SEARCH · Search bioRxiv

Results for “single cell RNA sequencing”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 217 records · Page 12Linked to original sources

bioRxiv2024

A single cell RNA sequence atlas of the early Drosophila larval eye

The Drosophila eye has been an important model to understand principles of differentiation, proliferation, apoptosis and tissue morphogenesis. However, a single cell RNA sequence resource that captures gene expression dynamics from the initiation of differentiation to the specification of different cell types in the larval eye disc is lacking. Here, we report transcriptomic data from 13,000 cells that cover six developmental stages of the larval eye. Our data show cell clusters that correspond to all major cell types present in the eye disc ranging from the initiation of the morphogenetic furrow to the differentiation of each photoreceptor cell type as well as early cone cells. We identify dozens of cell type-specific genes whose function in different aspects of eye development have not been reported. These single cell data will greatly aid research groups studying different aspects of early eye development and will facilitate a deeper understanding of the larval eye as a model system.

genomics↗

bioRxiv2024

CoupleVAE: coupled variational autoencoders for predicting perturbational single-cell RNA sequencing data

With the rapid advances in single-cell sequencing technology, it is now feasible to conduct in-depth genetic analysis in individual cells. Study on the dynamics of single cells in response to perturbations is of great significance for understanding the functions and behaviours of living organisms. However, the acquisition of post-perturbation cellular states via biological experiments is frequently cost-prohibitive. Predicting the single-cell perturbation responses poses a critical challenge in the field of computational biology. In this work, we propose a novel deep learning method called coupled variational autoencoders (CoupleVAE), devised to predict the post-perturbation single-cell RNA-Seq data. CoupleVAE is composed of two coupled VAEs connected by a coupler, initially extracting latent features for both controlled and perturbed cells via two encoders, subsequently engaging in mutual translation within the latent space through two nonlinear mappings via a coupler, and ultimately generating controlled and perturbed data by two separate decoders to process the encoded and translated features. CoupleVAE facilitates a more intricate state transformation of single cells within the latent space. Experiments in three real datasets on infection, stimulation and cross-species prediction show that CoupleVAE surpasses the existing comparative models in effectively predicting single-cell RNA-seq data for perturbed cells, achieving superior accuracy.

bioinformatics↗

bioRxiv2019

SciBet: An ultra-fast classifier for cell type identification using single cell RNA sequencing data

Fast, robust and technology-independent computational methods are needed for supervised cell type annotation of single-cell RNA sequencing data. We present SciBet, a Bayesian classifier that accurately predicts cell identity for newly sequenced cells or cell clusters. We enable web client deployment of SciBet for rapid local computation without uploading local data to the server. This user-friendly and cross-platform tool can be widely useful for single cell type identification.

bioinformatics↗

bioRxiv2026

Limitations of differential expression for cell-type marker discovery in single-cell RNA sequencing atlases

Cell-type marker genes are definitional characteristics of a cell type, and in single-cell RNA sequencing atlases they are commonly nominated through differential expression (DE) testing. The Wilcoxon rank-sum test is the default method for marker gene discovery in Seurat and a recommended method in Scanpy, two widely used computational software packages for scRNA-seq analysis. Structurally, Wilcoxon's standardized statistic grows with sample size at a fixed underlying effect, meaning computed rankings across clusters of unequal size are incomparable. We quantified this effect across three atlas datasets - lung, kidney, and brain. Mean Wilcoxon score scaled with log cluster size at Pearson correlation r = 0.90-0.95 in each organ. In contrast, a classification-based comparator, NS-Forest, showed low correlation with cluster size (r = 0.10-0.28). This DE-cluster size relationship remained essentially unchanged after controlling for classification performance through partial correlation, decayed monotonically under cluster size downsampling, and was reproduced in simulation with marker quality fixed. The Wilcoxon rank-sum test can be transformed into a rank-based effect size by normalizing by the target-background pair count, removing the positive size dependence. However, this effect size nominates an identical set of candidate genes as the original statistic, indicating the cluster size dependence and poor specificity of DE nominated genes are separate problems. Genes nominated under default settings were substantially less specific to their target clusters, with median on-target fractions of 0.204 (lung), 0.166 (kidney), and 0.083 (brain), compared to NS-Forest's marker genes (0.580, 0.687, 0.364), revealing the relatively poor specificity that highly scoring DE genes exhibit.

bioinformatics↗

bioRxiv2025

Single-cell RNA Sequencing Analysis of Sputum Cell Transcriptomes Reveals Pathways and Communication Networks That Contribute to the Pathogenesis of Asthma

BackgroundAsthma is driven by complex interactions amongst structural airway cells, cells of the immune system, and the environmental. While sputum cell characterization has been instrumental in studying asthma pathogenesis and refining treatment strategies, the nuances of cellular transcriptomes and intercellular communication in asthmatic sputum remain poorly understood. MethodsWe employed single-cell RNA sequencing to analyze cells isolated form the sputum from 16 asthma patients and 8 non-asthmatic controls. Cell identities were established using curated marker genes and SingleR annotation. We compared cell-specific gene expression and communication networks between asthmatic and control groups, correlating findings with distinct pathways that were dysregulated in asthma. Findings37,565 cellular transcriptomes were captured and analyzed. 15 distinct cell populations were identified, including various macrophages, monocytes, dendritic cells, and lymphocytes, along with rare cell types such as mast cells, innate lymphoid cells, bronchial epithelial cells, and eosinophils. Intercellular communication analysis indicated heightened signaling activity in asthma compared to controls, particularly in CD4+ T cells and dendritic cells which exhibited the most significant increases in RNA expression of outgoing signaling molecules. Notably, the ADAM12-SDC4 and CCL22-CCR4 ligand-receptor pathways demonstrated the strongest shifts between asthma and control subjects, particularly between dendritic cells and CD4 lymphocytes. InterpretationSC RNA seq profiling the asthma cellular transcriptome analysis of sputum highlights both innate and adaptive immune mechanisms that are significantly amplified in asthma. The elevated expression of ADAM12-SCD4 and CCL22-CC4 point to their critical role in asthma pathogenesis, suggesting potential avenues for targeted therapies and improved management of this chronic condition. Research in contextO_ST_ABSEvidence Before This StudyC_ST_ABSAsthma is a chronic inflammatory disease of the airways driven by intricate interactions between airway structural and immune cells. Previous transcriptomic studies have focused on bulk RNA samples from the airway, leaving significant gaps in our understanding of the cellular dynamics that characterize the disease. Added Value of This StudyThis study pioneers the use of single-cell RNA sequencing on sputum samples from patients with asthma, revealing a detailed landscape of cell phenotypes and dynamic communication patterns that distinguish asthmatic individuals from those without the disease. Notably, heightened intercellular communication was observed in asthma, particularly between CD4+ T cells and dendritic cells, confirming that there is a robust network of interactions between immune and structural cells. The notable increase of ADAM12-CCR4 communication from dendritic cells to other cell populations further emphasizes the dysregulation present in asthma. Implications of All Available EvidenceOur transcriptomic profiling illuminates distinct and amplified communication pathways involving CD4+ T cells and dendritic cells, aligning with established paradigms of both adaptive and innate immune responses in asthma pathogenesis. The identification of ADAM12 and CCR4 pathway dysregulation adds a critical layer to our understanding of the molecular mechanisms underpinning asthma, paving the way for potential therapeutic targets and personalized treatment strategies. Single cell profiling of the sputum has the capacity to characterize the breadth of cellular phenotypes, their functional status, and the communication in the airway at a level not previously attainable.

genomics↗

bioRxiv2025

DeepSeq: High-Throughput Single-Cell RNA Sequencing Data Labeling via Web Search-Augmented Agentic Generative AI Foundation Models

Generative AI foundation models offer transformative potential for processing structured biological data, particularly in single-cell RNA sequencing, where datasets are rapidly scaling toward billions of cells. We propose the use of agentic generative AI foundation models with real-time web search to automate the labeling of experimental data, achieving up to 82.5% accuracy. This addresses a key bottleneck in supervised learning for structured omics data by increasing annotation throughput without manual curation and human error. Our approach enables the development of virtual cell foundation models capable of downstream tasks such as cell-typing and perturbation prediction. As data volume grows, these models may surpass human performance in labeling, paving the way for reliable inference in large-scale perturbation screens. This application demonstrates domain-specific innovation in health monitoring and diagnostics, aligned with efforts like the Human Cell Atlas and Human Tumor Atlas Network.

bioinformatics↗

bioRxiv2024

Single-Cell Omics for Transcriptome CHaracterization (SCOTCH): isoform-level characterization of gene expression through long-read single-cell RNA sequencing

Recent advances in long-read single-cell transcriptome sequencing (lr-scRNA-Seq) enable full-length isoform profiling at single-cell resolution. We present SCOTCH (Single-Cell Omics for Transcriptome CHaracterization), an end-to-end, platform-independent pipeline for isoform characterization from lr-scRNA-Seq data, supporting Nanopore and PacBio sequencing as well as 10X Genomics and Parse Biosciences protocols. SCOTCH models isoforms as combinations of non-overlapping sub-exons and applies dynamic thresholding for robust isoform assignment while efficiently address ambiguous mapping issues. By refining sub-exon boundaries through integration of read coverage with existing annotations and applying an iterative clustering strategy to reconstruct novel transcripts, SCOTCH reliably recovers more true novel isoforms than existing splice-graph-based methods, with poly(A)-aware filtering further reducing false-positive structures. Extensive simulations demonstrate improved quantification of known isoforms and enhanced reconstruction of novel isoforms. Analyses of human blood and cerebral organoid datasets across multiple platforms further confirm SCOTCHs ability to resolve cell-type-specific transcriptome profiles and uncover experimentally supported novel isoforms.

bioinformatics↗

bioRxiv2024

Shrinkage estimation of gene interaction networks in single-cell RNA sequencing data

Gene interaction networks are graphs in which nodes represent genes and edges represent functional interactions between them. These interactions can be at multiple levels, for instance, gene regulation, protein-protein interaction, or metabolic pathways. To analyse gene interaction networks at a large scale, gene co-expression network analysis is often applied on high-throughput gene expression data such as RNA sequencing data. With the advance in sequencing technology, expression of genes can be measured in individual cells. Single-cell RNA sequencing (scRNAseq) provides insights of cellular development, differentiation and characteristics at transcriptomic level. High sparsity and high-dimensional data structure pose challenges in scRNAseq data analysis. In this study, a sparse inverse covariance matrix estimation framework for scRNAseq data is developed to capture direct functional interactions between genes. Comparative analyses highlight high performance and fast computation of Stein-type shrinkage in high-dimensional data using simulated scRNAseq data. Data transformation approaches also show improvement in performance of shrinkage methods in non-Gaussian distributed data. Zero-inflated modelling of scRNAseq data based on a negative binomial distribution enhances shrinkage performance in zero-inflated data without interference on non zeroinflated count data. The optimal zero-inflated Stein-type shrinkage framework is applied on experimental scRNAseq data which demonstrates its potential to construct sparser gene interaction networks with higher precision. Availability and implementationhttps://github.com/calathea24/ZINBGraphicalModel

systems biology↗

bioRxiv2021

A comprehensive temporal patterning gene network in Drosophila medulla neuroblasts revealed by single-cell RNA sequencing

During development, neural stem cells are temporally patterned to sequentially generate a variety of neural types before exiting the cell cycle. Temporal patterning is well-studied in Drosophila, where neural stem cells called neuroblasts sequentially express cascades of Temporal Transcription Factors (TTFs) to control the birth-order dependent neural specification. However, currently known TTFs were mostly identified through candidate approaches and may not be complete. In addition, many fundamental questions remain concerning the TTF cascade initiation, progression, and termination. It is also not known why temporal progression only happens in neuroblasts but not in their differentiated progeny. In this work, we performed single-cell RNA sequencing of Drosophila medulla neuroblasts of all ages to study the temporal patterning process with single-cell resolution. Our scRNA-seq data revealed that sets of genes involved in different biological processes show high to low or low to high gradients as neuroblasts age. We also identified a list of novel TTFs, and experimentally characterized their roles in the temporal progression and neural fate specification. Our study revealed a comprehensive temporal gene network that patterns medulla neuroblasts from start to end. Furthermore, we found that the progression and termination of this temporal cascade also require transcription factors differentially expressed along the differentiation axis (neuroblasts -> -> neurons). Lola proteins function as a speed modulator of temporal progression in neuroblasts; while Nerfin-1, a factor required to suppress de-differentiation in post-mitotic neurons, acts at the final temporal stage together with the last TTF of the cascade, to promote the switch to gliogenesis and the cell cycle exit. Our comprehensive study of the medulla neuroblast temporal cascade illustrated mechanisms that might be conserved in the temporal patterning of neural stem cells.

developmental biology↗

bioRxiv2017

Heterogeneity of hypothalamic Pro-opiomelanocortin-expressing neurons revealed by single-cell RNA sequencing

Arcuate proopiomelanocortin (POMC) neurons are critical nodes in the control of body weight. Often characterised simply as direct targets for leptin, recent data suggest a more complex architecture. Using single cell RNA sequencing, we have generated an atlas of gene expression in murine POMC neurons. Of 163 neurons, 118 expressed high levels of Pomc with little/no Agrp expression and were considered \"canonical\" POMC neurons (P+). The other 45/163 expressed low levels of Pomc and high levels of Agrp (A+P+). Unbiased clustering analysis of P+ neurons revealed four different classes, each with distinct cell surface receptor gene expression profiles. Further, only 12% (14/118) of P+ neurons expressed the leptin receptor (Lepr) compared with 58% (26/45) of A+P+ neurons. In contrast, the insulin receptor (Insr) was expressed at similar frequency on P+ and A+P+ neurons (64% and 55%, respectively). These data reveal arcuate POMC neurons to be a highly heterogeneous population.

neuroscience↗

bioRxiv2026

Mapping the response of primary human chondrocytes to an osteoarthritis-relevant stimulus by single-cell RNA sequencing

Objective: To determine how fibronectin fragment (Fn-f), a matrikine that induces an osteoarthritis (OA)-like phenotype, reshapes chondrocyte states and transcriptional programs at single-cell resolution. Design: Primary human ankle chondrocytes from two cadaveric donors without OA were treated with PBS or purified recombinant Fn-f for 18 hours and profiled by single-cell RNA sequencing. Transcriptional states were characterized by marker genes and functional enrichment, compared with classifications used for OA chondrocytes, and evaluated for Fn-f-dependent changes in abundance and transcriptional programs. Results: Five transcriptionally distinct populations were identified with features that matched previous classifications of effector, inflammatory, pre-hypertrophic/fibrochondrocyte, reparative, and homeostatic chondrocytes. Fn-f markedly redistributed cells among these states, enriching inflammatory chondrocytes while reducing effector and reparative populations. Within transcriptional states, Fn-f differentially altered matrix/cartilage, inflammatory, and stress/immediate-early programs. These responses did not uniformly correspond to assigned subtype identities, demonstrating that transcriptional programs can be dynamically deployed across chondrocyte states. Conclusions: Controlled perturbations of healthy chondrocytes complement observational OA atlases by resolving stimulus-dependent changes in chondrocyte state abundance and transcriptional programs. Discordance between subtype annotations and dynamic functional programs further highlights the need for harmonized definitions of chondrocytes across phenotypic states.

genetics↗

bioRxiv2025

Exploring the immune environment of glioblastoma through single-cell RNA sequencing in humanized mouse models

BackgroundGlioblastoma (GBM) is the deadliest primary brain tumor in adults, where current therapies fail to meaningfully extend survival. Available animal GBM tumor models, especially therapy-resistant and recurrent ones with unique immunological aspects, are restricted, impeding innovative treatment research. To confront this critical obstacle we established a unique GBM mouse model that utilizes patient-derived xenografts (PDXs) within humanized mice. MethodsWe selected two immune-deficient mouse models to facilitate the reconstitution of myeloid lineage cells. After undergoing myeloablation, mice received CD34+ hematopoietic stem progenitor cells derived from human umbilical cord blood for humanization. Upon confirming the reconstitution of human blood cells, mice were xenografted with PDXs resistant to radiation. Tumor profiles and immune cell infiltration were analyzed via flow cytometry, immunohistochemistry, and single-cell RNA sequencing (scRNA-seq). The findings were evaluated against scRNA-seq data from recurrent human GBM. ResultsA diverse range of human immune cells, including T, NK, and myeloid lineage cells, infiltrated PDX tumors in humanized mice. Notably, gene expression profiles in these immune cells resembled that of recurrent human GBM. Unlike conventional xenograft models, this model highlighted enhanced tumor diversity, particularly a high fraction of neural progenitor-like cells. ConclusionsOur humanized GBM mouse model displayed an immune cell signature similar to recurrent GBM. This model is a valuable resource for analyzing the tumor immune landscape and assessing new therapies, particularly immunotherapies. By enabling effective evaluation of novel treatments, our model has the potential to significantly advance GBM research. Key PointsO_LIA wide array of human immune cells infiltrates GBM-PDX tumors in humanized mice. C_LIO_LIThe tumor heterogeneity of GBM-PDX is more evident in humanized mouse xenografts. C_LI Importance of the studyCurrent GBM treatment development is primarily constrained by the absence of appropriate animal models that can enhance our understanding of human GBM biology and its immune dynamics. We created a specialized GBM mouse model using humanized mice with patient-derived xenografts resistant to radiation. Detailed immune cell analysis showed various human blood cells in the GBM-PDX derived from humanized mice. Single-cell RNA sequencing uncovered crucial T-cell subgroups, notably regulatory and exhausted T cells, which matched those observed in recurrent GBM patient tumors. Moreover, the model outperformed standard xenograft approaches in tumor heterogeneity, resembling the intricate diversity observed in recurrent GBM samples. Our humanized GBM mouse model provides a powerful platform to study the interactions between human tumors and immune cells. This tool will facilitate the discovery of immunosuppressive mechanisms and accelerate the assessment of immunotherapies applicable to various brain tumors.

immunology↗

bioRxiv2023

Using single-cell RNA sequencing to generate cell-type-specific split-GAL4 reagents throughout development

Cell-type-specific tools facilitate the identification and functional characterization of distinct cell types, which underly the complexity of neuronal circuits. A large collection of existing genetic tools in Drosophila relies on enhancer activity to label different subsets of cells. These enhancer-based GAL4 lines often fail to show a predicable expression pattern to reflect the expression of nearby gene(s), partly due to an incomplete capture of the full gene regulatory elements. While genetic intersectional technique such as the split-GAL4 system further improve cell-type-specificity, it requires significant time and resource to generate and screen through combinations of enhancer expression patterns. In addition, since existing enhancer-based split-GAL4 lines that show cell-type-specific labeling in adult are not necessarily active nor specific in early development, there is a relative lack of tools for the study of neural development. Here, we use an existing single-cell RNA sequencing (scRNAseq) dataset to select gene pairs and provide an efficient pipeline to generate cell-type-specific split-GAL4 lines based on the native genetic regulatory elements. These gene-specific split-GAL4 lines can be generated from a large collection of coding intronic MiMIC/CRIMIC lines either by embryo injection or in vivo cassette swapping crosses and/or CRISPR knock-in at the N or C terminal of the gene. We use the developing Drosophila visual system as a model to demonstrate the high prediction power of scRNAseq-guided gene specific split-GAL4 lines in targeting known cell types. The toolkit allows efficient cluster annotation in scRNAseq datasets but also the identification of novel cell types. Lastly, the gene-specific split-GAL4 lines are broadly applicable to Drosophila tissues. Our work opens new avenues for generating cell-type-specific tools for the targeted manipulation of distinct cell types throughout development and represents a valuable resource to the fly research community. Significance StatementUnderstanding the functional role of individual cell types in the nervous systems has remained a major challenge for neuroscience researchers, partly due to incomplete identification and characterization of underlying cell types. To study the development of individual cell types and their functional roles in health and disease, experimental access to a specific cell type is often a prerequisite. Here, we establish an experimental pipeline to generate gene-specific split-GAL4 guided by single-cell RNA sequencing datasets. These lines show high accuracy for labeling targeted cell types from early developmental stages to adulthood and can be applied to any tissues in Drosophila. The collection of gene-speicifc-split-GAL4 will provide a valuable resource to the entire fly research community.

developmental biology↗

bioRxiv2019

scPADGRN: A preconditioned ADMM approach for reconstructing dynamic gene regulatory network using single-cell RNA sequencing data

Disease development and cell differentiation both involve dynamic changes; therefore, the reconstruction of dynamic gene regulatory networks (DGRNs) is an important but difficult problem in systems biology. With recent technical advances in single-cell RNA sequencing (scRNA-seq), large volumes of scRNA-seq data are being obtained for various processes. However, most current methods of inferring DGRNs from bulk samples may not be suitable for scRNA-seq data. In this work, we present scPADGRN, a novel DGRN inference method using time-series scRNA-seq data. scPADGRN combines the preconditioned alternating direction method of multipliers with cell clustering for DGRN reconstruction. It exhibits advantages in accuracy, robustness and fast convergence. Moreover, a quantitative index called Differentiation Genes Interaction Enrichment (DGIE) is presented to quantify the interaction enrichment of genes related to differentiation. From the DGIE scores of relevant subnetworks, we infer that the functions of embryonic stem (ES) cells are most active initially and may gradually fade over time. The communication strength of known contributing genes that facilitate cell differentiation increases from ES cells to terminally differentiated cells. We also identify several genes responsible for the changes in the DGIE scores occurring during cell differentiation based on three real single-cell datasets. Our results demonstrate that single-cell analyses based on network inference coupled with quantitative computations can reveal key transcriptional regulators involved in cell differentiation and disease development.\n\nAuthor summarySingle-cell RNA sequencing (scRNA-seq) data are gaining popularity for providing access to cell-level measurements. Currently, time-series scRNA-seq data allow researchers to study dynamic changes during biological processes. This work proposes a novel method, scPADGRN, for application to time-series scRNA-seq data to construct dynamic gene regulatory networks, which are informative for investigating dynamic changes during disease development and cell differentiation. The proposed method shows satisfactory performance on both simulated data and three real datasets concerning cell differentiation. To quantify network dynamics, we present a quantitative index, DGIE, to measure the degree of activity of a certain set of genes in a regulatory network. Quantitative computations based on dynamic networks identify key regulators in cell differentiation and reveal the activity states of the identified regulators. Specifically, Bhlhe40, Msx2, Foxa2 and Dnmt3l might be important regulatory genes involved in differentiation from mouse ES cells to primitive endoderm (PrE) cells. For differentiation from mouse embryonic fibroblast cells to myocytes, Scx, Fos and Tcf12 are suggested to be key regulators. Sox5, Meis2, Hoxb3, Tcf7l1 and Plagl1 critically contribute during differentiation from human ES cells to definitive endoderm cells. These results may guide further theoretical and experimental efforts to understand cell differentiation processes and explore cell heterogeneity.

bioinformatics↗

bioRxiv2018

Leveraging single cell RNA sequencing experiments to model intra-tumor heterogeneity

PURPOSEMany cancers can be treated with targeted therapy. Almost inevitably, tumors develop resistance to targeted therapy, either from preexistence or by evolving new genotypes and traits. Intra-tumor heterogeneity serves as a reservoir for resistance, which often occurs due to selection of minor cellular sub-clones. On the level of gene expression, the clonal heterogeneity can only be revealed by high-dimensional single cell methods. We propose to use a general diversity index (GDI) to quantify heterogeneity on multiple scales and relate it to disease evolution.\n\nMETHODSWe focused on individual patient samples probed with single cell RNA sequencing to describe heterogeneity. We developed a pipeline to analyze single cell data, via sample normalization, clustering and mathematical interpretation using a generalized diversity measure, and exemplify the utility of this platform using single cell data.\n\nRESULTSWe focused on three sources of RNA sequencing data: two healthy bone marrow (BM) samples, two acute myeloid leukemia (AML) patients, each sampled before and after BM transplant (BMT), four samples of pre-sorted lineages, and six lung carcinoma patients with multi-region sampling. While healthy/normal samples scored low in diversity overall, GDI further quantified in which respect these samples differed. While a widely used Shannon diversity index sometimes reveals less differences, GDI exhibits differences in the number of potential key drivers or clonal richness. Comparing pre and post BMT AML samples did not reveal differences in heterogeneity, although they can be very different biologically.\n\nCONCLUSIONGDI can quantify cellular heterogeneity changes across a wide spectrum, even when standard measures, such as the Shannon index, do not. Our approach offers wide applications to quantify heterogeneity across samples and conditions.

cancer biology↗

bioRxiv2018

Random forest based similarity learning for single cell RNA sequencing data

MotivationGenome-wide transcriptome sequencing applied to single cells (scRNA-seq) is rapidly becoming an assay of choice across many fields of biological and biomedical research. Scientific objectives often revolve around discovery or characterization of types or sub-types of cells, and therefore obtaining accurate cell-cell similarities from scRNA-seq data is critical step in many studies. While rapid advances are being made in the development of tools for scRNA-seq data analysis, few approaches exist that explicitly address this task. Furthermore, abundance and type of noise present in scRNA-seq datasets suggest that application of generic methods, or of methods developed for bulk RNA-seq data, is likely suboptimal.\n\nResultsHere we present RAFSIL, a random forest based approach to learn cell-cell similarities from scRNA-seq data. RAFSIL implements a two-step procedure, where feature construction geared towards scRNA-seq data is followed by similarity learning. It is designed to be adaptable and expandable, and RAFSIL similarities can be used for typical exploratory data analysis tasks like dimension reduction, visualization, and clustering. We show that our approach compares favorably with current methods across a diverse collection of datasets, and that it can be used to detect and highlight unwanted technical variation in scRNA-seq datasets in situations where other methods fail. Overall, RAFSIL implements a flexible approach yielding a useful tool that improves the analysis of scRNA-seq data.\n\nAvailability and ImplementationThe RAFSIL R package is available online at www.kostkalab.net/software.html

bioinformatics↗

bioRxiv2022

Single-Cell RNA Sequencing Reveals the Effects of Chemotherapy on Human Pancreatic Adenocarcinoma and its Tumor Microenvironment

The tumor microenvironment (TME) in pancreatic ductal adenocarcinoma (PDAC) is a complex ecosystem that drives tumor progression; however, in-depth single cell characterization of the PDAC TME and its role in response to therapy is lacking. We performed single-cell RNA sequencing on freshly collected human PDAC samples either before or after chemotherapy. Overall, we found a heterogeneous mixture of basal and classical cancer cell subtypes, along with distinct cancer-associated fibroblast and macrophage subpopulations. Strikingly, classical and basal-like cancer cells exhibited similar transcriptional responses to chemotherapy, and did not demonstrate a shift towards a basal-like transcriptional program among treated samples. We observed decreased ligand-receptor interactions in treated samples, particularly TIGIT on CD8+ T cells and its receptor on cancer cells, and identified TIGIT as the major inhibitory checkpoint molecule of CD8+ T cells. Our results suggest that chemotherapy profoundly impacts the PDAC TME and may promote resistance to immunotherapy.

cancer biology↗

bioRxiv2020

Comprehensive benchmarking of single cell RNA sequencing technologies for characterizing cellular perturbation

Technological advances in transcriptome sequencing of single cells continues to provide an unprecedented view on tissue composition and cellular heterogeneity. While several studies have compared different single cell RNA-seq methods with respect to data quality and their ability to distinguish cell subpopulations, none of these studies investigated the heterogeneity of the cellular transcriptional response upon a chemical perturbation. In this study, we evaluated the transcriptional response of NGP neuroblastoma cells upon nutlin-3 treatment using the C1, ddSeq and Chromium single cell systems. These devices and library preparation methods are representative for the wide variety of platforms, ranging from microfluid chips to droplet-based systems and from full transcript sequencing to 3-prime end sequencing. In parallel, we used bulk RNA-seq for molecular characterization of the transcriptional response. Two complementary metrics to evaluate performance were applied: the first is the number and identity of differentially expressed genes as defined in consensus by two statistical models, and the second is the enrichment analysis of biological signals. Where relevant, to make the data more comparable, we downsampled sequencing library size, selected cell subpopulations based on specific RNA abundance features, or created pseudobulk samples. While the C1 detects the highest number of genes per cell and better resembles bulk RNA-seq, the Chromium identifies most differentially expressed genes, albeit still substantially fewer than bulk RNA-seq. Gene set enrichment analyses reveals that detection of a limited set of the most abundant genes in single cell RNA-seq experiments is sufficient for molecular phenotyping. Finally, single cell RNA-seq reveals a heterogeneous response of NGP neuroblastoma cells upon nutlin-3 treatment, revealing putative late-responder or resistant cells, both undetected in bulk RNA-seq experiments.

molecular biology↗