Search bioRxiv⌕ Search

Biology subjects

Wang, B.

Publications and source records attributed to Wang, B..

At least 433 records · Page 24Linked to original sources

scFormer: A Universal Representation Learning Approach for Single-Cell Data Using Transformers

AO_SCPLOWBSTRACTC_SCPLOWSingle-cell sequencing has emerged as a promising technique to decode cellular heterogeneity and analyze gene functions. With the high throughput of modern techniques and resulting large-scale sequencing data, deep learning has been used extensively to learn representations of individual cells for downstream tasks. However, most existing methods rely on fully connected networks and are unable to model complex relationships between both cell and gene representations. We hereby propose scFormer, a novel transformer-based deep learning framework to jointly optimize cell and gene embeddings for single-cell biology in an unsupervised manner. By drawing parallels between natural language processing and genomics, scFormer applies self-attention to learn salient gene and cell embeddings through masked gene modelling. scFormer provides a unified framework to readily address a variety of downstream tasks such as data integration, analysis of gene function, and perturbation response prediction. Extensive experiments using scFormer show state-of-the-art performance on seven datasets across the relevant tasks. The scFormer model implementation is available at https://github.com/bowang-lab/scFormer.

bioinformatics↗

A direct fiber approach to model sclera collagen architecture and biomechanics

Sclera collagen fiber microstructure and mechanical behavior are central to eye physiology and pathology. They are also complex, and are therefore often studied using modeling. Most models of sclera, however, have been built within a conventional continuum framework. In this framework, collagen fibers are incorporated as statistical distributions of fiber characteristics such as the orientation of a family of fibers. The conventional continuum approach, while proven successful for describing the macroscale behavior of the sclera, does not account for the sclera fibers are long, interwoven and interact with one another. Hence, by not considering these potentially crucial characteristics, the conventional approach has only a limited ability to capture and describe sclera structure and mechanics at smaller, fiber-level, scales. Recent advances in the tools for characterizing sclera microarchitecture and mechanics bring to the forefront the need to develop more advanced modeling techniques that can incorporate and take advantage of the newly available highly detailed information. Our goal was to create a new computational modeling approach that can represent the sclera fibrous microstructure more accurately than with the conventional continuum approach, while still capturing its macroscale behavior. In this manuscript we introduce the new modeling approach, that we call direct fiber modeling, in which the collagen architecture is built explicitly by long, continuous, interwoven fibers. The fibers are embedded in a continuum matrix representing the non-fibrous tissue components. We demonstrate the approach by doing direct fiber modeling of a rectangular patch of posterior sclera. The model integrated fiber orientations obtained by polarized light microscopy from coronal and sagittal cryosections of pig and sheep. The fibers were modeled using a Mooney- Rivlin model, and the matrix using a Neo-Hookean model. The fiber parameters were determined by inversely matching experimental equi-biaxial tensile data from the literature. After reconstruction, the direct fiber model orientations agreed well with the microscopy data both in the coronal plane (adjusted R2=0.8234) and in the sagittal plane (adjusted R2=0.8495) of the sclera. With the estimated fiber properties (C10=5746.9 MPa; C01=-5002.6MPa, matrix shear modulus 200kPa), the models stress-strain curves simultaneously fit the experimental data in radial and circumferential directions (adjusted R2s 0.9971 and 0.9508, respectively). The estimated fiber elastic modulus at 2.16% strain was 5.45GPa, in reasonable agreement with the literature. During stretch, the model exhibited stresses and strains at sub-fiber level, with interactions among individual fibers which are not accounted for by the conventional continuum methods. Our results demonstrate that direct fiber models can simultaneously describe the macroscale mechanics and microarchitecture of the sclera, and therefore that the approach can provide unique insight into tissue behavior questions inaccessible with continuum approaches. HighlightsCollagen fibers are the main load-bearing component of eye tissues. Conventional sclera modeling ignores that fibers are long, interwoven and interact. We demonstrate a direct fiber model with long, interwoven and interacting fibers. Collagen fiber mechanical properties were estimated using inverse fitting. The model captures simultaneously sclera fiber structure and macroscale mechanics.

bioengineering↗

Spatial Localization Ability of Planarians Identified Through the Light Maze Paradigm

Spatial localization ability is crucial for free-living animals to fit the environment. As shown by previous studies, planarians can be conditioned to discriminate directions. However, due to their simplicity and primitiveness, they had never been considered to have true spatial localization ability to retrieve locations of objects and places in the environment. Here, we introduce a light maze training paradigm to demonstrate that a planarian worm can navigate to a former recognized place from the start point, even if the worm is transferred into a newly produced maze. This finding identifies the spatial localization ability of planarians for the first time, which provides clues for the evolution of spatial learning. Since the planarians have a primitive brain with simple structures, this paradigm can also provide a simplified model for a detailed investigation of spatial learning.

animal behavior and cognition↗

Default But Not Rest: Topological Discrimination Defines the Default Mode Network

The default mode network (DMN), a set of transmodal cortical regions, has historically been argued to serve the internal functions of brain. The discovery of this network highlighted the brains intrinsic operations. The DMN generally decreases its activity during tasks and increases its activity during relaxed non-task states. It is important to investigate the nature of the DMN in order to understand the human brain in health and disease. In the current study, we discovered a task-related cortical network we called the Topological Discrimination Network (TDN), which was consistently revealed by contrasting activations from topological discrimination tasks with local geometric discrimination tasks. The TDN and the DMN consist of essentially the same group of brain regions and the fMRI response of topological discrimination in those regions exhibited consistent temporal dynamics with resting state. The robustness of the TDN is supported by multiple experiments performed at different field strengths (3T and 7T MRI scanner) as well as different types of signals measured (BOLD and CBF). The collective results suggest that the process of topological discrimination could almost be considered as a functional "default mode" of our brain. The TDN, like the DMN, could define the functional baseline of brain, with the advantage of functional consistency across participants and experimental sessions.

neuroscience↗

Early-life language deprivation affects specific neural mechanisms of semantic representations

One signature of the human brain is its ability to derive knowledge from language inputs, in addition to nonlinguistic sensory channels such as vision and touch. How does human language experience specifically modulate the way in which semantic knowledge is stored in the human brain? We investigated this question using a unique early-life language-deprivation human model: early deaf adults who are born to hearing parents and thus had delayed acquisition of any natural human language (speech or sign), with early deaf adults who acquired sign language from birth as nonlinguistic sensory experience controls. Neural responses in a meaning judgment task with 90 written words that were familiar to both groups were measured using fMRI. The early language-deprived group, compared with the deaf control group, showed reduced semantic sensitivity, in both multivariate pattern (semantic structure encoding) and univariate (abstractness effect) analyses, in the left dorsal anterior temporal lobe (dATL). These results provide positive, causal evidence that the neural semantic representation in dATL is specifically supported by language, as a unique mechanism of representing (abstract) semantic space, beyond the sensory-derived semantic representations distributed in the other cortical regions.

neuroscience↗

Oral Feeding of Nanoplastics reduces Brain function of Mice by Inducing Intestinal IL-1β-producing Macrophages

Background and AimsNanoplastics (NPs) as contaminants in food and water have drawn an increasing public attention. However, little is known about how NPs shape the gut immune landscape after entering the body. The objective of the study was to explore indirect effects caused by the interaction of NPs with the mammalian gut and whole immune system after entering the body. MethodsIn this study, we fabricated NPs ([~]500 nm) and microplastics (MPs) ([~]2 m) and aimed to evaluate their in vivo effects by feeding them in mice. The mechanism was then investigated by various technology including single-cell RNA sequencing of gut and brain tissue. ResultsThe results suggested that NPs showed a better ability to induce gut macrophage activation than did MPs. In addition, NPs triggered gut interleukin 1 beta (IL-1{beta})-producing macrophage reprogramming via inducing lysosomal damage after phagocytosis. More importantly, IL-1{beta} released from the intestine could affect brain immunity, leading to microglial activation and Th17 differentiation, all of which correlated with a decline in cognitive and short-term memory in NPs-fed mice ConclusionsThus, this study provides new insight into the mechanism of action of the gut-brain axis and delineates the way NPs reduce brain function, highlights the importance to fix the plastic pollution problem worldwide.

immunology↗

The conserved protein WhiA influences branched-chain fatty acid precursors in Bacillus subtilis

The conserved WhiA protein family is present in most Gram-positive bacteria and plays a role in cell division. WhiA contains a DNA-binding motive and has been identified as a transcription factor in actinomycetes. In Bacillus subtilis, the absence of WhiA influences cell division and chromosome segregation, however, it is still unclear how WhiA influences these processes, but the protein does not seem to function as transcription factor in this organism. To further investigate the function of WhiA in B. subtilis, we performed a yeast two-hybrid screen to find interaction partners, and a Hi-C experiment to reveal possible changes in chromosome conformation. The latter experiment indicated a reduction in short range chromosome interactions, but how this would affect either cell division or chromosome segregation is unclear. Based on adjacent genes, a role in carbon metabolism was put forward. To study this, we measured exometabolome fluxes during growth on different carbon sources. This revealed that in {Delta}whiA cells the pool of branched-chain fatty acid precursors is lower. However, the effect on the membrane fatty acid composition was minimal. Transcriptome data could not link the metabolome effects to gene regulatory differences. IMPORTANCEWhiA is a conserved DNA binding protein that influences cell division and chromosome segregation in the Gram-positive model bacterium B. subtilis. The molecular function of WhiA is still unclear, but a previous study has suggested that the protein does not function as a transcription factor. In this study, we used yeast two-hybrid screening, chromosome conformation capture analysis, metabolomics, transcriptomics and fatty acid analysis to obtain more information about the workings of this enigmatic protein.

microbiology↗

Integrated Multi-Omics Analysis of Brain Aging in Female Nonhuman Primates Reveals Altered Signaling Pathways Relevant to Age-Related Disorders

The prefrontal cortex (PFC) has been implicated as a key brain region responsible for age-related cognitive decline. Little is known about aging-related molecular changes in PFC that may mediate these effects. To date, no studies have used untargeted discovery methods with integrated analyses to determine PFC molecular changes in healthy female primates. We quantified PFC changes associated with healthy aging in female baboons by integrating multiple omics data types (transcriptomics, proteomics, metabolomics) from samples across the adult age span. Our integrated omics approach using unbiased weighted gene co-expression network analysis (WGCNA) to integrate data and treat age as a continuous variable, revealed highly interconnected known and novel pathways associated with PFC aging. We found GABA tissue content associated with these signaling pathways, providing one potential biomarker to assess PFC changes with age. These highly coordinated pathway changes during aging may represent early steps for aging-related decline in PFC functions, such as learning and memory, and provide potential biomarkers to assess cognitive status in humans.

systems biology↗

Identification of plant-derived microRNAs in human kidney

Plant-derived microRNAs (miRNAs) have been implicated as functional regulators in human diseases, although conclusive evidence of this effect remains to be reported. To examine their potential functional role, we profiled the plant-derived miRNAs in 139 blood exosome samples from renal transplantation patients and were able to identify 331 plant-derived miRNAs representing 149 families. According to their miRBase annotation, these miRNAs can be traced back to 76 plant species, most of which are foods common to the human diet (e.g., tomato, soybean, potato and rice). We also profiled 41 blood exosome samples from 22 patients with acute immune rejection (AR) of renal transplants and compared them to 21 samples from 11 patients with stable allograft function to explore possible roles of the functional plant miRNAs. We identified three plant-derived miRNAs (miR4995, miR2118/2218 and miR167) associated with allograft AR whose regulatory targets are mRNAs controlling immune response, T cell activation, and other cellular functions. miR4995 mimics were generated, transfected into HEK293T cells, and their function verified. Our findings not only demonstrate the presence of functional plant-derived miRNAs in human cells, but also provide initial evidence that these miRNAs may be involved in malfunction of renal transplantation.

genetics↗

Expansion Spatial Transcriptomics

Capture array-based spatial transcriptomics methods have been widely used to resolve gene expression in diverse tissue contexts, however, their spatial resolution is limited by array density. We present Expansion Spatial Transcriptomics (Ex-ST) to overcome this limitation by clearing and expanding tissue prior to capturing the entire transcriptome. This approach allows us to achieve near cellular resolution and higher capture efficiency of lowly-expressed genes, which we demonstrate using mouse brain samples.

genomics↗

Transcriptome analyses using regulon, functional category, and operon information with GINtool

When analysing transcriptome data, threshold values are chosen to decide whether the regulation of a gene is relevant or not, however this may result in the loss of valuable information. To overcome this problem it can be useful to analyse regulons instead of individual genes, to harness the statistical power of combining genes. Another advantage of a regulon-based analysis is that it provides direct insights into the activity of regulatory pathways, which is the essence of transcriptome analyses. We have developed a software tool called GINtool that can use regulon information to analyse transcriptome data. GINtool includes the option to take the activity mode of a regulator in account, which is important when a regulator can function both as an activator and repressor. GINtool also contains two novel graphical representations that greatly facilitate the visual inspection of regulon-based transcriptome analyses. Additional features of GINtool includes the evaluation of transcriptome data using functional categories, and the analysis of gene expression differences within operons. To ease the analyses and downstream processing of figures, GINtool has been developed as an add-in for Excel.

bioinformatics↗

The differential impacts of dataset imbalance in single-cell data integration

Single-cell transcriptomic data measured across distinct samples has led to a surge in computational methods for data integration. Few studies have explicitly examined the common case of cell-type imbalance between datasets to be integrated, and none have characterized its impact on downstream analyses. To address this gap, we developed the Iniquitate pipeline for assessing the stability of single-cell RNA sequencing (scRNA-seq) integration results after perturbing the degree of imbalance between datasets. Through benchmarking 5 state-of-the-art scRNA-seq integration techniques in 1600 perturbed integration scenarios for a multi-sample peripheral blood mononuclear cell (PBMC) dataset, our results indicate that sample imbalance has significant impacts on downstream analyses and the biological interpretation of integration results. We observed significant variation in clustering, cell-type classification, marker gene-based annotation, and query-to-reference mapping in imbalanced settings. Two key factors were found to lead to quantitation differences after scRNA-seq integration - the cell-type imbalance within and between samples (relative cell-type support) and the relatedness of cell-types across samples (minimum cell-type center distance). To account for evaluation gaps in imbalanced contexts, we developed novel clustering metrics robust to sample imbalance, including the balanced Adjusted Rand Index (bARI) and balanced Adjusted Mutual Information (bAMI). Our analysis quantifies biologically-relevant effects of dataset imbalance in integration scenarios and introduces guidelines and novel metrics for integration of disparate datasets. The Iniquitate pipeline and balanced clustering metrics are available at https://github.com/hsmaan/Iniquitate and https://github.com/hsmaan/balanced-clustering, respectively.

bioinformatics↗

Embracing fine-root system complexity to improve the predictive understanding of ecosystem functioning

Projecting the functioning of the biosphere requires a holistic consideration of whole-ecosystem processes. Although improving leaf and canopy processes has been the focus of ecosystem model development since the 1970s, the arbitrary homogenization of fine-root systems into a single pool is at odds with observations. This discrepancy has increased in the last two decades as accelerated conceptual and empirical advances have revealed functional differentiation and cooperation conferred by the hierarchical structure of fine-root orders and associations with mycorrhizal fungi in fine-root systems. To close this model-data gap, we propose a 3-pool structure comprising Transport and Absorptive fine roots with Mycorrhizal fungi (TAM) to model vertically resolved fine-root systems across organizational and spatial-temporal scales. A comparison of TAM to the single fine-root structure in a state-of-the-art Earth System Model using the big-leaf approach demonstrates robust impacts on carbon cycling in temperate forests, lending further quantitative support to the empirical and theoretical basis for TAM. Strong support in both theory and practice therefore suggests a move beyond the useful but incorrect paradigm of single-pool homogenization, echoing a broad trend of embracing ecological complexities in terrestrial ecosystem modelling. Although challenges lay ahead towards realizing TAM in ecologically realistic demography models simulating emergent functioning from pattern and diversity, adoption of TAM by both modelers and empiricists holds promise to build a better predictive understanding of ecosystem functioning in the context of global change.

ecology↗

Flaver: mining transcription factors in genome-wide transcriptome profiling data using weighted rank correlation statistics

BackgroundMining key transcription factors (TFs) in genome-wide transcriptome profiling data has been an active research area for many years and it has been partially solved by mathematically modelling the ranking orders of genes in the target gene-set for the TF of interest in the gene-list ranked by expression values, called gene-set enrichment analysis (GSEA). However, in some application scenarios the gene-set itself also has a rank attribute, such as the putative target gene-set predicted by the Grit software and other alternatives like FIMO and Pscan. New algorithms must be developed to analyze these data properly. Methodology/Principal FindingsBy implementing the weighted Kendalls tau statistic, we proposed a method for genome-wide transcriptome profiling data mining that can identify the key TFs orchestrating a profile. Theoretical properties of the proposed method were established, and its advantages over the GSEA approach were demonstrated when analyzing the RNA-Atlas datasets. The results showed that the top-rated TFs by our method always have experimentally supported evidences in the literatures. Benchmarking using gene ontology (GO) annotations in the AmiGO database indicated that the geometry performance (SQR_P) of our method is higher than GSEA in more than 14% of the cases. SignificanceThe developed method is suitable for analyzing the significance of overrepresentation of ranked gene-sets in a ranked gene-list. A software implementing the method, called "Flaver", was developed and is publicly available at http://www.thua45.cn/flaver under an academic free license. Author SummaryIdentification of the regulation roles of TFs in the transcriptome is fundamental in understanding various biological processes. Improve the performance of the prediction tools is important because accurate TF-mining in transcriptome data can finely improve the efficiency of wet-lab experiments. Also, genome wide TF-mining can provide new target genes for transcriptome regulation analysis in system biology perspective. This study developed a new TF-mining tool based on weighted rank correlation statistical method. The tool has better performance in analyzing ranked gene-set and ranked gene-list than its competitor, the GSEA tool. It can help the researchers in identification of the most important TFs in transcriptional data.

bioinformatics↗

Entropic control of the free energy landscape of an archetypal biomolecular machine

Biomolecular machines are complex macromolecular assemblies that utilize thermal and chemical energy to perform essential, multistep, cellular processes. Despite possessing different architectures and functions, an essential feature of the mechanisms-of-action of all such machines is that they require dynamic rearrangements of structural components. Surprisingly, biomolecular machines generally possess only a limited set of such motions, suggesting that these dynamics must be repurposed to drive different mechanistic steps. Although ligands that interact with these machines are known to drive such repurposing, the physical and structural mechanisms through which ligands achieve this remain unknown. Using temperature-dependent, single-molecule measurements analyzed with a time-resolution-enhancing algorithm, here we dissect the free energy landscape of an archetypal biomolecular machine, the bacterial ribosome, to reveal how its dynamics are repurposed to drive distinct steps during ribosome-catalyzed protein synthesis. Specifically, we show that the free energy landscape of the ribosome encompasses a network of allosterically coupled structural elements that coordinates the motions of these elements. Moreover, we reveal that ribosomal ligands which participate in disparate steps of the protein synthesis pathway repurpose this network by differentially modulating the structural flexibility of the ribosomal complex (i.e., the entropic component of the free energy landscape). We propose that such ligand-dependent entropic control of free energy landscapes has evolved as a general strategy through which ligands may regulate the functions of all biomolecular machines. Such entropic control is therefore an important driver in the evolution of naturally occurring biomolecular machines and a critical consideration for the design of synthetic molecular machines.

biophysics↗

Structure-function analysis of 110 phosphorylation sites on the circadian clock protein FRQ identifies clusters determining period length and temperature compensation

In the negative feedback loop driving the Neurospora circadian oscillator, the negative element, FREQUENCY (FRQ), inhibits its own expression by promoting phosphorylation of its heterodimeric transcriptional activators, White Collar-1 (WC-1) and WC-2. FRQ itself also undergoes extensive time-of-day-specific phosphorylation with over 100 phosphosites previously documented. Although disrupting individual or certain clusters of phosphorylation sites has been shown to alter circadian period lengths to some extent, how all the phosphorylations on FRQ control its activity is still elusive. In this study, we systematically investigated the role in period determination of all 110 phosphorylated residues reported on FRQ by mutagenetic and luciferase reporter assays. Surprisingly, robust FRQ phosphorylation is still detected even when 84 phosphosites were eliminated altogether; further mutating another 26 phosphoresidues completely abolished FRQ phosphorylation. To identify phosphoresidue(s) on FRQ impacting circadian period length, series of clustered frq phosphomutants covering all the 110 phosphosites were generated and examined for period changes. When phosphosites in the N-terminal and middle regions of FRQ were eliminated, longer periods were mostly seen while removal of phosphorylation in the C-terminal tail result in extremely short periods, among the shortest reported. Interestingly, abolishing the 11 phosphosites in the C-terminal tail of FRQ does not only result in an extremely short period, but also causes an over-compensated circadian oscillator under a range of physiological temperatures. When different groups of phosphomutations on FRQ were combined intramolecularly, an additive effect was observed as expected; unexpectedly, arrhythmicity resulting from one cluster frq phosphorylation mutants was restored by eliminating phosphorylation at another group of sites, suggesting an epistatic effect between phosphoevents. ImportanceExisting in most eukaryotes, circadian clocks are built based on cell-autonomous, auto-regulatory feedback loops in which negative elements feed back to depress their own expression by repressing the positive elements that drive their synthesis. In Neurospora, the WCC transcription activator drives expression of FRQ, which complexes with FRH and CK1 to repress the DNA-binding activity of WCC by promoting phosphorylation at a group of residues of WCC. The phosphorylation status of FRQ determines the circadian period length, acting independent of effects of phosphorylation on FRQ half-life. Reflecting this dominant role of phosphorylation, FRQ is subject to substantial phosphorylation at over 100 sites in a time-of-day-specific manner. However, how this plethora of phosphoevents on FRQ controls its activity in a circadian cycle is still elusive, and prior work had shown limited effects of individual phosphosite point mutants. In this study, a series of frq mutants targeting multisite phosphorylation within domains of FRQ were generated and analyzed in order to define their roles in period determination. A clear pattern of period altering effects was observed in these frq mutants; certain mutants display strong temperature compensation phenotypes, and interestingly, a novel epistatic relationship on rhythmicity between phosphogroups emerged.

genetics↗

Structure-Based Pathogenicity Relationship Identifier (SPRI): A Novel Structure-Guided Method to Evaluate Pathological Effect of Missense Mutations

We report the Structure-based Pathogenicity Relationship Identifier (SPRI), a novel computational tool for accurate evaluation of pathological effects of missense single mutations and prediction of higher-order spatially organized units of mutational clusters. SPRI can effectively extract properties determining pathogenicity encoded in protein structures, and can identify deleterious missense mutations of germ line origin associated with Mendelian diseases, as well as mutations of somatic origin associated with cancer drivers. It compares favorably to other methods in predicting deleterious mutations. Furthermore, SPRI can discover spatially organized pathogenic higher-order spatial clusters (patHOS) of deleterious mutations, including those of low recurrence, and can be used for discovery of candidate cancer driver genes and driver mutations. We further demonstrate that SPRI can take advantage of AlphaFold2 predicted structures and can be deployed for saturation mutation analysis of the whole human proteome.

bioinformatics↗

Massively-parallel Microbial mRNA Sequencing (M3-Seq) reveals heterogenous behaviors in bacteria at single-cell resolution

Bacterial populations are highly adaptive. They can respond to stress and survive in shifting environments. How the behaviors of individual bacteria vary during stress, however, is poorly understood. To identify and characterize rare bacterial subpopulations, technologies for single-cell transcriptional profiling have been developed. Existing approaches, though, are all limited in some technical capacity (e.g., number of cells or transcripts that can be profiled). Due in part to these limitations, few conditions have yet been studied with these tools. Here, we develop Massively-parallel Microbial mRNA sequencing (M3-Seq), a single-cell RNA-sequencing platform for bacteria that pairs combinatorial cell indexing with post hoc rRNA depletion. We show that M3-Seq can profile hundreds of thousands of bacterial cells from different species under a range of conditions in single experiments. We then apply M3-Seq to reveal rare populations, insights into bet hedging strategies during stress responses, and host responses to phage infection.

microbiology↗