Biology subjects

Cai, J. J.

Publications and source records attributed to Cai, J. J..

At least 19 recordsLinked to original sources

bioRxiv2025

STAT1-Mediated Regulation of IL-17A/CEBPB/NF-κB Axisin HIV-1 Infected Human Cerebral Organoids RevealsTherapeutic Targets for Neuroprotection

HIV-associated neurological complications remain a major concern in people living with HIV (PLWH), even under effective viral suppression with combination antiretroviral therapy (cART), underscoring unresolved mechanisms driving HIV-related neurodegeneration. To address this gap, we used HIV-1-infected human cerebral organoids (hCOs) containing microglia as a physiologically relevant three-dimensional model of the central nervous system (CNS) and validated key findings in simian immunodeficiency virus (SIV)-infected, cART-treated rhesus macaque model. Integrating single-nucleus RNA sequencing with ATAC sequencing, we identified cell-type-specific alterations in dual innate immune-driven inflammatory signaling pathways across neural and glial populations. Microglia were preferentially infected and activated, initiating IL-17A-mediated cascades involving IFN-{gamma}/STAT1, CEBPB, and NF-{kappa}B pathways, which converged with cGAS-STING and Notch-NEURL1 signaling, both associated with neuroinflammation and synaptic dysfunction. cART suppressed viral replication, reduced IL-17A-driven neuroinflammation, improved mitochondrial function, and restored synaptic gene expression. Collectively, these findings identify IL-17A, STAT1, CEBPB, and cGAS-STING as key molecular drivers of HIV-associated neuroinflammation and establish a novel mechanistic link between IL-17A signaling and microglial remodeling, providing a translational framework for the development of CNS-targeted therapies in PLWH.

neuroscience↗

bioRxiv2025

Mapping Gene Impact on Single-cell Transcriptomic Networks via Perturbation Response Scanning

Gene knockout experiments are essential for dissecting gene function, and CRISPR has made targeted gene disruption more accessible than ever. Single-cell CRISPR screening enables the construction of rich genetic perturbation landscapes, facilitating the identification of genes whose perturbation strongly reshapes cellular states. However, due to the nonlinear dependencies within gene networks, identifying the most impactful tangible genes remains challenging. Existing virtual knockout methods estimate downstream effects of single-gene deletions but do not evaluate whether such perturbations disrupt global information flow or compromise cellular robustness. To address this limitation, we adapt a perturbation-response framework originally developed for protein structural dynamics to identify gene modules most susceptible to perturbation. We introduce the single-cell Perturbation Impact Index (scPII), a data-driven metric derived from gene regulatory networks that quantifies system-level responses to gene perturbations, without using any CRISPR screening information. Our results demonstrate that scPII effectively identifies genes whose perturbation has the greatest system-wide impact. Analysis of single-cell RNA-seq data from the Cancer Cell Line Encyclopedia revealed a strong correlation between scPII-derived gene impact scores and gene effect scores from genome-wide CRISPR screens. These findings indicate that scPII provides a robust metric for quantifying gene knockout effects. More broadly, integrating perturbation response scanning with gene regulatory networks offers a powerful framework for advancing single-cell data analytics in biomedical research.

bioinformatics↗

bioRxiv2025

Constructing Gene Regulatory Network using Chatterjee's Rank Correlation with Single-cell Transcriptomic Data

Discovering gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data is critical for understanding cellular function. Still, existing methods are limited by strong theoretical assumptions or high computational complexity. We introduce a multiple testing framework for GRN construction using Chatterjee's rank correlation coefficient, a nonparametric measure of dependence. Our approach overcomes the limitations of traditional methods while offering a transparent, scalable, and computationally efficient alternative to recent black-box machine learning models. Crucially, to address the non-independence of cellular observations inherent to scRNA-seq, we develop a data-driven algorithm for estimating robust testing cutoffs. Furthermore, we exploit the asymmetric nature of Chatterjee's correlation to propose a new test for active regulation, enabling the construction of biologically meaningful and directionally informed GRNs. We demonstrate that our method matches or outperforms state-of-the-art approaches in recovering true gene-gene dependencies and directed regulatory interactions from both simulated and real datasets, particularly for complex, non-linear dependencies, providing a powerful tool for dissecting complex GRNs.

bioinformatics↗

bioRxiv2025

Genotype-Phenotype Distinctions in Spastic Paraplegia 4 Reveal HDAC6 as a Therapeutic Target

Spastic Paraplegia 4 (SPG4) is the most prevalent form of Hereditary Spastic Paraplegia (HSP), a neurodegenerative disorder characterized by progressive lower limb spasticity and debilitating gait impairment, primarily driven by axonal degeneration of corticospinal motor neurons (CSMNs). Caused by mutations in the SPAST gene encoding spastin, an AAA-ATPase involved in microtubule severing and intracellular organelle function, SPG4 accounts for 40-50% of autosomal dominant HSP cases, yet without effective treatments. Although reduced microtubule acetylation has emerged as a key pathological mechanism, whether and how distinct mutations lead to microtubule deacetylation and subsequent neurodegeneration remains unclear. To address this, we generated isogenic human induced pluripotent stem cell (hiPSC) lines with two distinct heterozygous SPAST mutations - SPASTWT/C448Y (missense) and SPASTWT/S245X (truncation). Employing an innovative differentiation protocol, we created human motor cortical organoids enriched in CSMNs, providing a robust platform to study SPG4 pathophysiology. These organoids revealed striking genotype-phenotype distinctions, with mutation-specific variations in CSMN loss, axonal degeneration and neuronal activities, mirroring clinical heterogeneity. Mechanistic studies identified aberrant activation of histone deacetylase 6 (HDAC6), a major neuronal microtubule deacetylase, as a key driver of SPG4 pathology. This dysregulation was specifically attributed to mutant M1-spastin, the longer isoform of spastin. Remarkably, pharmacological inhibition of HDAC6 with Tubastatin A restored microtubule acetylation status and mitigated axonal degeneration in both SPAST-mutant organoids, with corresponding improvements in corticospinal tract integrity and gait deficits validated in SPG4-transgenic mice. Collectively, our study establishes isogenic hiPSC-derived motor cortical organoids as a robust human model for corticospinal motor neuron degeneration and identifies HDAC6 hyperactivation as a central pathogenic mechanism and viable therapeutic target in SPG4.

neuroscience↗

bioRxiv2025

Aerobic exercise decreases the number and transcript expression of inflammatory M1 macrophages and CD8+ T cells in the epicardial adipose tissue of female pigs

BackgroundEpicardial adipose tissue (EAT) regulates coronary artery function via lipid metabolism and immune cell recruitment. Increased EAT is a risk factor for coronary artery disease (CAD), but aerobic exercise mitigates CAD. The effect of aerobic exercise on immune cells in EAT is unknown. We hypothesized that aerobic exercise creates an anti-inflammatory environment characterized by increased M2 macrophages and up-regulation of anti-inflammatory cytokine transcripts in EAT. MethodsFemale Yucatan pigs (n=7) were allocated to sedentary or exercised groups. To mimic CAD, a coronary artery was chronically occluded or remained non-occluded. EAT samples were processed for bulk and single nuclei transcriptomic sequencing. ResultsSub-clustering identified immune, endothelial, smooth muscle, adipocytes, adipocyte progenitor cells (APSCs), and neuronal cells, with adipocytes and APSCs being dominant. Non-occluded sedentary EAT had the largest percentage of M1 macrophages and CD8+ T cells. Irrespective of occlusion, sedentary EAT had the largest fraction of cells expressing genes in the tumor necrosis factor (TNF) superfamily. Irrespective of occlusion, exercise upregulated peroxisome proliferator-activated receptor (PPAR) gamma (G) expression and enriched PPAR signaling pathways in adipocytes, macrophages, and T cells. However, PPARG expression was lowest in CD8+ T cells from non-occluded exercised EAT. The greatest number of significant cell-cell communications between adipocytes and immune cells via growth factors and adhesion molecules occurred in occluded sedentary EAT. ConclusionAerobic exercise mitigates the proinflammatory nature of EAT in CAD via modulation of immune cell subpopulations, decreased TNF superfamily and increased PPARG gene expression, and decreased growth factor communication between adipocytes and immune cells.

cell biology↗

bioRxiv2025

Role of RSPO3 in Estrogen-mediated Sex Differences in Body Fat Distribution: A Single-cell Data-driven Study

Obesity is a global health concern affecting more than one-third of the adult population worldwide. Epidemiological studies indicate that body fat distribution, more than overall adiposity, is a key risk factor for obesity-related diseases. Genome-wide association studies revealed that RSPO3 is associated with body fat distribution. Additionally, marked sex differences in body fat distribution are largely influenced by estrogen, suggesting that exploring the relationship between estrogen and RSPO3 expression may illuminate the mechanisms underlying the sexual dimorphism of fat distribution. We revisited a comprehensive set of gene expression data from humans and mice to explore these mechanisms. We assessed RSPO3 expression in male and female adipose tissues and utilized single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics to examine RSPO3s function within different tissues. Our findings confirmed significantly higher RSPO3 expression in male adipose tissue compared to female. Furthermore, our data for the first time suggest that estrogen plays a crucial role in fat distribution, as blocking estrogen signaling upregulates RSPO3 expression. In mouse liver, we observed a correlation between sex-specific ADK overexpression and RSPO3 expression, resulting in increased obesity and hepatocyte expansion via WNT/{beta}-catenin signaling in male mice. This study sheds light on how estrogen influences sexual dimorphism in fat distribution and underscores RSPO3s vital role in this process. These insights open new avenues for targeted therapeutic interventions addressing obesity and metabolic diseases with a focus on sex-specific factors.

bioinformatics↗

bioRxiv2024

Functional diversification of the MADS-box gene family in fine-tuning the dimorphic transition of Talaromyces marneffei

The dynamic transition between yeast and hypha is a crucial adaptive mechanism for many human pathogenic fungi, including Talaromyces marneffei, a thermodimorphic fungus responsible for causing fatal talaromycosis. In the current study, we elucidated the roles of the MADS-box gene family in fine-tuning the dimorphic transition in T. marneffei through functional diversification. Utilizing adaptive laboratory evolution, we identified an enrichment of MADS-box genes in mutants deficient in yeast-to- mycelium transition. Further phylogenetic analyses revealed a significant expansion of MADS-box gene family within T. marneffei. Functional genetic manipulations revealed that overexpression of mads9, as opposed to its paralog mads10, effectively impeded the hyphal-to-yeast transition. Through integrating RNA sequencing (RNA-seq) and chromatin immunoprecipitation sequencing (ChIP-seq), we demonstrated that mads9 and the previously characterized madsA modulated the rate of hyphal-to-yeast conversion by orchestrating metabolic pathways and membrane dynamics, respectively, with mutual regulation via shared target genes. Our findings illuminated the distinct functional roles of the MADS-box family in regulating dimorphic transitions in T. marneffei, offering new insights into fungal adaptability.

microbiology↗

bioRxiv2024

Beyond Differential Expression: Embracing Cell-to-Cell Variability in Single-Cell Gene Expression Data Analysis

Single-cell RNA sequencing (scRNA-seq) has revolutionized our understanding of cellular variability by capturing gene expression profiles of individual cells. The importance of cell-to-cell variability in determining and shaping cell function has been widely appreciated. Nevertheless, differential expression (DE) analysis remains a cornerstone method in analytic practice, by focusing exclusively on mean expressional differences, current computational analyses overlook the rich information encoded by variability within the single-cell gene expression data. Thus, there is a need for more relevant approaches to assess data variability rather than the mean in order to offer a deeper understanding of cellular systems. Here we present spline-DV, a novel statistical framework for differential variability (DV) analysis using scRNA-seq data. The spline-DV method identifies genes exhibiting significantly increased or decreased expression variability among cells derived from two experimental conditions. Case studies show that DV genes identified using spline-DV are representative, and functionally relevant in the tested cellular conditions including obesity, fibrosis, and cancer.

bioinformatics↗

bioRxiv2024

Gene Function Revealed at the Moment of Stochastic Gene Silencing

Gene expression is a dynamic and stochastic process characterized by transcriptional bursting followed by periods of silence. Single-cell RNA sequencing (scRNA-seq) is a powerful tool to measure transcriptional bursting and silencing at the individual cell level. In this study, we introduce the single-cell Stochastic Gene Silencing (scSGS) method, which leverages the natural variability in single-cell gene expression to decipher gene function. For a target gene g under investigation, scSGS classifies cells into transcriptionally active (g+) and silenced (g-) samples. It then compares these cell samples to identify differentially expressed genes, referred to as SGS-responsive genes, which are used to infer the function of the target gene g. Analysis of real data demonstrates that scSGS can reveal regulatory relationships up- and downstream of target genes, circumventing the survivorship bias that often affects gene knockout and perturbation studies. scSGS thus offers an efficient approach for gene function prediction, with significant potential to reduce the use of genetically modified animals in gene function research.

bioinformatics↗

bioRxiv2024

Ephrin Forward Signaling Controls Interspecies Cell Competition in Pluripotent Stem Cells

In the animal kingdom, evolutionarily conserved mechanisms known as cell competition eliminate unfit cells during development. Interestingly, cell competition also leads to apoptosis of donor cells upon direct contact with host cells from a different species during interspecies chimera formation. The mechanisms underlying how host animal cells recognize and transmit cell death signals to adjacent xenogeneic human cells remain incompletely understood. In this study, we developed an interspecies cell contact reporter system to dissect the mechanisms underlying competitive interactions between mouse and human pluripotent stem cells (PSCs). Through single-cell RNA-seq analyses, we discovered that Ephrin A ligands in mouse cells play a crucial role in signaling cell death to adjacent human cells that express EPHA receptors during interspecies PSC co-culture. We also demonstrated that blocking the Ephrin A-EPHA receptor interaction pharmacologically, and inhibiting Ephrin forward signaling genetically in the mouse cells, enhances the survival of human PSCs and promotes chimera formation both in vitro and in vivo. Our findings elucidate key mechanisms of interspecies PSC competition during early embryogenesis and open new avenues for generating humanized tissues or organs in animals, potentially revolutionizing regenerative medicine.

cell biology↗

bioRxiv2024

Controlled Noise: Evidence of Epigenetic Regulation of Single-Cell Expression Variability

MotivationUnderstanding single-cell expression variability (scEV) or gene expression noise among cells of the same type and state is crucial for delineating population-level cellular function. While epigenetic mechanisms are widely implicated in gene expression regulation, a definitive link between chromatin accessibility and scEV remains elusive. Advances in single-cell techniques now enable simultaneous measurement of scATAC-seq and scRNA-seq within individual cells, presenting an unprecedented opportunity to address this gap. ResultsThis paper introduces an innovative testing pipeline to investigate the association between chromatin accessibility and scEV. The pipeline hinges on comparing the prediction performance of scATAC-seq data on gene expression levels between highly variable genes (HVGs) and non-highly variable genes (non-HVGs). Applying this pipeline to paired scATAC-seq and scRNA-seq data from human hematopoietic stem and progenitor cells, we observed a significantly superior prediction performance of scATAC-seq data for HVGs compared to non-HVGs. Notably, there was substantial overlap between well-predicted genes and HVGs. The gene pathways enriched from well-predicted genes are highly pertinent to cell type-specific functions. Our findings support the notion that scEV largely stems from cell-to-cell variability in chromatin accessibility, providing compelling evidence for the epigenetic regulation of scEV and offering promising avenues for investigating gene regulation mechanisms at the single-cell level. Availability and implementationThe source code and data used in this paper can be found at https://github.com/SiweiCui/EpigeneticControlOfSingle-CellExpressionVariability.

bioinformatics↗

bioRxiv2023

Single-nucleus transcriptomics of epicardial adipose tissue from females reveals exercise control of innate and adaptive immune cells

Coronary artery disease (CAD) is a leading cause of death in women. Although exercise mitigates CAD, the mechanisms by which exercise impacts epicardial adipose tissue (EAT) are unknown. We hypothesized that exercise promotes an anti-inflammatory microenvironment in EAT from female pigs. Yucatan pigs (n=7) were assigned to sedentary (Sed) or exercise (Ex) treatments and coronary arteries were occluded (O) with an ameroid to mimic CAD or remained non-occluded (N). EAT was collected for bulk and single nucleus transcriptomic sequencing (snRNA-seq). Exercise upregulated G-protein coupled receptor, S100 family, and FAK pathways and downregulated the coagulation pathway. Exercise increased the interaction between immune, endothelial, and mesenchymal cells in the insulin-like growth factor pathway and between endothelial and other cell types in the platelet endothelial cell adhesion molecule 1 pathway. Sub- clustering revealed nine cell types in EAT with fibroblast and macrophage populations predominant in O-Ex EAT and T cell population predominant in N-Ex EAT. Coronary occlusion impacted the largest number of genes in T and endothelial cells. Genes related to fatty acid metabolism were the most highly upregulated in non-immune cells from O-Ex EAT. Sub-clustering of endothelial cells revealed that N-Ex EAT separated from other treatments. In conclusion, aerobic exercise increased interaction amongst immune and mesenchymal and endothelial cells in female EAT. Exercise was minimally effective at reversing alterations in gene expression in endothelial and mesenchymal cells in EAT surrounding occluded arteries. These findings lay the foundation for future work focused on the impact of exercise on cell types in EAT. Significance StatementCoronary artery disease (CAD) is a leading cause of death in women. However, the role of epicardial adipose tissue (EAT) in the development of CAD in females and how exercise, which is recommended to slow CAD progression, impacts EAT are unknown. The effect of aerobic exercise on gene expression in EAT was investigated with RNA-sequencing, revealing significant alterations in fatty acid processing and immunoregulatory processes. This study provides valuable insights into the molecular and cellular changes induced in EAT by exercise in the context of chronic ischemic heart disease in females. These findings bolster current understanding of the impact of aerobic exercise on cardiac health in females and provide a foundation for future research in the field of exercise science.

physiology↗

bioRxiv2023

Interpretable modeling of time-resolved single-cell gene-protein expression using CrossmodalNet

Cell-surface proteins play a critical role in cell function and are primary targets for therapeutics. CITE-seq is a single-cell technique that enables simultaneous measurement of gene and surface protein expression. It is powerful but costly and technically challenging. Computational methods have been developed to predict surface protein expression using gene expression information such as from single-cell RNA sequencing (scRNA-seq) data. Existing methods however are computationally demanding and lack the interpretability to reveal underlying biological processes. We propose CrossmodalNet, an interpretable machine learning model, to predict surface protein expression from scRNA-seq data. Our model with a customized adaptive loss accurately predicts surface protein abundances. When samples from multiple time points are given, our model encodes temporal information into an easy-to-interpret time embedding to make prediction in a time point-specific manner able to uncover noise-free causal gene-protein relationships. Using two publicly available time-resolved CITE-seq data sets, we validate the performance of our model by comparing it to benchmarking methods and evaluate its interpretability. Together, we show our method accurately and interpretably profiles surface protein expression using scRNA-seq data, thereby expanding the capacity of CITE-seq experiments for investigating molecular mechanisms involving surface proteins.

bioinformatics↗

bioRxiv2022

A developmental program that regulates mammalian organ size offsets evolutionary distance

Pigs are evolutionarily more distant from humans than mice, but their physiological organs are closest to humans. The molecular program leading to a more than 1,000-fold increase in organ size in pigs and humans over that of mice across evolution has not been elucidated. We generated large-scale transcriptional landscapes throughout swine lung development. Our cross-species single-cell molecular atlas let us discover swine progenitor identities, stage-specific markers, and a core organ-size regulation program (COSRP), well-conserved in swine and humans but less so in mice. Across eight mammalian species, human COSRP promoters showed higher homologies to evolutionary-distant large animals, including pigs, than evolutionary-close small animals. Our study provides a molecular foundation during swine lung development that unveils animal size regulation conserved in the COSRP promoter, independent of genome-wide evolution. COSRP is a critical paradigm for studying thousands-fold changes in biological sizes in evolution, development, cancer, zoology, respirology, organoids, and biotechnology, particularly human-compatible organ generation. One Sentence SummaryA cross-species developmental molecular atlas identified the indicator of lung and animal size beyond evolution

evolutionary biology↗

bioRxiv2021

rPanglaoDB: an R package to download and merge labeled single-cell RNA-seq data from the PanglaoDB database

MotivationCharacterizing cells with rare molecular phenotypes is one of the promises of high throughput single-cell RNA sequencing (scRNA-seq) techniques. However, collecting enough cells with the desired molecular phenotype in a single experiment is challenging, requiring several samples preprocessing steps to filter and collect the desired cells experimentally before sequencing. Data integration of multiple public single-cell experiments stands as a solution for this problem, allowing the collection of enough cells exhibiting the desired molecular signatures. By increasing the sample size of the desired cell type, this approach enables a robust cell type transcriptome characterization. ResultsHere, we introduce rPanglaoDB, an R package to download and merge the uniformly processed and annotated scRNA-seq data provided by the PanglaoDB database. To show the potential of rPanglaoDB for collecting rare cell types by integrating multiple public datasets, we present a biological application collecting and characterizing a set of 157 fibrocytes. Fibrocytes are a rare monocyte-derived cell type, that exhibits both the inflammatory features of macrophages and the tissue remodeling properties of fibroblasts. This constitutes the first fibrocytes unbiased transcriptome profile report. We compared the transcriptomic profile of the fibrocytes against the fibroblasts collected from the same tissue samples and confirm their associated relationship with healing processes in tissue damage and infection through the activation of the prostaglandin biosynthesis and regulation pathway. Availability and ImplementationrPanglaoDB is implemented as an R package available through the CRAN repositories https://CRAN.R-project.org/package=rPanglaoDB. Contactdaniecos@uio.no Supplementary informationCode to replicate the case example and figure 1 is available at https://github.com/dosorio/rPanglaoDB O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=161 SRC="FIGDIR/small/446161v1_fig1.gif" ALT="Figure 1"> View larger version (44K): org.highwire.dtl.DTLVardef@b4e67forg.highwire.dtl.DTLVardef@88a46aorg.highwire.dtl.DTLVardef@e25681org.highwire.dtl.DTLVardef@19d5bb7_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 1.C_FLOATNO Characterization of the fibrocytes transcriptome. (A) Identification of cells expressing marker genes that differentiate fibrocytes identity from macrophages and fibroblasts (CD34, ACTA2, FN1, Collagen V, FAP and SIRPA). (B) Cross validation of the identified cells expressing CD34+, ACTA2+, COL5A1+, COL5A2+, COL5A3+, FN1+, FAP+, SIRPA+, PTPRC+, MME+, and SEMA7A+ by kernel density estimation through the Nebulosa package. (C) Volcano plot displaying the differential expression between fibrocytes and the fibroblasts collected in the same merged samples. (D) Enrichment of the prostaglandin biosynthesis and regulation pathway using GSEA through the fgsea package. (E) Enrichment of the prostaglandin biosynthesis and regulation pathway using ssGSEA through the GSVA package. C_FIG

bioinformatics↗

bioRxiv2021

scTenifoldKnk: a machine learning workflow performing virtual knockout experiments on single-cell gene regulatory networks

Bigger PictureO_LIGene knockout (KO) experiments, using genetically altered animals, are a proven powerful approach to elucidate the role of a gene in a biological process. However, systematic KO experiments targeting many genes are usually prohibitive due to limited experimental and animal resources. Here, we present scTenifoldKnk, an efficient virtual KO tool that allows the systematic deletion of many genes individually. scTenifoldKnk uses single-cell RNA sequencing (scRNAseq) data from wild-type (WT) samples to predict gene function in a cell type-specific manner. We show that predictions made by scTenifoldKnk recapitulate findings from real-animal KO experiments. scTenifoldKnk has proven to be a powerful and effective approach for elucidating gene function, prioritizing KO targets, predicting experimental outcomes before real-animal KO experiments are conducted. C_LI HighlightsO_LIscTenifoldKnk performs virtual KO experiments using scRNAseq data. C_LIO_LIscTenifoldKnk only requires data from WT samples; no data is needed from KO samples. C_LIO_LIPredictions made by scTenifoldKnk recapitulate findings from real-animal KO experiments. C_LI Data Science Maturity Level O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=14 SRC="FIGDIR/small/436484v2_ufig1.gif" ALT="Figure 1"> View larger version (8K): org.highwire.dtl.DTLVardef@157126aorg.highwire.dtl.DTLVardef@17a2e76org.highwire.dtl.DTLVardef@805c4dorg.highwire.dtl.DTLVardef@12bdc1b_HPS_FORMAT_FIGEXP M_FIG C_FIG eTOC blurbscTenifoldKnk is a machine learning workflow performing virtual KO experiments to predict gene function. It constructs gene regulatory networks using single-cell RNA sequencing data from wild-type samples and then computationally deletes target genes. Real-data applications demonstrate that scTenifoldKnk recapitulates findings of real-animal KO experiments and accurately predicts gene function in analyzed cells. SummaryGene knockout (KO) experiments are a proven, powerful approach for studying gene function. However, systematic KO experiments targeting a large number of genes are usually prohibitive due to the limit of experimental and animal resources. Here, we present scTenifoldKnk, an efficient virtual KO tool that enables systematic KO investigation of gene function using data from single-cell RNA sequencing (scRNAseq). In scTenifoldKnk analysis, a gene regulatory network (GRN) is first constructed from scRNAseq data of wild-type samples, and a target gene is then virtually deleted from the constructed GRN. Manifold alignment is used to align the resulting reduced GRN to the original GRN to identify differentially regulated genes, which are used to infer target gene functions in analyzed cells. We demonstrate that the scTenifoldKnk-based virtual KO analysis recapitulates the main findings of real-animal KO experiments and recovers the expected functions of genes in relevant cell types.

bioinformatics↗

bioRxiv2020

Systematic determination of the mitochondrial proportion in human and mice tissues for single-cell RNA sequencing data quality control

MotivationQuality control (QC) is a critical step in single-cell RNA-seq (scRNA-seq) data analysis. Low-quality cells are removed from the analysis during the QC process to avoid misinterpretation of the data. One of the important QC metrics is the mitochondrial proportion (mtDNA%), which is used as a threshold to filter out low-quality cells. Early publications in the field established a threshold of 5% and since then, it has been used as a default in several software packages for scRNA-seq data analysis and adopted as a standard in many scRNA-seq studies. However, the validity of using a uniform threshold across different species, single-cell technologies, tissues, and cell types has not been adequately assessed. ResultsWe systematically analyzed 5,530,106 cells reported in 1,349 annotated datasets available in the PanglaoDB database and found that the average mtDNA% in scRNA-seq data across human tissues is significantly higher than in mouse tissues. This difference is not confounded by the platform used to generate the data. Based on this finding, we propose new reference values of the mtDNA% for 121 tissues of mice and 44 tissues of humans. In general, for mouse tissues, the 5% threshold performs well to distinguish between healthy and low-quality cells. However, for human tissues, the 5% threshold should be reconsidered as it fails to accurately discriminate between healthy and low-quality cells in 29.5% (13 of 44) tissues analyzed. We conclude that omitting the mtDNA% QC filter or adopting a suboptimal mtDNA% threshold may lead to erroneous biological interpretations of scRNA-seq data. AvailabilityThe code used to download datasets, perform the analyzes, and produce the figures is available at https://github.com/dosorio/mtProportion Contactdcosorioh@tamu.edu Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2020

scTenifoldNet: a machine learning workflow for constructing and comparing transcriptome-wide gene regulatory networks from single-cell data

Constructing and comparing gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNAseq) data has the potential to reveal critical components in the underlying regulatory networks regulating different cellular transcriptional activities. Here, we present a robust and powerful machine learning workflow--scTenifoldNet--for comparative GRN analysis of single cells. The scTenifoldNet workflow, consisting of principal component regression, low-rank tensor approximation, and manifold alignment, constructs and compares transcriptome-wide single-cell GRNs (scGRNs) from different samples to identify gene expression signatures shifting with cellular activity changes such as those associated with pathophysiological processes and responses to environmental perturbations. We used simulated data to benchmark scTenifoldNets performance, and then applied scTenifoldNet to several real data sets. In real-data applications, scTenifoldNet identified highly specific changes in gene regulation in response to acute morphine treatment, an antibody anticancer drug, gene knockout, double-stranded RNA stimulus, and amyloid-beta plaques in various types of mouse and human cells. We anticipate that scTenifoldNet can help achieve breakthroughs through constructing and comparing scGRNs in poorly characterized biological systems, by deciphering the full cellular and molecular complexity of the data. HighlightsO_LIscTenifoldNet is a machine learning workflow built upon principal component regression, low-rank tensor approximation, and manifold alignment C_LIO_LIscTenifoldNet uses single-cell RNA sequencing (scRNAseq) data to construct single-cell gene regulatory networks (scGRNs) C_LIO_LIscTenifoldNet compares scGRNs of different samples to identify differentially regulated genes C_LIO_LIReal-data applications demonstrate that scTenifoldNet accurately detects specific signatures of gene expression relevant to the cellular systems tested. C_LI Short abstractWe present scTenifoldNet--a machine learning workflow built upon principal component regression, low-rank tensor approximation, and manifold alignment--for constructing and comparing single-cell gene regulatory networks (scGRNs) using data from single-cell RNA sequencing (scRNAseq). scTenifoldNet reveals regulatory changes in gene expression between samples by comparing the constructed scGRNs. With real data, scTenifoldNet identifies specific gene expression programs associated with different biological processes, providing critical insights into the underlying mechanism of regulatory networks governing cellular transcriptional activities.

bioinformatics↗