor binds pRib-AMP and is recognized by the NLR ADR1-L2, thereby activating stomatal immunity.

plant biology↗

bioRxiv2024

Comparative spatiotemporal single cell transcriptomes reveal rewiring of pre-existing regulations during emergence of Kranz anatomy in C4 grasses

Many of the worlds most productive food and bioenergy crops use C4 photosynthesis, which have high photosynthetic efficiency due to a Kranz anatomy-based CO2 concentrating mechanism. Here, we took a comparative transcriptomics approach using single cell spatial transcriptomes of leaf primordia for maize (Zea mays) and single cell RNA-seq (scRNA-seq) atlases of corresponding leaf tissues for three C4 species (Zea mays, Sorghum bicolor, Setaria viridis) and one C3 rice (Oryza sativa) to study regulatory networks involved in the development and evolution of Kranz anatomy. We show that the formation of Kranz anatomy involves extensive recruitment and modification of pre-existing regulatory modules, especially the SHR-SCR module and the auxin signaling pathway. Members of the INDETERMINATE DOMAIN (IDD) family transcription factors, including IDD7 and IDDP1, contribute to the modification of the involved modules. These extensive recruitment and modification of existing genetic regulatory programs underlie the repeated emergence of C4 photosynthesis during evolution.

plant biology↗

bioRxiv2024

Strong segregation promotes self-destructive cooperation

Self-destructive cooperator (SDC), where individuals sacrifice themselves for others, poses significant personal costs yet is widespread in nature. Traditional group selection theory struggles to explain its persistence, as the extreme cost of self-destruction typically exceeds the benefit to these individuals. Here, we predict that this behavior can endure in structured environments with strong segregation, where populations are divided into homogenous groups originating from one or two founding members. In such contexts, the benefits derived from SDC are primarily confined within the homogenous SDC groups, thus preserving the value of this altruistic sacrifice and ensuring its maintenance. To validate our hypothesis, we employ a synthetic system with engineered bacteria to mimic both SDC and non-sacrificing cheaters. We then conduct automated experiments facilitated by biofoundry technology to monitor and operate the subgroups with diverse growth behaviors stemming from a restricted number of initial cells because of strong segregation. Eventually, we demonstrate that SDC indeed maintains under strong segregation, and high stress benefits SDC because it reduces the benefits received by cheaters in heterogenous subgroups. This study extends the group selection theory to encompass even the most extreme manifestations of altruism and highlights the potential of automation in evolutionary research.

microbiology↗

bioRxiv2024

A global resource constrained model to predict metabolic flux dynamics in fluctuating environments

Environmental changes often induce global variations on bacterial gene expression, frequently accompanied by alterations in growth rates. Integrating non-metabolic cellular processes, such as gene expression and macromolecule synthesis, with metabolic modeling remains a significant challenge in systems biology due to the lack of mechanistic representation of gene regulation. Here, we introduce a novel constraint-based modeling framework called dynamic Constrained Allocation Flux Balance Analysis (dCAFBA), which comprehensively integrates metabolism, cellular resource allocation, and gene regulation. We employ a quasi-steady-state assumption, positing that reaction fluxes achieve balance at each time step, adapting more rapidly than protein synthesis and growth dilution. This approach enables the prediction of reaction flux dynamics and protein allocation necessary to achieve cellular objectives, such as optimizing growth during various growth shifts (including carbon, amino acid, and transcriptional signal), without detailing molecular gene regulations. Our model offers a new method for interpreting proteome allocation and cellular metabolism in complex and transient environments, providing mechanistic insights valuable for metabolic engineering.

systems biology↗

bioRxiv2024

Identification of Cellular Signatures Associated with Chinese Hamster Ovary (CHO) Cell Adaptation for Secretion of Antibodies

The secretory capacity of Chinese hamster ovary (CHO) cells remains a fundamental bottleneck in the manufacturing of protein-based therapeutics. Unconventional biological drugs with complex structures and processing requirements are particularly problematic. Although engineered vector DNA elements can achieve rapid and high-level therapeutic protein production, a high metabolic and protein folding burden is imposed on the host cell. Cellular adaptations to these conditions include differential gene expression profiles that can in turn influence the productivity and quality control of recombinant proteins. In this study, we used quantitative transcriptomics and proteomics analyses to investigate how biological pathways change with antibody titre. Gene and protein expression profiles of CHO pools and clones producing a panel of different monoclonal and bispecific antibodies were analysed during fed-batch production. Antibody-expressing CHO pools were heterogeneous, resulting in few discernible genetic signatures. Clonal lines derived from these pools, selected for high and low production, yielded a small number of differentially expressed proteins that correlated with productivity and were shared across biotherapeutics. However, the dominant feature associated with higher protein production was transgene copy number and resulting mRNA expression level. Moreover, variability between clones suggested that the process of cellular adaptation is variable with diverse cellular changes associated with individual adaptation events.

bioengineering↗

bioRxiv2024

A Versatile Light Field Microscopy Platform for Multi-purpose Dynamic Volumetric Bio-imaging

Light field microscopy (LFM) has emerged in recent years as a unique solution for fast, scan-free volumetric imaging of dynamic biological samples. This is achieved by using a microlens array in the detection path to record both the lateral and angular information of the light fields coming from the sample, capturing a 3-dimensional (3D) volume in a single 2-dimensional (2D) snapshot. In post-acquisition, the 3D sample volume is computationally reconstructed from the recorded 2D images, thus enabling unprecedented 3D capture speed, not limited by the typical constraint of physically scanning the focal plane over the sample volume. Up to date, most published LFM imaging setups have been specialized single-purpose platforms, optimized for a narrow performance window in field of view and resolution, thus hampering widespread adoption of LFM for biomedical research. Here, we present a versatile LFM platform for fast 3D imaging across multiple scales, enabling applications from cell to system-level biology on the same imaging setup. Our multiscale LFM is built as an add-on module to a conventional commercially available wide field microscope, and the various imaging applications, with different ranges of field of view and resolution, are achieved by simply switching between the standard microscope objectives available on the wide field microscope. Importantly, we provide an open-source end-to-end software package for calculating the system performance parameters, processing the experimentally measured point spread function, and light field 3D image reconstruction. We demonstrate the performance of our multiscale LFM platform through imaging the whole-brain activity map of seizures in larval zebrafish, calcium dynamics in ex vivo mouse pancreatic islets, and subcellular protein dynamics in cultured cells.

bioengineering↗

bioRxiv2024

Neural representation of vocalization self-monitoring in the bat auditory midbrain

During acoustic interaction, mammals, including humans, not only monitor sounds from external sources but also their own vocalizations, the latter of which is known as vocalization feedback monitoring. Yet, it is unclear whether subcortical auditory regions process vocalization feedback. Here, we established an experimental paradigm that allows recording chronically the single-unit activity of the auditory midbrain (inferior colliculus) in freely vocalizing bats. We report that most collicular neurons represented self-produced biosonar vocalizations distinctively from pure tones and vocalization playbacks. Some neurons showed robust excitatory responses to vocalization feedback yet lacked any excitatory response to pure tones covering the entire frequency range of the vocalizations. Surprisingly, some collicular neurons even reversed the response polarity, from vocalization-induced suppression to vocalization-playback-induced excitation. Moreover, approximately a third of the neurons responded faster to self-produced vocalizations than to both vocalization playbacks and pure tones. These findings show that the midbrain inferior colliculus is involved in vocalization feedback processing, an ability suggested to arise locally in the auditory cortex.

neuroscience↗

bioRxiv2024

scDrugAtlas: an integrative single-cell drug response atlas for unraveling tumor heterogeneity in therapeutic efficacy

Tumor heterogeneity often leads to substantial differences in responses to same drug treatment. The presence of pre-existing or acquired drug-resistant cell sub-populations within a tumor survive and proliferate, ultimately resulting in tumor relapse and metastasis. The drug resistance is the leading cause of failure in clinical tumor therapy. Therefore, accurate identification of drug-resistant tumor cell subpopulations could greatly facilitate the precision medicine and novel drug development. However, the scarcity of single-cell drug response data significantly hinders the exploration of tumor cell resistance mechanisms and the development of computational predictive methods. In this paper, we propose scDrugAtlas, a comprehensive database devoted to integrating the drug response data at single-cell level. We manually compiled more than 100 datasets containing single-cell drug responses from various public resources. The current version comprises large-scale single-cell transcriptional profiles and drug response labels from 1,023 samples (cell line, mouse, PDX models, patients and bacterium), across 78 unique drugs and 24 major cancer types. Particularly, we assigned a confidence level to each response label based on the tissue source (primary or relapse/metastasis), drug exposure time and drug-induced cell phenotype. We believe scDrugAtlas could greatly facilitate the Bioinformatics community for developing computational models and biologists for identifying drug-resistant tumor cells and underlying molecular mechanism. The scDrugAtlas database is available at: http://drug.hliulab.tech/scDrugAtlas/.

bioinformatics↗

bioRxiv2024

Three-dimension transcriptomics maps of whole mouse embryo during organogenesis

Understanding mammalian development heavily relies on classical animal models like the house mouse (Mus musculus). Advanced spatial transcriptomics has enabled biologists to break new ground in studies of molecular dynamics and cellular patterning during embryonic development with spatiotemporal resolution. To construct a comprehensive developmental trajectory, current three-dimensional (3D) spatial transcriptomic profiling leverages mouse embryos from gastrulation (E5.5) to organogenesis (E13.5) in continuity. However, a crucial phase for early organogenesis between E9.5 and E11.5 was deficient. To unveil the mystery of this stage, we present the 3D transcriptomics of mouse embryos at E9.5 and E11.5, with a widely applicable reconstruction workflow that bypasses sophisticated bioinformatic calculations. As our 3D atlas is generated at single-cell resolution, we demonstrate how organogenetic processes can be interpreted at different levels of granularity, from local cellular interactions to whole embryonic regionalization. We release the open-access database MOSTA3D (Mouse Organogenesis Spatiotemporal Transcriptomic Atlas in Three-dimensional) and hope a broader community will contribute to extending this framework from conception to senility in the near future.

developmental biology↗

bioRxiv2024

Deep Learning Driven Cell-Type-Specific Embedding for Inference of Single-Cell Co-expression Networks

The inference of gene co-expression module in specific cell types is important for understanding of cell-type-specific biological processes. We have developed DeepCSCN, an unsupervised deep-learning framework, to infer gene co-expression modules from single-cell RNA sequencing (scRNA-seq) data. Utilizing a global-to-local network construction approach, DeepCSCN can infer co-expression at the whole sample level and construct cell-type-specific co-expression networks. Systematic evaluations on eight public scRNA-seq datasets show that DeepCSCN significantly outperforms eight existing methods in co-expression network construction. Furthermore, DeepCSCN effectively identifies cell-type-specific co-expression networks that are more enriched for cell-specific functional pathways compared to current methods. Finally, application of DeepCSCN on public scRNA-seq data revealed 280 cell-specific gene modules across 27 cell types, including epithelial cells, immune cells, and myonuclei, demonstrating its versatility and accuracy in elucidating cell-type-specific gene co-expression regulation. DeepCSCN offers a powerful tool for researchers to dissect the intricate gene co-expression networks within distinct cell types.

bioinformatics↗

bioRxiv2024

Exploiting functional regions in the viral RNA genome as druggable entities

RNA-targeting compounds or small interfering RNAs (siRNAs) offer a potent means for controlling viral infections. An essential prerequisite to their design depends on identifying conserved and functional viral RNA structures in cells. Techniques that probe RNA structures in situ have been developed recently including SHAPE-MaP, which has been helpful in analyzing the secondary structures of RNA. In this study, we report on the application of SHAPE-MaP to the Porcine Epidemic Diarrhoea Virus (PEDV) RNA genome to categorize different functional regions including potential quadruplex forming sequence and target sites of siRNA. Our results show that these structures can be exploited to inhibit viral proliferation and that SHAPE-MaP is an effective method to the identification of secondary structures in RNA genomes.

genomics↗

bioRxiv2024

Loss of RREB1 reduces adipogenesis and improves insulin sensitivity in mouse and human adipocytes

There are multiple independent genetic signals at the Ras-responsive element binding protein 1 (RREB1) locus associated with type 2 diabetes risk, fasting glucose, ectopic fat, height, and bone mineral density. We have previously shown that loss of RREB1 in pancreatic beta cells reduces insulin content and impairs islet cell development and function. However, RREB1 is a widely expressed transcription factor and the metabolic impact of RREB1 loss in vivo remains unknown. Here, we show that male and female global heterozygous knockout (Rreb1+/-) mice have reduced body length, weight, and fat mass on high-fat diet. Rreb1+/- mice have sex- and diet-specific decreases in adipose tissue and adipocyte size; male mice on high-fat diet had larger gonadal adipocytes, while males on standard chow and females on high-fat diet had smaller, more insulin sensitive subcutaneous adipocytes. Mouse and human precursor cells lacking RREB1 have decreased adipogenic gene expression and activated transcription of genes associated with osteoblast differentiation, which was associated with Rreb1+/- mice having increased bone mineral density in vivo. Finally, human carriers of RREB1 T2D protective alleles have smaller adipocytes, consistent with RREB1 loss-of-function reducing diabetes risk.

physiology↗

bioRxiv2024

Prediction of protein subcellular localization in single cells

The subcellular localization of a protein is important for its function and interaction with other molecules, and its mislocalization is linked to numerous diseases. While atlas-scale efforts have been made to profile protein localization across various cell lines, existing datasets only contain limited pairs of proteins and cell lines which do not cover all human proteins. We present a method that uses both protein sequences and cellular landmark images to perform Predictions of Unseen Proteins Subcellular localization (PUPS), which can generalize to both proteins and cell lines not used for model training. PUPS combines a protein language model and an image inpainting model to utilize both protein sequence and cellular images for protein localization prediction. The protein sequence input enables generalization to unseen proteins and the cellular image input enables cell type specific prediction that captures single-cell variability. PUPS ability to generalize to unseen proteins and cell lines enables us to assess the variability in protein localization across cell lines as well as across single cells within a cell line and to identify the biological processes associated with the proteins that have variable localization. Experimental validation shows that PUPS can be used to predict protein localization in newly performed experiments outside of the Human Protein Atlas used for training. Collectively, PUPS utilizes both protein sequences and cellular images to predict protein localization in unseen proteins and cell lines with the ability to capture single-cell variability.

cell biology↗

bioRxiv2024

Microglia-astrocyte interplay mitigates Aβ toxicity in a novel human 3D neurosphere model of Alzheimer's Disease

BackgroundAlzheimers Disease (AD) is characterized by progressive amyloid beta (A{beta}) deposition in the brain, with eventual widespread neurodegeneration. While the cell-specific molecular signature of end-stage AD is reasonably well characterized through autopsy material, less is known about the molecular pathways in the human brain involved in the earliest exposure to A{beta}. Human model systems that not only replicate the pathological features of AD but also the transcriptional landscape in neurons, astrocytes and microglia are crucial for understanding disease mechanisms and for identifying novel therapeutic targets. MethodsIn this study, we used a human 3D iPSC-derived neurosphere model to explore how resident neurons, microglia and astrocytes and their interplay are modified by chronic amyloidosis induced over 3 to 5 weeks by supplementing media with synthetic A{beta}1-42 oligomers. Neurospheres under chronic A{beta} exposure were grown with or without microglia to investigate the functional roles of microglia. Neuronal activity and oxidative stress were monitored using genetically encoded indicators, including GCaMP6f and roGFP1, respectively. Single nuclei RNA sequencing (snRNA-seq) was performed to profile A{beta} and microglia driven transcriptional changes in neurons and astrocytes, providing a comprehensive analysis of cellular responses. ResultsMicroglia efficiently phagocytosed A{beta} inside neurospheres and significantly reduced neurotoxicity, mitigating amyloidosis-induced oxidative stress and neurodegeneration following different exposure times to A{beta}. The neuroprotective effects conferred by the presence of microglia was associated with unique gene expression profiles in astrocytes and neurons, including several known AD-associated genes such as APOE. These findings reveal how microglia can directly alter the molecular landscape of AD. ConclusionsOur human 3D neurosphere culture system with chronic A{beta} exposure reveals how microglia may be essential for the cellular and transcriptional responses in AD pathogenesis. Microglia are not only neuroprotective in neurospheres but also act as key drivers of A{beta}-dependent APOE expression suggesting critical roles for microglia in regulating APOE in the AD brain. This novel, well characterized, functional in vitro platform offers unique opportunities to study the roles and responses of microglia to A{beta} modelling key aspects of human AD. This tool will help identify new therapeutic targets, accelerating the transition from discovery to clinical applications. HighlightsO_LIWell-characterized functional human iPSC-derived 3D neurospheres (hiNS) consisting of neurons and astrocytes can be supplemented with microglia/macrophages (hiMG) C_LIO_LIChronic amyloidosis in the presence of hiMG recapitulate key features and gene expression profiles of AD C_LIO_LIhiMG within the model phagocytose A{beta} and mitigate A{beta}-induced neurotoxicity, reducing oxidative stress and neuronal damage C_LIO_LIhiMG are essential for A{beta} to upregulate AD-like gene expression signatures in astrocytes C_LIO_LIImmunohistochemical analysis reveals hiMG-dependent colocalization of A{beta} and APOE C_LI

neuroscience↗

bioRxiv2024

RNA-ligand interaction scoring via data perturbation and augmentation modeling

RNA-targeting drug discovery is undergoing an unprecedented revolution. Despite recent advances in this field, developing data-driven deep learning models remains challenging due to the limited availability of validated RNA-small molecule interactions and the scarcity of known RNA structures. In this context, we introduce RNAsmol, a novel sequence-based deep learning framework that incorporates data perturbation with augmentation, graph-based molecular feature representation and attention-based feature fusion modules to predict RNA-small molecule interactions. RNAsmol employs perturbation strategies to balance the bias between true negative and unknown interaction space thereby elucidating the intrinsic binding patterns between RNA and small molecules. The resulting model demonstrates accurate predictions of the binding between RNA and small molecules, outperforming other methods with average improvements of [~]8% (AUROC) in 10-fold cross-validation, [~]16% (AUROC) in cold evaluation (on unseen datasets), and [~]30% (ranking score) in decoy evaluation. Moreover, we use case studies to validate molecular binding hotspots in the prediction of RNAsmol, proving the models interpretability. In particular, we demonstrate that RNAsmol, without requiring structural input, can generate reliable predictions and be adapted to many RNA-targeting drug design scenarios.

bioinformatics↗

bioRxiv2024

The Ptch methylase installs an m3U modification on 28S rRNA for efficient protein synthesis in flies and men

The ribosomal RNA (rRNA) is one of the most heavily modified RNA species in nature. Although we have advanced knowledge of the sites, functions and the enzymology of many of the rRNA modifications from all kingdoms, we lack basic understanding on many of those that are not universally present. A single N3 modified Uridine base (m3U) was identified on the 28S rRNA from human and frog over thirty years ago, which is absent in bacteria or yeast. Here we show that the equivalent m3U is present in Drosophila, and that the Ptch enzyme and its human homolog are both necessary and sufficient for carrying out the modification. The Ptch-modified U is at a functional center of the large ribosome, and consistently ptch-mutant cells suffer loss of ribosomal functions. Ptch, proposed to be the most druggable RNA methyltransferases in human, represents a unique target where ribosomal functions could be specifically compromised in cancer cells.

genetics↗