A balanced measure shows superior performance of pseudobulk methods in single-cell RNA-sequencing analysis

Identification and Pharmacological Targeting of Treatment-Resistant, Stem-like Breast Cancer Cells for Combination Therapy

Tumors frequently harbor isogenic yet epigenetically distinct subpopulations of multi-potent cells with high tumor-initiating potential-often called Cancer Stem-Like Cells (CSLCs). These can display preferential resistance to standard-of-care chemotherapy. Single-cell analyses can help elucidate Master Regulator (MR) proteins responsible for governing the transcriptional state of these cells, thus revealing complementary dependencies that may be leveraged via combination therapy. Interrogation of single-cell RNA sequencing profiles from seven metastatic breast cancer patients, using perturbational profiles of clinically relevant drugs, identified drugs predicted to invert the activity of MR proteins governing the transcriptional state of chemoresistant CSLCs, which were then validated by CROP-seq assays. The top drug, the anthelmintic albendazole, depleted this subpopulation in vivo without noticeable cytotoxicity. Moreover, sequential cycles of albendazole and paclitaxel-a commonly used chemotherapeutic -displayed significant synergy in a patient-derived xenograft (PDX) from a TNBC patient, suggesting that network-based approaches can help develop mechanism-based combinatorial therapies targeting complementary subpopulations.

Cell-type and dynamic state govern genetic regulation of gene expression in heterogeneous differentiating cultures

Identifying the molecular effects of human genetic variation across cellular contexts is crucial for understanding the mechanisms underlying disease-associated loci, yet many cell-types and developmental stages remain underexplored. Here we harnessed the potential of heterogeneous differentiating cultures ( HDCs ), an in vitro system in which pluripotent cells asynchronously differentiate into a broad spectrum of cell-types. We generated HDCs for 53 human donors and collected single-cell RNA-sequencing data from over 900,000 cells. We identified expression quantitative trait loci in 29 cell-types and characterized regulatory dynamics across diverse differentiation trajectories. This revealed novel regulatory variants for genes involved in key developmental and disease-related processes while replicating known effects from primary tissues, and dynamic regulatory effects associated with a range of complex traits.

Single cell regulatory architecture of human pancreatic islets suggests sex differences in β cell function and the pathogenesis of type 2 diabetes

Biological sex affects the pathogenesis of type 2 and type 1 diabetes (T2D, T1D) including the development of β cell failure observed more often in males. The mechanisms that drive sex differences in β cell failure is unknown. Studying sex differences in islet regulation and function represent a unique avenue to understand the sex-specific heterogeneity in β cell failure in diabetes. Here, we examined sex and race differences in human pancreatic islets from up to 52 donors with and without T2D (including 37 donors from the Human Pancreas Analysis Program [HPAP] dataset) using an orthogonal series of experiments including single cell RNA-seq (scRNA-seq), single nucleus assay for transposase-accessible chromatin sequencing (snATAC-seq), dynamic hormone secretion, and bioenergetics. In cultured islets from nondiabetic (ND) donors, in the absence of the in vivo hormonal environment, sex differences in islet cell type gene accessibility and expression predominantly involved sex chromosomes. Of particular interest were sex differences in the X-linked KDM6A and Y-linked KDM5D chromatin remodelers in female and male islet cells respectively. Islets from T2D donors exhibited similar sex differences in differentially expressed genes (DEGs) from sex chromosomes. However, in contrast to islets from ND donors, islets from T2D donors exhibited major sex differences in DEGs from autosomes. Comparing β cells from T2D and ND donors revealed that females had more DEGs from autosomes compared to male β cells. Gene set enrichment analysis of female β cell DEGs showed a suppression of oxidative phosphorylation and electron transport chain pathways, while male β cell had suppressed insulin secretion pathways. Thus, although sex-specific differences in gene accessibility and expression of cultured ND human islets predominantly affect sex chromosome genes, major differences in autosomal gene expression between sexes appear during the transition to T2D and which highlight mitochondrial failure in female β cells.

Modulator of TMB-associated immune infiltration (MOTIF) predicts immunotherapy response and guides combination therapy

Patients with high tumor mutational burden (TMB) levels do not consistently respond to immune checkpoint inhibitors (ICIs), possibly because a high TMB level does not necessarily result in adequate infiltration of CD8+ T cells. Using bulk ribonucleic acid sequencing (RNA-seq) data from 9311 tumor samples across 30 cancer types, we developed a novel tool called the modulator of TMB-associated immune infiltration (MOTIF), which comprises genes that can determine the extent of CD8+ T cell infiltration prompted by a certain TMB level. We confirmed that MOTIF can accurately reflect the integrity and defects of the cancer-immunity cycle. By analyzing 84 human single-cell RNA-seq datasets from 32 types of solid tumors, we revealed that MOTIF can provide insights into the diverse roles of various cell types in the modulation of CD8+ T cell infiltration. Using pretreatment RNA-seq data from 13 ICI-treated cohorts, we validated the use of MOTIF in predicting CD8+ T cell infiltration and ICI efficacy. Among the components of MOTIF, we identified EMC3 as a negative regulator of CD8+ T cell infiltration, which was validated via in vivo studies. Additionally, MOTIF provided guidance for the potential combinations of programmed death 1 blockade with certain immunostimulatory drugs to facilitate CD8+ T cell infiltration and improve ICI efficacy.

Functional impairment of "helpless" CD8 + memory T cells is transient and driven by prolonged but finite cognate antigen presentation

Generation of functional CD8 + T cell memory typically requires engagement of CD4 + T cells. However, in certain scenarios, such as acutely-resolving viral infections, effector (T E ) and subsequent memory (T M ) CD8 + T cell formation appear impervious to a lack of CD4 + T cell help during priming. Nonetheless, such "helpless" CD8 + T M respond poorly to pathogen rechallenge. At present, the origin and long-term evolution of helpless CD8 + T cell memory remain incompletely understood. Here, we demonstrate that helpless CD8 + T E differentiation is largely normal but a multiplicity of helpless CD8 T M defects, consistent with impaired memory maturation, emerge as a consequence of prolonged yet finite exposure to cognate antigen. Importantly, these defects resolve over time leading to full restoration of CD8 + T M potential and recall capacity. Our findings provide a unified explanation for helpless CD8 + T cell memory and emphasize an unexpected CD8 + T M plasticity with implications for vaccination strategies and beyond.

Differential detection workflows for multi-sample single-cell RNA-seq data

In single-cell transcriptomics, differential gene expression (DE) analyses typically focus on testing differences in the average expression of genes between cell types or conditions of interest. Single-cell transcriptomics, however, also has the promise to prioritise genes for which the expression differ in other aspects of the distribution. Here we develop a workflow for assessing differential detection (DD), which tests for differences in the average fraction of samples or cells in which a gene is detected. After benchmarking eight different DD data analysis strategies, we provide a unified workflow for jointly assessing DE and DD. Using simulations and two case studies, we show that DE and DD analysis provide complementary information, both in terms of the individual genes they report and in the functional interpretation of those genes.

Avoiding false discoveries in single-cell RNA-seq by revisiting the first Alzheimer's disease dataset

Mathys et al. conducted the first single-nucleus RNA-seq (snRNA-seq) study of Alzheimer's disease (AD) (Mathys et al., 2019). With bulk RNA-seq, changes in gene expression across cell types can be lost, potentially masking the differentially expressed genes (DEGs) across different cell types. Through the use of single-cell techniques, the authors benefitted from increased resolution with the potential to uncover cell type-specific DEGs in AD for the first time. However, there were limitations in both their data processing and quality control and their differential expression analysis. Here, we correct these issues and use best-practice approaches to snRNA-seq differential expression, resulting in 549 times fewer DEGs at a false discovery rate of 0.05. Thus, this study highlights the impact of quality control and differential analysis methods on the discovery of disease-associated genes and aims to refocus the AD research field away from spuriously identified genes.

NRTPredictor: identifying rice root cell state in single-cell RNA-seq via ensemble learning

BACKGROUND

Single-cell RNA sequencing (scRNA-seq) measurements of gene expression show great promise for studying the cellular heterogeneity of rice roots. How precisely annotating cell identity is a major unresolved problem in plant scRNA-seq analysis due to the inherent high dimensionality and sparsity.

RESULTS

To address this challenge, we present NRTPredictor, an ensemble-learning system, to predict rice root cell stage and mine biomarkers through complete model interpretability. The performance of NRTPredictor was evaluated using a test dataset, with 98.01% accuracy and 95.45% recall. With the power of interpretability provided by NRTPredictor, our model recognizes 110 marker genes partially involved in phenylpropanoid biosynthesis. Expression patterns of rice root could be mapped by the above-mentioned candidate genes, showing the superiority of NRTPredictor. Integrated analysis of scRNA and bulk RNA-seq data revealed aberrant expression of Epidermis cell subpopulations in flooding, Pi, and salt stresses.

CONCLUSION

Taken together, our results demonstrate that NRTPredictor is a useful tool for automated prediction of rice root cell stage and provides a valuable resource for deciphering the rice root cellular heterogeneity and the molecular mechanisms of flooding, Pi, and salt stresses. Based on the proposed model, a free webserver has been established, which is available at https://www.cgris.net/nrtp .

CRISPR screening in hematology research: from bulk to single-cell level

The CRISPR genome editing technology has revolutionized the way gene function is studied. Genome editing can be achieved in single genes or for thousands of genes simultaneously in sensitive genetic screens. While conventional genetic screens are limited to bulk measurements of cell behavior, recent developments in single-cell technologies make it possible to combine CRISPR screening with single-cell profiling. In this way, cell behavior and gene expression can be monitored simultaneously, with the additional possibility of including data on chromatin accessibility and protein levels. Moreover, the availability of various Cas proteins leading to inactivation, activation, or other effects on gene function further broadens the scope of such screens. The integration of single-cell multi-omics approaches with CRISPR screening open the path to high-content information on the impact of genetic perturbations at single-cell resolution. Current limitations in cell throughput and data density need to be taken into consideration, but new technologies are rapidly evolving and are likely to easily overcome these limitations. In this review, we discuss the use of bulk CRISPR screening in hematology research, as well as the emergence of single-cell CRISPR screening and its added value to the field.

Challenges in the real world use of classification accuracy metrics: From recall and precision to the Matthews correlation coefficient

The accuracy of a classification is fundamental to its interpretation, use and ultimately decision making. Unfortunately, the apparent accuracy assessed can differ greatly from the true accuracy. Mis-estimation of classification accuracy metrics and associated mis-interpretations are often due to variations in prevalence and the use of an imperfect reference standard. The fundamental issues underlying the problems associated with variations in prevalence and reference standard quality are revisited here for binary classifications with particular attention focused on the use of the Matthews correlation coefficient (MCC). A key attribute claimed of the MCC is that a high value can only be attained when the classification performed well on both classes in a binary classification. However, it is shown here that the apparent magnitude of a set of popular accuracy metrics used in fields such as computer science medicine and environmental science (Recall, Precision, Specificity, Negative Predictive Value, J, F1, likelihood ratios and MCC) and one key attribute (prevalence) were all influenced greatly by variations in prevalence and use of an imperfect reference standard. Simulations using realistic values for data quality in applications such as remote sensing showed each metric varied over the range of possible prevalence and at differing levels of reference standard quality. The direction and magnitude of accuracy metric mis-estimation were a function of prevalence and the size and nature of the imperfections in the reference standard. It was evident that the apparent MCC could be substantially under- or over-estimated. Additionally, a high apparent MCC arose from an unquestionably poor classification. As with some other metrics of accuracy, the utility of the MCC may be overstated and apparent values need to be interpreted with caution. Apparent accuracy and prevalence values can be mis-leading and calls for the issues to be recognised and addressed should be heeded.

Bayesian-frequentist hybrid inference framework for single cell RNA-seq analyses

Background

Single cell RNA sequencing technology (scRNA-seq) has been proven useful in understanding cell-specific disease mechanisms. However, identifying genes of interest remains a key challenge. Pseudo-bulk methods that pool scRNA-seq counts in the same biological replicates have been commonly used to identify differentially expressed genes. However, such methods may lack power due to the limited sample size of scRNA-seq datasets, which can be prohibitively expensive.

Results

Motivated by this, we proposed to use the Bayesian-frequentist hybrid (BFH) framework to increase the power.

Conclusion

In our idiopathic pulmonary fibrosis (IPF) case study, we demonstrated that with a proper informative prior, the BFH approach identified more genes of interest. Furthermore, these genes were reasonable based on the current knowledge of IPF. Thus, the BFH offers a unique and flexible framework for future scRNA-seq analyses.

Multi-omics segregate different transcriptomic impacts of anti-IL-17A blockade on type 17 T-cells and regulatory immune cells in psoriasis skin

Durable psoriasis improvement has been reported in a subset of psoriasis patients after treatment withdrawal of biologics blocking IL-23/Type 17 T-cell (T17) autoimmune axis. However, it is not well understood if systemic blockade of the IL-23/T17 axis promotes immune tolerance in psoriasis skin. The purpose of the study was to find translational evidence that systemic IL-17A blockade promotes regulatory transcriptome modification in human psoriasis skin immune cell subsets. We analyzed human psoriasis lesional skin 6 mm punch biopsy tissues before and after systemic IL-17A blockade using the muti-genomics approach integrating immune cell-enriched scRNA-seq (n = 18), microarray (n = 61), and immunohistochemistry (n = 61) with repository normal control skin immune cell-enriched scRNA-seq (n = 10) and microarray (n = 8) data. For the T17 axis transcriptome, systemic IL-17A blockade depleted 100% of IL17A + T-cells and 95% of IL17F + T-cells in psoriasis skin. The expression of IL23A in DC subsets was also downregulated by IL-17A blockade. The expression of IL-17-driven inflammatory mediators (IL36G, S100A8, DEFB4A, and DEFB4B) in suprabasal keratinocytes was correlated with psoriasis severity and was downregulated by IL-17A blockade. For the regulatory DC transcriptome, the proportion of regulatory semimature DCs expressing regulatory DC markers of BDCA-3 (THBD) and DCIR (CLEC4A) was increased in posttreatment psoriasis lesional skin compared to pretreatment psoriasis lesional skin. In addition, IL-17A blockade induced higher expression of CD1C and CD14, which are markers of CD1c+ CD14+ dendritic cell (DC) subset that suppresses antigen-specific T-cell responses, in posttreatment regulatory semimature DCs compared to pretreatment regulatory semimature DCs. In conclusion, systemic IL-17A inhibition not only blocks the entire IL-23/T17 cell axis but also promotes regulatory gene expression in regulatory DCs in human psoriasis skin.

Modeling group heteroscedasticity in single-cell RNA-seq pseudo-bulk data

Group heteroscedasticity is commonly observed in pseudo-bulk single-cell RNA-seq datasets and its presence can hamper the detection of differentially expressed genes. Since most bulk RNA-seq methods assume equal group variances, we introduce two new approaches that account for heteroscedastic groups, namely voomByGroup and voomWithQualityWeights using a blocked design (voomQWB). Compared to current gold-standard methods that do not account for group heteroscedasticity, we show results from simulations and various experiments that demonstrate the superior performance of voomByGroup and voomQWB in terms of error control and power when group variances in pseudo-bulk single-cell RNA-seq data are unequal.

Gene function and cell surface protein association analysis based on single-cell multiomics data

Single-cell transcriptomics provides researchers with a powerful tool to resolve the transcriptome heterogeneity of individual cells. However, this method falls short in revealing cellular heterogeneity at the protein level. Previous single-cell multiomics studies have focused on data integration rather than exploiting the full potential of multiomics data. Here we introduce a new analysis framework, gene function and protein association (GFPA), that mines reliable associations between gene function and cell surface protein from single-cell multimodal data. Applying GFPA to human peripheral blood mononuclear cells (PBMCs), we observe an association of epithelial mesenchymal transition (EMT) with the CD99 protein in CD4 T cells, which is consistent with previous findings. Our results show that GFPA is reliable across multiple cell subtypes and PBMC samples. The GFPA python packages and detailed tutorials are freely available at https://github.com/studentiz/GFPA.