Bias, robustness and scalability in single-cell differential expression analysis

scCODE: an R package for data-specific differentially expressed gene detection on single-cell RNA-sequencing data

Differential expression (DE) gene detection in single-cell ribonucleic acid (RNA)-sequencing (scRNA-seq) data is a key step to understand the biological question investigated. Filtering genes is suggested to improve the performance of DE methods, but the influence of filtering genes has not been demonstrated. Furthermore, the optimal methods for different scRNA-seq datasets are divergent, and different datasets should benefit from data-specific DE gene detection strategies. However, existing tools did not take gene filtering into consideration. There is a lack of metrics for evaluating the optimal method on experimental datasets. Based on two new metrics, we propose single-cell Consensus Optimization of Differentially Expressed gene detection, an R package to automatically optimize DE gene detection for each experimental scRNA-seq dataset.

hECA: The cell-centric assembly of a cell atlas

The accumulation of massive single-cell omics data provides growing resources for building biomolecular atlases of all cells of human organs or the whole body. The true assembly of a cell atlas should be cell-centric rather than file-centric. We developed a unified informatics framework for seamless cell-centric data assembly and built the human Ensemble Cell Atlas (hECA) from scattered data. hECA v1.0 assembled 1,093,299 labeled human cells from 116 published datasets, covering 38 organs and 11 systems. We invented three new methods of atlas applications based on the cell-centric assembly: “in data” cell sorting for targeted data retrieval with customizable logic expressions, “quantitative portraiture” for multi-view representations of biological entities, and customizable reference creation for generating references for automatic annotations. Case studies on agile construction of user-defined sub-atlases and “in data” investigation of CAR-T off-targets in multiple organs showed the great potential enabled by the cell-centric ensemble atlas.

•

A unified informatics framework for seamless cell-centric assembly of massive single-cell data

•

Built the general-purpose human Ensemble Cell Atlas (hECA) V1.0 from scattered data

•

Three new methods of applications enabling “in data” cell experiments and portraiture

•

Case studies of agile atlas reconstruction and target therapies side-effect discovery

Variant-specific vaccination induces systems immune responses and potent in vivo protection against SARS-CoV-2

Lipid nanoparticle (LNP)-mRNA vaccines offer protection against COVID-19; however, multiple variant lineages caused widespread breakthrough infections. Here, we generate LNP-mRNAs specifically encoding wild-type (WT), B.1.351, and B.1.617 SARS-CoV-2 spikes, and systematically study their immune responses. All three LNP-mRNAs induced potent antibody and T cell responses in animal models; however, differences in neutralization activity have been observed between variants. All three vaccines offer potent protection against in vivo challenges of authentic viruses of WA-1, Beta, and Delta variants. Single-cell transcriptomics of WT- and variant-specific LNP-mRNA-vaccinated animals reveal a systematic landscape of immune cell populations and global gene expression. Variant-specific vaccination induces a systemic increase of reactive CD8 T cells and altered gene expression programs in B and T lymphocytes. BCR-seq and TCR-seq unveil repertoire diversity and clonal expansions in vaccinated animals. These data provide assessment of efficacy and direct systems immune profiling of variant-specific LNP-mRNA vaccination in vivo.

Regulatory analysis of single cell multiome gene expression and chromatin accessibility data with scREG

Technological development has enabled the profiling of gene expression and chromatin accessibility from the same cell. We develop scREG, a dimension reduction methodology, based on the concept of cis-regulatory potential, for single cell multiome data. This concept is further used for the construction of subpopulation-specific cis-regulatory networks. The capability of inferring useful regulatory network is demonstrated by the two-fold increment on network inference accuracy compared to the Pearson correlation-based method and the 27-fold enrichment of GWAS variants for inflammatory bowel disease in the cis-regulatory elements. The R package scREG provides comprehensive functions for single cell multiome data analysis.

The impact of methodological choices when developing predictive models using urinary metabolite data

The continuous evolution of metabolomics over the past two decades has stimulated the search for metabolic biomarkers of many diseases. Metabolomic data measured from urinary samples can provide rich information of the biological events triggered by organ rejection in pediatric kidney transplant recipients. With additional validation, metabolic markers can be used to build clinically useful diagnostic tools. However, there are many methodological steps ranging from data processing to modeling that can influence the performance of the resulting metabolomic classifiers. In this study we focus on the comparison of various classification methods that can handle the complex structure of metabolomic data, including regularized classifiers, partial least squares discriminant analysis, and nonlinear classification models. We also examine the effectiveness of a physiological normalization technique widely used in the clinical and biochemical literature but not extensively analyzed and compared in urine metabolomic studies. While the main objective of this work is to interrogate metabolomic data of pediatric kidney transplant recipients to improve the diagnosis of T cell-mediated rejection (TCMR), we also analyze three independent datasets from other disease conditions to investigate the generalizability of our findings.

Benchmarking of a Bayesian single cell RNAseq differential gene expression test for dose-response study designs

The application of single-cell RNA sequencing (scRNAseq) for the evaluation of chemicals, drugs, and food contaminants presents the opportunity to consider cellular heterogeneity in pharmacological and toxicological responses. Current differential gene expression analysis (DGEA) methods focus primarily on two group comparisons, not multi-group dose-response study designs used in safety assessments. To benchmark DGEA methods for dose-response scRNAseq experiments, we proposed a multiplicity corrected Bayesian testing approach and compare it against 8 other methods including two frequentist fit-for-purpose tests using simulated and experimental data. Our Bayesian test method outperformed all other tests for a broad range of accuracy metrics including control of false positive error rates. Most notable, the fit-for-purpose and standard multiple group DGEA methods were superior to the two group scRNAseq methods for dose-response study designs. Collectively, our benchmarking of DGEA methods demonstrates the importance in considering study design when determining the most appropriate test methods.

A Markov random field model-based approach for differentially expressed gene detection from single-cell RNA-seq data

The development of single-cell RNA-sequencing (scRNA-seq) technologies has offered insights into complex biological systems at the single-cell resolution. In particular, these techniques facilitate the identifications of genes showing cell-type-specific differential expressions (DE). In this paper, we introduce MARBLES, a novel statistical model for cross-condition DE gene detection from scRNA-seq data. MARBLES employs a Markov Random Field model to borrow information across similar cell types and utilizes cell-type-specific pseudobulk count to account for sample-level variability. Our simulation results showed that MARBLES is more powerful than existing methods to detect DE genes with an appropriate control of false positive rate. Applications of MARBLES to real data identified novel disease-related DE genes and biological pathways from both a single-cell lipopolysaccharide mouse dataset with 24 381 cells and 11 076 genes and a Parkinson's disease human data set with 76 212 cells and 15 891 genes. Overall, MARBLES is a powerful tool to identify cell-type-specific DE genes across conditions from scRNA-seq data.

NISC: Neural Network-Imputation for Single-Cell RNA Sequencing and Cell Type Clustering

Single-cell RNA sequencing (scRNA-seq) reveals the transcriptome diversity in heterogeneous cell populations as it allows researchers to study gene expression at single-cell resolution. The latest advances in scRNA-seq technology have made it possible to profile tens of thousands of individual cells simultaneously. However, the technology also increases the number of missing values, i. e, dropouts, from technical constraints, such as amplification failure during the reverse transcription step. The resulting sparsity of scRNA-seq count data can be very high, with greater than 90% of data entries being zeros, which becomes an obstacle for clustering cell types. Current imputation methods are not robust in the case of high sparsity. In this study, we develop a Neural Network-based Imputation for scRNA-seq count data, NISC. It uses autoencoder, coupled with a weighted loss function and regularization, to correct the dropouts in scRNA-seq count data. A systematic evaluation shows that NISC is an effective imputation approach for handling sparse scRNA-seq count data, and its performance surpasses existing imputation methods in cell type identification.

Spatially informed cell-type deconvolution for spatial transcriptomics

Many spatially resolved transcriptomic technologies do not have single-cell resolution but measure the average gene expression for each spot from a mixture of cells of potentially heterogeneous cell types. Here, we introduce a deconvolution method, conditional autoregressive deconvolution (CARD), that combines cell type–specific expression information from single-cell RNA sequencing (scRNA-seq) with correlation in cell type composition across tissue locations. Modeling spatial correlation allows us to borrow the cell-type composition information across locations, improving accuracy of deconvolution even with a mismatched scRNA-seq reference. CARD can also impute cell type compositions and gene expression levels at unmeasured tissue locations, enable the construction of a refined spatial tissue map with a resolution arbitrarily higher than that measured in the original study, and perform deconvolution without a scRNA-seq reference. Applications to four datasets including a pancreatic cancer dataset identified multiple cell types and molecular markers with distinct spatial localization that define the progression, heterogeneity, and compartmentalization of pancreatic cancer.

Selection drives convergent gene expression changes during transitions to co-sexuality in haploid sexual systems

Co-sexuality has evolved repeatedly from unisexual (dioicous) ancestors across a wide range of taxa. However, the molecular changes underpinning this important transition remain unknown, particularly in organisms with haploid sexual systems such as bryophytes, red algae and brown algae. Here we explore four independent events of emergence of co-sexuality from unisexual ancestors in brown algal clades to examine the nature, evolution and degree of convergence of gene expression changes that accompany the breakdown of dioicy. The amounts of male versus female phenotypic differences in dioicous species were not correlated with the extent of sex-biased gene expression, in stark contrast to what is observed in animals. Although sex-biased genes exhibited a high turnover rate during brown alga diversification, some of their predicted functions were conserved across species. Transitions to co-sexuality consistently involved adaptive gene expression shifts and rapid sequence evolution, particularly for male-biased genes. Gene expression in co-sexual species was more similar to that in females rather than males of related dioicous species, suggesting that co-sexuality may have arisen from ancestral females. Finally, extensive convergent gene expression changes, driven by selection, were associated with the transition to co-sexuality. Together, our observations provide insights on how co-sexual systems arise from ancestral, haploid UV sexual systems.

Single-cell characterization of leukemic and non-leukemic immune repertoires in CD8+ T-cell large granular lymphocytic leukemia

T cell large granular lymphocytic leukemia (T-LGLL) is a rare lymphoproliferative disorder of mature, clonally expanded T cells, where somatic-activating STAT3 mutations are common. Although T-LGLL has been described as a chronic T cell response to an antigen, the function of the non-leukemic immune system in this response is largely uncharacterized. Here, by utilizing single-cell RNA and T cell receptor profiling (scRNA+TCRαβ-seq), we show that irrespective of STAT3 mutation status, T-LGLL clonotypes are more cytotoxic and exhausted than healthy reactive clonotypes. In addition, T-LGLL clonotypes show more active cell communication than reactive clones with non-leukemic immune cells via costimulatory cell-cell interactions, monocyte-secreted proinflammatory cytokines, and T-LGLL-clone-secreted IFNγ. Besides the leukemic repertoire, the non-leukemic T cell repertoire in T-LGLL is also more mature, cytotoxic, and clonally restricted than in other cancers and autoimmune disorders. Finally, 72% of the leukemic T-LGLL clonotypes share T cell receptor similarities with their non-leukemic repertoire, linking the leukemic and non-leukemic repertoires together via possible common target antigens. Our results provide a rationale to prioritize therapies that target the entire immune repertoire and not only the T-LGLL clonotype.

acorde unravels functionally interpretable networks of isoform co-usage from single cell data

Alternative splicing (AS) is a highly-regulated post-transcriptional mechanism known to modulate isoform expression within genes and contribute to cell-type identity. However, the extent to which alternative isoforms establish co-expression networks that may be relevant in cellular function has not been explored yet. Here, we present acorde, a pipeline that successfully leverages bulk long reads and single-cell data to confidently detect alternative isoform co-expression relationships. To achieve this, we develop and validate percentile correlations, an innovative approach that overcomes data sparsity and yields accurate co-expression estimates from single-cell data. Next, acorde uses correlations to cluster co-expressed isoforms into a network, unraveling cell type-specific alternative isoform usage patterns. By selecting same-gene isoforms between these clusters, we subsequently detect and characterize genes with co-differential isoform usage (coDIU) across cell types. Finally, we predict functional elements from long read-defined isoforms and provide insight into biological processes, motifs, and domains potentially controlled by the coordination of post-transcriptional regulation. The code for acorde is available at https://github.com/ConesaLab/acorde .

Application of individualized differential expression analysis in human cancer proteome

Liquid chromatography–mass spectrometry-based quantitative proteomics can measure the expression of thousands of proteins from biological samples and has been increasingly applied in cancer research. Identifying differentially expressed proteins (DEPs) between tumors and normal controls is commonly used to investigate carcinogenesis mechanisms. While differential expression analysis (DEA) at an individual level is desired to identify patient-specific molecular defects for better patient stratification, most statistical DEP analysis methods only identify deregulated proteins at the population level. To date, robust individualized DEA algorithms have been proposed for ribonucleic acid data, but their performance on proteomics data is underexplored. Herein, we performed a systematic evaluation on five individualized DEA algorithms for proteins on cancer proteomic datasets from seven cancer types. Results show that the within-sample relative expression orderings (REOs) of protein pairs in normal tissues were highly stable, providing the basis for individualized DEA for proteins using REOs. Moreover, individualized DEA algorithms achieve higher precision in detecting sample-specific deregulated proteins than population-level methods. To facilitate the utilization of individualized DEA algorithms in proteomics for prognostic biomarker discovery and personalized medicine, we provide Individualized DEP Analysis IDEPAXMBD (XMBD: Xiamen Big Data, a biomedical open software initiative in the National Institute for Data Science in Health and Medicine, Xiamen University, China.) (https://github.com/xmuyulab/IDEPA-XMBD), which is a user-friendly and open-source Python toolkit that integrates individualized DEA algorithms for DEP-associated deregulation pattern recognition.

Dissecting the transcriptome in cardiovascular disease

The human transcriptome comprises a complex network of coding and non-coding RNAs implicated in a myriad of biological functions. Non-coding RNAs exhibit highly organized spatial and temporal expression patterns and are emerging as critical regulators of differentiation, homeostasis, and pathological states, including in the cardiovascular system. This review defines the current knowledge gaps, unmet methodological needs, and describes the challenges in dissecting and understanding the role and regulation of the non-coding transcriptome in cardiovascular disease. These challenges include poor annotation of the non-coding genome, determination of the cellular distribution of transcripts, assessment of the role of RNA processing and identification of cell-type specific changes in cardiovascular physiology and disease. We highlight similarities and differences in the hurdles associated with the analysis of the non-coding and protein-coding transcriptomes. In addition, we discuss how the lack of consensus and absence of standardized methods affect reproducibility of data. These shortcomings should be defeated in order to make significant scientific progress and foster the development of clinically applicable non-coding RNA-based therapeutic strategies to lessen the burden of cardiovascular disease.

A Systematic Evaluation of Supervised Machine Learning Algorithms for Cell Phenotype Classification Using Single-Cell RNA Sequencing Data

The new technology of single-cell RNA sequencing (scRNA-seq) can yield valuable insights into gene expression and give critical information about the cellular compositions of complex tissues. In recent years, vast numbers of scRNA-seq datasets have been generated and made publicly available, and this has enabled researchers to train supervised machine learning models for predicting or classifying various cell-level phenotypes. This has led to the development of many new methods for analyzing scRNA-seq data. Despite the popularity of such applications, there has as yet been no systematic investigation of the performance of these supervised algorithms using predictors from various sizes of scRNA-seq datasets. In this study, 13 popular supervised machine learning algorithms for cell phenotype classification were evaluated using published real and simulated datasets with diverse cell sizes. This benchmark comprises two parts. In the first, real datasets were used to assess the computing speed and cell phenotype classification performance of popular supervised algorithms. The classification performances were evaluated using the area under the receiver operating characteristic curve, F1-score, Precision, Recall, and false-positive rate. In the second part, we evaluated gene-selection performance using published simulated datasets with a known list of real genes. The results showed that ElasticNet with interactions performed the best for small and medium-sized datasets. The NaiveBayes classifier was found to be another appropriate method for medium-sized datasets. With large datasets, the performance of the XGBoost algorithm was found to be excellent. Ensemble algorithms were not found to be significantly superior to individual machine learning methods. Including interactions in the ElasticNet algorithm caused a significant performance improvement for small datasets. The linear discriminant analysis algorithm was found to be the best choice when speed is critical; it is the fastest method, it can scale to handle large sample sizes, and its performance is not much worse than the top performers.

Guidelines for bioinformatics of single-cell sequencing data analysis in Alzheimer's disease: review, recommendation, implementation and application

Alzheimer's disease (AD) is the most common form of dementia, characterized by progressive cognitive impairment and neurodegeneration. Extensive clinical and genomic studies have revealed biomarkers, risk factors, pathways, and targets of AD in the past decade. However, the exact molecular basis of AD development and progression remains elusive. The emerging single-cell sequencing technology can potentially provide cell-level insights into the disease. Here we systematically review the state-of-the-art bioinformatics approaches to analyze single-cell sequencing data and their applications to AD in 14 major directions, including 1) quality control and normalization, 2) dimension reduction and feature extraction, 3) cell clustering analysis, 4) cell type inference and annotation, 5) differential expression, 6) trajectory inference, 7) copy number variation analysis, 8) integration of single-cell multi-omics, 9) epigenomic analysis, 10) gene network inference, 11) prioritization of cell subpopulations, 12) integrative analysis of human and mouse sc-RNA-seq data, 13) spatial transcriptomics, and 14) comparison of single cell AD mouse model studies and single cell human AD studies. We also address challenges in using human postmortem and mouse tissues and outline future developments in single cell sequencing data analysis. Importantly, we have implemented our recommended workflow for each major analytic direction and applied them to a large single nucleus RNA-sequencing (snRNA-seq) dataset in AD. Key analytic results are reported while the scripts and the data are shared with the research community through  GitHub. In summary, this comprehensive review provides insights into various approaches to analyze single cell sequencing data and offers specific guidelines for study design and a variety of analytic directions. The review and the accompanied software tools will serve as a valuable resource for studying cellular and molecular mechanisms of AD, other diseases, or biological systems at the single cell level.

CellMeSH: probabilistic cell-type identification using indexed literature

MOTIVATION

Single-cell RNA sequencing (scRNA-seq) is widely used for analyzing gene expression in multi-cellular systems and provides unprecedented access to cellular heterogeneity. scRNA-seq experiments aim to identify and quantify all cell types present in a sample. Measured single-cell transcriptomes are grouped by similarity and the resulting clusters are mapped to cell types based on cluster-specific gene expression patterns. While the process of generating clusters has become largely automated, annotation remains a laborious ad hoc effort that requires expert biological knowledge.

RESULTS

Here, we introduce CellMeSH-a new automated approach to identifying cell types for clusters based on prior literature. CellMeSH combines a database of gene-cell-type associations with a probabilistic method for database querying. The database is constructed by automatically linking gene and cell-type information from millions of publications using existing indexed literature resources. Compared to manually constructed databases, CellMeSH is more comprehensive and is easily updated with new data. The probabilistic query method enables reliable information retrieval even though the gene-cell-type associations extracted from the literature are noisy. CellMeSH is also able to optionally utilize prior knowledge about tissues or cells for further annotation improvement. CellMeSH achieves top-one and top-three accuracies on a number of mouse and human datasets that are consistently better than existing approaches.

AVAILABILITY AND IMPLEMENTATION

Web server at https://uncurl.cs.washington.edu/db_query and API at https://github.com/shunfumao/cellmesh.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

SCRIP: an accurate simulator for single-cell RNA sequencing data

MOTIVATION

Recent advancements in single-cell RNA sequencing (scRNA-seq) have enabled time-efficient transcriptome profiling in individual cells. To optimize sequencing protocols and develop reliable analysis methods for various application scenarios, solid simulation methods for scRNA-seq data are required. However, due to the noisy nature of scRNA-seq data, currently available simulation methods cannot sufficiently capture and simulate important properties of real data, especially the biological variation. In this study, we developed scRNA-seq information producer (SCRIP), a novel simulator for scRNA-seq that is accurate and enables simulation of bursting kinetics.

RESULTS

Compared to existing simulators, SCRIP showed a significantly higher accuracy of stimulating key data features, including mean-variance dependency in all experiments. SCRIP also outperformed other methods in recovering cell-cell distances. The application of SCRIP in evaluating differential expression analysis methods showed that edgeR outperformed other examined methods in differential expression analyses, and ZINB-WaVE improved the AUC at high dropout rates. Collectively, this study provides the research community with a rigorous tool for scRNA-seq data simulation.

AVAILABILITY AND IMPLEMENTATION

https://CRAN.R-project.org/package=SCRIP.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Beware to ignore the rare: how imputing zero-values can improve the quality of 16S rRNA gene studies results

BACKGROUND

16S rRNA-gene sequencing is a valuable approach to characterize the taxonomic content of the whole bacterial population inhabiting a metabolic and spatial niche, providing an important opportunity to study bacteria and their role in many health and environmental mechanisms. The analysis of data produced by amplicon sequencing, however, brings very specific methodological issues that need to be properly addressed to obtain reliable biological conclusions. Among these, 16S count data tend to be very sparse, with many null values reflecting species that are present but got unobserved due to the multiplexing constraints. However, current data workflows do not consider a step in which the information about unobserved species is recovered.

RESULTS

In this work, we evaluate for the first time the effects of introducing in the 16S data workflow a new preprocessing step, zero-imputation, to recover this lost information. Due to the lack of published zero-imputation methods specifically designed for 16S count data, we considered a set of zero-imputation strategies available for other frameworks, and benchmarked them using in silico 16S count data reflecting different experimental designs. Additionally, we assessed the effect of combining zero-imputation and normalization, i.e. the only preprocessing step in current 16S workflow. Overall, we benchmarked 35 16S preprocessing pipelines assessing their ability to handle data sparsity, identify species presence/absence, recovery sample proportional abundance distributions, and improve typical downstream analyses such as computation of alpha and beta diversity indices and differential abundance analysis.

CONCLUSIONS

The results clearly show that 16S data analysis greatly benefits from a properly-performed zero-imputation step, despite the choice of the right zero-imputation method having a pivotal role. In addition, we identify a set of best-performing pipelines that could be a valuable indication for data analysts.

Correlation between targeted RNAseq signature of breast cancer CTCs and onset of bone-only metastases

BACKGROUND

Bone is the most frequent site of metastases from breast cancer (BC), but no biomarkers are yet available to predict skeletal dissemination.

METHODS

We attempted to identify a gene signature correlated with bone metastasis (BM) onset in circulating tumour cells (CTCs), isolated by a DEPArray-based protocol from 40 metastatic BC patients and grouped according to metastasis sites, namely "BM" (bone-only), "ES" (extra-skeletal) or BM + ES (bone + extra-skeletal).

RESULTS

A 134-gene panel was first validated through targeted RNA sequencing (RNAseq) on sub-clones of the MDA-MB-231 BC cell line with variable organotropism, which successfully shaped their clustering. The panel was then applied to CTC groups and, in particular, the "BM" vs "ES" CTC comparison revealed 31 differentially expressed genes, including MAF, CAPG, GIPC1 and IL1B, playing key prognostic roles in BC.

CONCLUSION

Such evidence confirms that CTCs are suitable biological sources for organotropism investigation through targeted RNAseq and might deserve future applications in wide-scale prospective studies.

Robust Transcriptional Profiling and Identification of Differentially Expressed Genes With Low Input RNA Sequencing of Adult Hippocampal Neural Stem and Progenitor Populations

Multipotent neural stem cells (NSCs) are found in several isolated niches of the adult mammalian brain where they have unique potential to assist in tissue repair. Modern transcriptomics offer high-throughput methods for identifying disease or injury associated gene expression signatures in endogenous adult NSCs, but they require adaptation to accommodate the rarity of NSCs. Bulk RNA sequencing (RNAseq) of NSCs requires pooling several mice, which impedes application to labor-intensive injury models. Alternatively, single cell RNAseq can profile hundreds to thousands of cells from a single mouse and is increasingly used to study NSCs. The consequences of the low RNA input from a single NSC on downstream identification of differentially expressed genes (DEGs) remains insufficiently explored. Here, to clarify the role that low RNA input plays in NSC DEG identification, we directly compared DEGs in an oxidative stress model of cultured NSCs by bulk and single cell sequencing. While both methods yielded DEGs that were replicable, single cell sequencing using the 10X Chromium platform yielded DEGs derived from genes with higher relative transcript counts compared to non-DEGs and exhibited smaller fold changes than DEGs identified by bulk RNAseq. The loss of high fold-change DEGs in the single cell platform presents an important limitation for identifying disease-relevant genes. To facilitate identification of such genes, we determined an RNA-input threshold that enables transcriptional profiling of NSCs comparable to standard bulk sequencing and used it to establish a workflow for in vivo profiling of endogenous NSCs. We then applied this workflow to identify DEGs after lateral fluid percussion injury, a labor-intensive animal model of traumatic brain injury. Our work joins an emerging body of evidence suggesting that single cell RNA sequencing may underestimate the diversity of pathologic DEGs. However, our data also suggest that population level transcriptomic analysis can be adapted to capture more of these DEGs with similar efficacy and diversity as standard bulk sequencing. Together, our data and workflow will be useful for investigators interested in understanding and manipulating adult hippocampal NSC responses to various stimuli.

AGAMEMNON: an Accurate metaGenomics And MEtatranscriptoMics quaNtificatiON analysis suite

We introduce AGAMEMNON ( https://github.com/ivlachos/agamemnon ) for the acquisition of microbial abundances from shotgun metagenomics and metatranscriptomic samples, single-microbe sequencing experiments, or sequenced host samples. AGAMEMNON delivers accurate abundances at genus, species, and strain resolution. It incorporates a time and space-efficient indexing scheme for fast pattern matching, enabling indexing and analysis of vast datasets with widely available computational resources. Host-specific modules provide exceptional accuracy for microbial abundance quantification from tissue RNA/DNA sequencing, enabling the expansion of experiments lacking metagenomic/metatranscriptomic analyses. AGAMEMNON provides an R-Shiny application, permitting performance of investigations and visualizations from a graphics interface.

Statistics or biology: the zero-inflation controversy about scRNA-seq data

Researchers view vast zeros in single-cell RNA-seq data differently: some regard zeros as biological signals representing no or low gene expression, while others regard zeros as missing data to be corrected. To help address the controversy, here we discuss the sources of biological and non-biological zeros; introduce five mechanisms of adding non-biological zeros in computational benchmarking; evaluate the impacts of non-biological zeros on data analysis; benchmark three input data types: observed counts, imputed counts, and binarized counts; discuss the open questions regarding non-biological zeros; and advocate the importance of transparent analysis.

Accurate and fast cell marker gene identification with COSG

Accurate cell classification is the groundwork for downstream analysis of single-cell sequencing data, yet how to identify true marker genes for different cell types still remains a big challenge. Here, we report COSine similarity-based marker Gene identification (COSG) as a cosine similarity-based method for more accurate and scalable marker gene identification. COSG is applicable to single-cell RNA sequencing data, single-cell ATAC sequencing data and spatially resolved transcriptome data. COSG is fast and scalable for ultra-large datasets of million-scale cells. Application on both simulated and real experimental datasets showed that the marker genes or genomic regions identified by COSG have greater cell-type specificity, demonstrating the superior performance of COSG in terms of both accuracy and efficiency as compared with other available methods.

Long non-coding RNAs (lncRNAs) NEAT1 and MALAT1 are differentially expressed in severe COVID-19 patients: An integrated single-cell analysis

Hyperactive and damaging inflammation is a hallmark of severe rather than mild Coronavirus disease 2019 (COVID-19). To uncover key inflammatory differentiators between severe and mild COVID-19, we applied an unbiased single-cell transcriptomic analysis. We integrated two single-cell RNA-seq datasets with COVID-19 patient samples, one that sequenced bronchoalveolar lavage (BAL) cells and one that sequenced peripheral blood mononuclear cells (PBMCs). The combined cell population was then analyzed with a focus on genes associated with disease severity. The immunomodulatory long non-coding RNAs (lncRNAs) NEAT1 and MALAT1 were highly differentially expressed between mild and severe patients in multiple cell types. Within those same cell types, the concurrent detection of other severity-associated genes involved in cellular stress response and apoptosis regulation suggests that the pro-inflammatory functions of these lncRNAs may foster cell stress and damage. Thus, NEAT1 and MALAT1 are potential components of immune dysregulation in COVID-19 that may provide targets for severity related diagnostic measures or therapy.

Case Report: Hypoglycemia Due to a Novel Activating Glucokinase Variant in an Adult - a Molecular Approach

We present a case of an obese 22-year-old man with activating GCK variant who had neonatal hypoglycemia, re-emerging with hypoglycemia later in life. We investigated him for asymptomatic hypoglycemia with a family history of hypoglycemia. Genetic testing yielded a novel GCK missense class 3 variant that was subsequently found in his mother, sister and nephew and reclassified as a class 4 likely pathogenic variant. Glucokinase enables phosphorylation of glucose, the rate-limiting step of glycolysis in the liver and pancreatic β cells. It plays a crucial role in the regulation of insulin secretion. Inactivating variants in GCK cause hyperglycemia and activating variants cause hypoglycemia. Spleen-preserving distal pancreatectomy revealed diffuse hyperplastic islets, nuclear pleomorphism and periductular islets. Glucose stimulated insulin secretion revealed increased insulin secretion in response to glucose. Cytoplasmic calcium, which triggers exocytosis of insulin-containing granules, revealed normal basal but increased glucose-stimulated level. Unbiased gene expression analysis using 10X single cell sequencing revealed upregulated INS and CKB genes and downregulated DLK1 and NPY genes in β-cells. Further studies are required to see if alteration in expression of these genes plays a role in the metabolic and histological phenotype associated with glucokinase pathogenic variant. There were more large islets in the patient's pancreas than in control subjects but there was no difference in the proportion of β cells in the islets. His hypoglycemia was persistent after pancreatectomy, was refractory to diazoxide and improved with pasireotide. This case highlights the variable phenotype of GCK mutations. In-depth molecular analyses in the islets have revealed possible mechanisms for hyperplastic islets and insulin hypersecretion.

CellDepot: A unified repository for scRNA-seq data and visual exploration

CellDepot containing over 270 datasets from 8 species and many tissues serves as an integrated web application to empower scientists in exploring single-cell RNA-seq (scRNA-seq) datasets and comparing the datasets among various studies through a user-friendly interface with advanced visualization and analytical capabilities. To begin with, it provides an efficient data management system that users can upload single cell datasets and query the database by multiple attributes such as species and cell types. In addition, the graphical multi-logic, multi-condition query builder and convenient filtering tool backed by MySQL database system, allows users to quickly find the datasets of interest and compare the expression of gene(s) across these. Moreover, by embedding the cellxgene VIP tool, CellDepot enables fast exploration of individual dataset in the manner of interactivity and scalability to gain more refined insights such as cell composition, gene expression profiles, and differentially expressed genes among cell types by leveraging more than 20 frequently applied plotting functions and high-level analysis methods in single cell research. In summary, the web portal available at http://celldepot.bxgenomics.com, prompts large scale single cell data sharing, facilitates meta-analysis and visualization, and encourages scientists to contribute to the single-cell community in a tractable and collaborative way. Finally, CellDepot is released as open-source software under MIT license to motivate crowd contribution, broad adoption, and local deployment for private datasets.

Stem-like intestinal Th17 cells give rise to pathogenic effector T cells during autoimmunity

While intestinal Th17 cells are critical for maintaining tissue homeostasis, recent studies have implicated their roles in the development of extra-intestinal autoimmune diseases including multiple sclerosis. However, the mechanisms by which tissue Th17 cells mediate these dichotomous functions remain unknown. Here, we characterized the heterogeneity, plasticity, and migratory phenotypes of tissue Th17 cells in vivo by combined fate mapping with profiling of the transcriptomes and TCR clonotypes of over 84,000 Th17 cells at homeostasis and during CNS autoimmune inflammation. Inter- and intra-organ single-cell analyses revealed a homeostatic, stem-like TCF1+ IL-17+ SLAMF6+ population that traffics to the intestine where it is maintained by the microbiota, providing a ready reservoir for the IL-23-driven generation of encephalitogenic GM-CSF+ IFN-γ+ CXCR6+ T cells. Our study defines a direct in vivo relationship between IL-17+ non-pathogenic and GM-CSF+ and IFN-γ+ pathogenic Th17 populations and provides a mechanism by which homeostatic intestinal Th17 cells direct extra-intestinal autoimmune disease.

Integrated single-cell transcriptome analysis reveals heterogeneity of esophageal squamous cell carcinoma microenvironment

The tumor microenvironment is a highly complex ecosystem of diverse cell types, which shape cancer biology and impact the responsiveness to therapy. Here, we analyze the microenvironment of esophageal squamous cell carcinoma (ESCC) using single-cell transcriptome sequencing in 62,161 cells from blood, adjacent nonmalignant and matched tumor samples from 11 ESCC patients. We uncover heterogeneity in most cell types of the ESCC stroma, particularly in the fibroblast and immune cell compartments. We identify a tumor-specific subset of CST1⁺ myofibroblasts with prognostic values and potential biological significance. CST1⁺ myofibroblasts are also highly tumor-specific in other cancer types. Additionally, a subset of antigen-presenting fibroblasts is revealed and validated. Analyses of myeloid and T lymphoid lineages highlight the immunosuppressive nature of the ESCC microenvironment, and identify cancer-specific expression of immune checkpoint inhibitors. This work establishes a rich resource of stromal cell types of the ESCC microenvironment for further understanding of ESCC biology.

Human embryonic genome activation initiates at the one-cell stage

In human embryos, the initiation of transcription (embryonic genome activation [EGA]) occurs by the eight-cell stage, but its exact timing and profile are unclear. To address this, we profiled gene expression at depth in human metaphase II oocytes and bipronuclear (2PN) one-cell embryos. High-resolution single-cell RNA sequencing revealed previously inaccessible oocyte-to-embryo gene expression changes. This confirmed transcript depletion following fertilization (maternal RNA degradation) but also uncovered low-magnitude upregulation of hundreds of spliced transcripts. Gene expression analysis predicted embryonic processes including cell-cycle progression and chromosome maintenance as well as transcriptional activators that included cancer-associated gene regulators. Transcription was disrupted in abnormal monopronuclear (1PN) and tripronuclear (3PN) one-cell embryos. These findings indicate that human embryonic transcription initiates at the one-cell stage, sooner than previously thought. The pattern of gene upregulation promises to illuminate processes involved at the onset of human development, with implications for epigenetic inheritance, stem-cell-derived embryos, and cancer.

•

Gene expression initiates at the one-cell stage in human embryos

•

Expression is of low magnitude but remains elevated until the eight-cell stage

•

Upregulated transcripts are spliced and correspond to embryonic processes

•

Upregulation is disrupted in morphologically abnormal one-cell embryos

splatPop: simulating population scale single-cell RNA sequencing data

Population-scale single-cell RNA sequencing (scRNA-seq) is now viable, enabling finer resolution functional genomics studies and leading to a rush to adapt bulk methods and develop new single-cell-specific methods to perform these studies. Simulations are useful for developing, testing, and benchmarking methods but current scRNA-seq simulation frameworks do not simulate population-scale data with genetic effects. Here, we present splatPop, a model for flexible, reproducible, and well-documented simulation of population-scale scRNA-seq data with known expression quantitative trait loci. splatPop can also simulate complex batch, cell group, and conditional effects between individuals from different cohorts as well as genetically-driven co-expression.

CIDER: an interpretable meta-clustering framework for single-cell RNA-seq data integration and evaluation

Clustering of joint single-cell RNA-Seq (scRNA-Seq) data is often challenged by confounding factors, such as batch effects and biologically relevant variability. Existing batch effect removal methods typically require strong assumptions on the composition of cell populations being near identical across samples. Here, we present CIDER, a meta-clustering workflow based on inter-group similarity measures. We demonstrate that CIDER outperforms other scRNA-Seq clustering methods and integration approaches in both simulated and real datasets. Moreover, we show that CIDER can be used to assess the biological correctness of integration in real datasets, while it does not require the existence of prior cellular annotations.

A Comprehensive Survey of Statistical Approaches for Differential Expression Analysis in Single-Cell RNA Sequencing Studies

Single-cell RNA-sequencing (scRNA-seq) is a recent high-throughput sequencing technique for studying gene expressions at the cell level. Differential Expression (DE) analysis is a major downstream analysis of scRNA-seq data. DE analysis the in presence of noises from different sources remains a key challenge in scRNA-seq. Earlier practices for addressing this involved borrowing methods from bulk RNA-seq, which are based on non-zero differences in average expressions of genes across cell populations. Later, several methods specifically designed for scRNA-seq were developed. To provide guidance on choosing an appropriate tool or developing a new one, it is necessary to comprehensively study the performance of DE analysis methods. Here, we provide a review and classification of different DE approaches adapted from bulk RNA-seq practice as well as those specifically designed for scRNA-seq. We also evaluate the performance of 19 widely used methods in terms of 13 performance metrics on 11 real scRNA-seq datasets. Our findings suggest that some bulk RNA-seq methods are quite competitive with the single-cell methods and their performance depends on the underlying models, DE test statistic(s), and data characteristics. Further, it is difficult to obtain the method which will be best-performing globally through individual performance criterion. However, the multi-criteria and combined-data analysis indicates that DECENT and EBSeq are the best options for DE analysis. The results also reveal the similarities among the tested methods in terms of detecting common DE genes. Our evaluation provides proper guidelines for selecting the proper tool which performs best under particular experimental settings in the context of the scRNA-seq.

Uncovering cell identity through differential stability with Cepo

The use of single-cell RNA-sequencing (scRNA-seq) allows observation of different cells at multi-tiered complexity in the same microenvironment. To get insights into cell identity using scRNA-seq data, we present Cepo, which generates cell-type-specific gene statistics of differentially stable genes from scRNA-seq data to define cell identity. When applied to multiple datasets, Cepo outperforms current methods in assigning cell identity and enhances several cell identification applications such as cell-type characterisation, spatial mapping of single cells and lineage inference of single cells.

Feature selection revisited in the single-cell era

Recent advances in single-cell biotechnologies have resulted in high-dimensional datasets with increased complexity, making feature selection an essential technique for single-cell data analysis. Here, we revisit feature selection techniques and summarise recent developments. We review their application to a range of single-cell data types generated from traditional cytometry and imaging technologies and the latest array of single-cell omics technologies. We highlight some of the challenges and future directions and finally consider their scalability and make general recommendations on each type of feature selection method. We hope this review stimulates future research and application of feature selection in the single-cell era.

Differential analysis of binarized single-cell RNA sequencing data captures biological variation

Single-cell RNA sequencing data is characterized by a large number of zero counts, yet there is growing evidence that these zeros reflect biological variation rather than technical artifacts. We propose to use binarized expression profiles to identify the effects of biological variation in single-cell RNA sequencing data. Using 16 publicly available and simulated datasets, we show that a binarized representation of single-cell expression data accurately represents biological variation and reveals the relative abundance of transcripts more robustly than counts.

Discriminative feature of cells characterizes cell populations of interest by a small subset of genes

Organisms are composed of various cell types with specific states. To obtain a comprehensive understanding of the functions of organs and tissues, cell types have been classified and defined by identifying specific marker genes. Statistical tests are critical for identifying marker genes, which often involve evaluating differences in the mean expression levels of genes. Differentially expressed gene (DEG)-based analysis has been the most frequently used method of this kind. However, in association with increases in sample size such as in single-cell analysis, DEG-based analysis has faced difficulties associated with the inflation of P-values. Here, we propose the concept of discriminative feature of cells (DFC), an alternative to using DEG-based approaches. We implemented DFC using logistic regression with an adaptive LASSO penalty to perform binary classification for discriminating a population of interest and variable selection to obtain a small subset of defining genes. We demonstrated that DFC prioritized gene pairs with non-independent expression using artificial data and that DFC enabled characterization of the muscle satellite/progenitor cell population. The results revealed that DFC well captured cell-type-specific markers, specific gene expression patterns, and subcategories of this cell population. DFC may complement DEG-based methods for interpreting large data sets. DEG-based analysis uses lists of genes with differences in expression between groups, while DFC, which can be termed a discriminative approach, has potential applications in the task of cell characterization. Upon recent advances in the high-throughput analysis of single cells, methods of cell characterization such as scRNA-seq can be effectively subjected to the discriminative methods.

Statistical methods for detecting differences in individual gene expression are indispensable for understanding cell types. However, conventional statistical methods, such as differentially expressed gene (DEG)-based analysis, have faced difficulties associated with the inflation of P-values because of both the large sample size and selection bias introduced by exploratory data analysis such as single-cell transcriptomics. Here, we propose the concept of discriminative feature of cells (DFC), an alternative to using DEG-based approaches. We implemented DFC using logistic regression with an adaptive LASSO penalty to perform binary classification for the discrimination of a population of interest and variable selection to obtain a small subset of defining genes. We demonstrated that DFC prioritized gene pairs with non-independent expression using artificial data, and that it enabled characterization of the muscle satellite/progenitor cell population. The results revealed that DFC well captured cell-type-specific markers, specific gene expression patterns, and subcategories of this cell population. DFC may complement differentially expressed gene-based methods for interpreting large data sets.

Maximizing statistical power to detect differentially abundant cell states with scPOST

To estimate a study design's power to detect differential abundance, we require a framework that simulates many multi-sample single-cell datasets. However, current simulation methods are challenging for large-scale power analyses because they are computationally resource intensive and do not support easy simulation of multi-sample datasets. Current methods also lack modeling of important inter-sample variation, such as the variation in the frequency of cell states between samples that is observed in single-cell data. Thus, we developed single-cell POwer Simulation Tool (scPOST) to address these limitations and help investigators quickly simulate multi-sample single-cell datasets. Users may explore a range of effect sizes and study design choices (such as increasing the number of samples or cells per sample) to determine their effect on power, and thus choose the optimal study design for their planned experiments.

scPower accelerates and optimizes the design of multi-sample single cell transcriptomic studies

Single cell RNA-seq has revolutionized transcriptomics by providing cell type resolution for differential gene expression and expression quantitative trait loci (eQTL) analyses. However, efficient power analysis methods for single cell data and inter-individual comparisons are lacking. Here, we present scPower; a statistical framework for the design and power analysis of multi-sample single cell transcriptomic experiments. We modelled the relationship between sample size, the number of cells per individual, sequencing depth, and the power of detecting differentially expressed genes within cell types. We systematically evaluated these optimal parameter combinations for several single cell profiling platforms, and generated broad recommendations. In general, shallow sequencing of high numbers of cells leads to higher overall power than deep sequencing of fewer cells. The model, including priors, is implemented as an R package and is accessible as a web tool. scPower is a highly customizable tool that experimentalists can use to quickly compare a multitude of experimental designs and optimize for a limited budget.

Dietary palmitic acid promotes a prometastatic memory via Schwann cells

Fatty acid uptake and altered metabolism constitute hallmarks of metastasis^1,2, yet evidence of the underlying biology, as well as whether all dietary fatty acids are prometastatic, is lacking. Here we show that dietary palmitic acid (PA), but not oleic acid or linoleic acid, promotes metastasis in oral carcinomas and melanoma in mice. Tumours from mice that were fed a short-term palm-oil-rich diet (PA), or tumour cells that were briefly exposed to PA in vitro, remained highly metastatic even after being serially transplanted (without further exposure to high levels of PA). This PA-induced prometastatic memory requires the fatty acid transporter CD36 and is associated with the stable deposition of histone H3 lysine 4 trimethylation by the methyltransferase Set1A (as part of the COMPASS complex (Set1A/COMPASS)). Bulk, single-cell and positional RNA-sequencing analyses indicate that genes with this prometastatic memory predominantly relate to a neural signature that stimulates intratumoural Schwann cells and innervation, two parameters that are strongly correlated with metastasis but are aetiologically poorly understood^3,4. Mechanistically, tumour-associated Schwann cells secrete a specialized proregenerative extracellular matrix, the ablation of which inhibits metastasis initiation. Both the PA-induced memory of this proneural signature and its long-term boost in metastasis require the transcription factor EGR2 and the glial-cell-stimulating peptide galanin. In summary, we provide evidence that a dietary metabolite induces stable transcriptional and chromatin changes that lead to a long-term stimulation of metastasis, and that this is related to a proregenerative state of tumour-activated Schwann cells.

The NF-κB Transcriptional Footprint Is Essential for SARS-CoV-2 Replication

SARS-CoV-2, the etiological agent of COVID-19, is characterized by a delay in type I interferon (IFN-I)-mediated antiviral defenses alongside robust cytokine production. Here, we investigate the underlying molecular basis for this imbalance and implicate virus-mediated activation of NF-κB in the absence of other canonical IFN-I-related transcription factors. Epigenetic and single-cell transcriptomic analyses show a selective NF-κB signature that was most prominent in infected cells. Disruption of NF-κB signaling through the silencing of the NF-κB transcription factor p65 or p50 resulted in loss of virus replication that was rescued upon reconstitution. These findings could be further corroborated with the use of NF-κB inhibitors, which reduced SARS-CoV-2 replication in vitro. These data suggest that the robust cytokine production in response to SARS-CoV-2, despite a diminished IFN-I response, is the product of a dependency on NF-κB for viral replication.

IMPORTANCE The COVID-19 pandemic has caused significant mortality and morbidity around the world. Although effective vaccines have been developed, large parts of the world remain unvaccinated while new SARS-CoV-2 variants keep emerging. Furthermore, despite extensive efforts and large-scale drug screenings, no fully effective antiviral treatment options have been discovered yet. Therefore, it is of the utmost importance to gain a better understanding of essential factors driving SARS-CoV-2 replication to be able to develop novel approaches to target SARS-CoV-2 biology.

Single-cell normalization and association testing unifying CRISPR screen and gene co-expression analyses with Normalisr

Single-cell RNA sequencing (scRNA-seq) provides unprecedented technical and statistical potential to study gene regulation but is subject to technical variations and sparsity. Furthermore, statistical association testing remains difficult for scRNA-seq. Here we present Normalisr, a normalization and statistical association testing framework that unifies single-cell differential expression, co-expression, and CRISPR screen analyses with linear models. By systematically detecting and removing nonlinear confounders arising from library size at mean and variance levels, Normalisr achieves high sensitivity, specificity, speed, and generalizability across multiple scRNA-seq protocols and experimental conditions with unbiased p-value estimation. The superior scalability allows us to reconstruct robust gene regulatory networks from trans-effects of guide RNAs in large-scale single cell CRISPRi screens. On conventional scRNA-seq, Normalisr recovers gene-level co-expression networks that recapitulated known gene functions.

The multicellular signalling network of ovarian cancer metastases

BACKGROUND

Transcoelomic spread is the major route of metastasis of ovarian high-grade serous carcinoma (HGSC) with the omentum as the major metastatic site. Its unique tumour microenvironment with its large populations of adipocytes, mesothelial cells and immune cells establishes an intercellular signaling network that is instrumental for metastatic growth yet poorly understood.

METHODS

Based on transcriptomic analysis of tumour cells, tumour-associated immune and stroma cells we defined intercellular signaling pathways for 284 cytokines and growth factors and their cognate receptors after bioinformatic adjustment for contaminating cell types. The significance of individual components of this network was validated by analysing clinical correlations and potentially pro-metastatic functions, including tumour cell migration, pro-inflammatory signal transduction and TAM expansion.

RESULTS

The data show an unexpected prominent role of host cells, and in particular of omental adipocytes, mesothelial cells and fibroblasts (CAF), in sustaining this signaling network. These cells, rather than tumour cells, are the major source of most cytokines and growth factors in the omental microenvironment (n = 176 vs. n = 13). Many of these factors target tumour cells, are linked to metastasis and are associated with a short survival. Likewise, tumour stroma cells play a major role in extracellular-matrix-triggered signaling. We have verified the functional significance of our observations for three exemplary instances. We show that the omental microenvironment (i) stimulates tumour cell migration and adhesion via WNT4 which is highly expressed by CAF; (ii) induces pro-tumourigenic TAM proliferation in conjunction with high CSF1 expression by omental stroma cells and (iii) triggers pro-inflammatory signaling, at least in part via a HSP70-NF-κB pathway.

CONCLUSIONS

The intercellular signaling network of omental metastases is majorly dependent on factors secreted by immune and stroma cells to provide an environment that supports ovarian HGSC progression. Clinically relevant pathways within this network represent novel options for therapeutic intervention.

Single-Cell RNA-Sequencing Identifies Infrapatellar Fat Pad Macrophage Polarization in Acute Synovitis/Fat Pad Fibrosis and Cell Therapy

The pathogenesis and progression of knee inflammatory pathologies is modulated partly by residing macrophages in the infrapatellar fat pad (IFP), thus, macrophage polarization towards pro-inflammatory (M1) or anti-inflammatory (M2) phenotypes is important in joint disease pathologies. Alteration of M1/M2 balance contributes to the initiation and progression of joint inflammation and can be potentially altered with mesenchymal stem cell (MSC) therapy. In an acute synovial/IFP inflammation rat model a single intra-articular injection of IFP-MSC was performed, having as controls (1) diseased rats not receiving IFP-MSC and (2) non-diseased rats. After 4 days, cell specific transcriptional profiling via single-cell RNA-sequencing was performed on isolated IFP tissue from each group. Eight transcriptomically distinct cell populations were identified within the IFP across all three treatment groups with a noted difference in the proportion of myeloid cells across the groups. Largely myeloid cells consisted of macrophages (>90%); one M1 sub-cluster highly expressing pro-inflammatory markers and two M2 sub-clusters with one of them expressing higher levels of canonical M2 markers. Notably, the diseased samples (11.9%) had the lowest proportion of cells expressing M2 markers relative to healthy (14.8%) and MSC treated (19.4%) samples. These results suggest a phenotypic polarization of IFP macrophages towards the pro-inflammatory M1 phenotype in an acute model of inflammation, which are alleviated by IFP-MSC therapy inducing a switch towards an alternate M2 status. Understanding the IFP cellular heterogeneity and associated transcriptional programs may offer insights into novel therapeutic strategies for disabling joint disease pathologies.

A Markov random field model for network-based differential expression analysis of single-cell RNA-seq data

BACKGROUND

Recent development of single cell sequencing technologies has made it possible to identify genes with different expression (DE) levels at the cell type level between different groups of samples. In this article, we propose to borrow information through known biological networks to increase statistical power to identify differentially expressed genes (DEGs).

RESULTS

We develop MRFscRNAseq, which is based on a Markov random field (MRF) model to appropriately accommodate gene network information as well as dependencies among cell types to identify cell-type specific DEGs. We implement an Expectation-Maximization (EM) algorithm with mean field-like approximation to estimate model parameters and a Gibbs sampler to infer DE status. Simulation study shows that our method has better power to detect cell-type specific DEGs than conventional methods while appropriately controlling type I error rate. The usefulness of our method is demonstrated through its application to study the pathogenesis and biological processes of idiopathic pulmonary fibrosis (IPF) using a single-cell RNA-sequencing (scRNA-seq) data set, which contains 18,150 protein-coding genes across 38 cell types on lung tissues from 32 IPF patients and 28 normal controls.

CONCLUSIONS

The proposed MRF model is implemented in the R package MRFscRNAseq available on GitHub. By utilizing gene-gene and cell-cell networks, our method increases statistical power to detect differentially expressed genes from scRNA-seq data.