Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Apr 7.
Published in final edited form as: Nat Cardiovasc Res. 2023 Jan 18;2(1):20–34. doi: 10.1038/s44161-022-00205-7

Current and Future Perspectives of Single-Cell Multi-omics Technologies in Cardiovascular Research

Wilson Lek Wen Tan a, Wei Qiang Seow a, Angela Zhang a,b, Siyeon Rhee a,b, Wing H Wong c,d, William J Greenleaf e, Joseph C Wu a,b,f,*
PMCID: PMC11974510  NIHMSID: NIHMS1991624  PMID: 39196210

Abstract

Single-cell technology has become an indispensable tool in cardiovascular research since its first introduction in 2009. Here, we highlight the recent remarkable progress in using single-cell technology to study transcriptomic and epigenetic heterogeneity in cardiac disease and development. We then introduce the key concepts in single-cell multi-omics modalities that apply to cardiovascular research. Lastly, we discuss some of the trending concepts in single-cell technology that are expected to propel cardiovascular research to the next phase of single-cell research.

Introduction

Single-cell technology has propelled novel discoveries in all niches of biomedical research1. Since first being described in 2009, single-cell RNA-sequencing (scRNA-seq) has evolved rapidly and has been adopted by most research laboratories to uncover cell-type-specific gene regulation2. Today, it is possible to profile the transcriptome of tens of thousands of cells in a single experiment3. The conventional assay of bulk RNA-seq suffers from the averaging effect, which masks the differential contribution from heterogeneous cell populations4. While scRNA-seq has been widely adopted in cardiovascular research, novel single-cell methodologies emerge to study different aspects of cell biology, such as DNA modification, chromatin accessibility, transcription factor (TF) binding, and proteomics5–8. By parallel profiling of genomic, epigenetic, and transcriptomic readout at single-cell resolution, researchers can now study the complex relationship between epigenetic modifications and gene expression at unprecedented resolution (Figure 1)9. This review article will provide an overview of the application of single-cell technologies in the human heart and vascular tissues. Specifically, we offer an in-depth discussion and appraisal of single-cell technology’s impact on the cardiovascular field. Next, we discuss various multi-omics methodologies that are relevant to cardiovascular research. Lastly, we discuss possible innovations of single-cell technologies that can revolutionize cardiovascular research.

Figure 1. Illustration of the single-cell sequencing technologies and the measured modalities.

Figure 1

Techniques have been developed to measure single or multiple modalities simultaneously. For example, some popular techniques to measure transcriptome and epigenome in parallel use an experimental protocol such as 10X GEX + ATAC Multiome, Pair-seq, sci-CAR, SNARE-seq, and ASTAR-seq. Novel experimental protocols also allow investigation of the correlation between proteome and transcriptome simultaneously, such as CITE-seq, REAP-sq, and ECCITE-seq. There are also techniques to investigate the roles of genome-wide perturbation at the single-cell level, such as Perturb-seq, CROP-seq, and Spear-ATAC.

I. Single-cell Methodologies

The transcriptome is dynamic and tightly linked to cell states and functions10. In efforts to deconvolve bulk heart tissue at single-cell resolution, scRNA-seq has been widely adopted in cardiovascular research11 (Table 1). Single-cell transcriptomic sequencing performed on large cohorts of human hearts has unveiled the main cardiac cell types with diverse molecular signatures and functions12–19. Beyond the realm of transcriptomic landscape, cardiovascular research begins to witness the application of single-cell technology for epigenetics analysis, such as single-cell ATAC-seq (scATAC-seq), to uncover the chromatin biology of heart diseases7,18,20. Several single-cell methods have been developed, each with a unique approach to cell capture and amplification, as well as differences in mRNA transcript length, target cell capture number, and read depth per cell. Each method has distinct merits and limitations. Nevertheless, most single-cell techniques developed to study transcriptome and epigenome follow a similar workflow. Here, we will describe and evaluate the common single-cell methodologies applied in cardiovascular research.

Table 1. Single-cell sequencing in cardiovascular development and disease.

sci-RNA-seq3, single-cell combinatorial indexing RNA sequencing; sci-ATAC-seq3, single-cell combinatorial ATAC sequencing; CEL-seq2, Cell Expression by Linear amplification and sequencing; DCM, Dilated cardiomyopathy; ACM, Arrhythmogenic cardiomyopathy; HCM, Hypertrophic cardiomyopathy; iPSC-EC, induced pluripotent stem cell-derived endothelial cell; GWAS, genome wide association study.

Assay Tissue Cell Types Number of Donors Method Age Number of Cells Findings Year Data Availability Ref
scRNA-seq / snRNA-seq Whole heart All cell types 79 10x Genomics adult 881,081 snRNAseq of human ventricular samples illuminated cell types and states, molecular signals, and intercellular communications that characterize DCM and ACM. 2022 EGAS00001006374 19
Whole heart All cell types 42 10x Genomics adult 592,689 Comprehensive characterizion of the transcriptional landscape at the cellular level from the human heart in health and in patients with DCM or HCM. 2022 SCP1303 13
Whole heart All cell types 23 10x Genomics adult 191,795 An integrative molecular map of human myocardial infarction, represents an essential reference for the field and paves the way for advanced mechanistic and therapeutic studies of cardiac disease. 2022 Cellxgene 18
Septum All cell types 5 10x Genomics adult 88,536 Human cell atlas of the hypertrophied heart highlights the importance of intercellular crosstalk in disease pathogenesis 2022 E-MTAB-11268 16
Left ventricle All cell types 45 10x Genomics adult 220,752 Unsupervised clustering, differential gene expression and trajectory analyses revealed cell type-specific transcriptional programs and emergence of disease-associated cell states in the context of heart failure. 2022 GSE183852 15
Multiple tissues (including fetal hearts) All cell types 121 sci-RNA-seq3 fetus 4,979,593 Organ-specific and cell-type specific gene regulations (including heart) 2020 GSE156793 110
Whole heart All cell types 14 10x Genomics adult 363,213 nuclei 45,870 cells 78,023 CD45+ enriched cells Cellular heterogeneity of cardiomyocytes, pericytes and fibroblasts, and revealed distinct atrial and ventricular subsets of cells with diverse developmental origins and specialized properties. 2020 ERP123138 12
Whole heart All cell types 7 10X Genomics adult 287,269 Transcriptional and cellular diversity in non-failing human heart. 2020 phs001539.v1.p1 14
Whole heart All cell types 24 ICELL8 adult 21,422 Identification 9 major cardiac cell types and marked differences in cell type transcription by chamber, laterality and sex. 2020 GSE109816 / GSE121893 111
Vasti lateralis muscle Satellite cells 8 10x Genomics adult 68,108 Identification of novel subpopulations of human satellite cells and their associated surface markers 2020 Q65X273X 112
hiPSC-derived Endothelial Cells hiPSC-derived Endothelial Cells 1 10x Genomics adult 5,673 Bona fide iPSC-EC cluster during the differentiation process, which exclusively expressed endothelial-specific genes 2020 GSE116555 113
Ascending aortic wall tissue All cell types 3 10x Genomics adult 48,128 Transcription factor ERG a potential therapeutic target for ascending thoracic aortic aneurysms 2020 GSE155468 114
Human Atherosclerotic Carotid Artery All cell types 3 10x Genomics adult 8,867 Smooth muscle cells transitioned to an intermediate cell state during atherosclerosis. 2020 GSE155512 115
iPSC-derived Cardiomyocytes hiPSC-derived Cardiomyocytes 2 Drop-seq / DroNc-seq Multiple Time Points 42,704 Benchmarking experiments to establish the applicability of single-nucleus RNA-seq for the Human Cell Atlas (HCA) 2020 GSE129096 116
Carotid Endarterectomy Tissues All cell types 18 CEL-seq2 adult 3,282 A transcriptome-based cellular landscape of human atherosclerotic plaques that highlights cellular plasticity and intercellular communication at the site of disease. 2020 Upon Request 117
Right coronary artery All cell types 4 10x Genomics adult 11,756 Protective role for both TCF21 and SMC phenotypic modulation in coronary artery diseases 2019 GSE131778 58
Whole heart All cell types 18 Mouth pipette fetus 4,948 Mechanims of in vivo human cardiac development and potential clues to understand cardiac regeneration 2019 GSE106118 118
Left Ventricle Cardiomyocytes 10 Smart-seq2 adult 411 Conservation of pathogenic gene programs in mouse and human cardiomyocytes provides the potential for assessing and regulating cardiomyocyte remodeling in heart failure. 2018 GSE95143 119
hiPSC-derived cardiomyocytes hiPSC-derived Cardiomyocytes 2 10x Genomics Multiple Time Points 10,376 In vitro cellular heterogeneity of cardiomyocytes 2018 GSE81585 120
Left ventricles Cardiomyocytes 5 C1 Fluidigm adult 129 Long non-coding RNA regulating dedifferentiation and cell cycle genes in cardiomyocytes 2017 SRP049944 25
scATAC-seq Coronary arteries All cell types 41 10x Genomics adult 28,316 Landscape of ~320,000 accessible sites across all cells in coronary artery, identified cell-type-specific elements and transcription factors, and prioritized functional CAD risk variants. 2022 GSE175621 / GSE188422 6
Atherosclerotic lesions All cell types 3 10x Genomics adult 7,000 Single-cell epigenomics and functional fine-mapping of atherosclerosis GWAS 2021 phs000424.v8.p2 7
Multiple tissues (including fetal hearts) All cell types 59 sci-ATAC-seq3 fetus 1,568,018 Organ-specific and cell-type specific gene regulations (including heart) 2020 GSE149683 8
scRNA-seq + scATAC-seq Whole heart All cell types 23 10x Genomics adult 191,795 An integrative molecular map of human myocardial infarction, represents an essential reference for the field and paves the way for advanced mechanistic and therapeutic studies of cardiac disease. 2022 Cellxgene 18
Whole Heart All cell types 4 10x Genomics adult 79,515 Cell-type specific epigenetics regulation in human heart hearts 2021 GSE165839 20
Spatial Transcriptomics Whole Heart All cell types 4 Codelink-activated microscope glass slides Embryonic NA Spatial annotation of embryonic cardiac gene expression to facilitate future studies on human cardiogenesis 2019 EGAS00001003996 8

Microfluidic-based Technique

One of the more commonly used scRNA-seq methodologies depends on the droplet-based microfluidic technology21. For instance, the C1 Fluidigm system isolates single cells into individual reaction chambers in the exclusive integrated fluidic circuit (IFC) (Figure 2a)22. The optically clear IFC allows users to automatically stain captured cells and examine them by microscopy for viability, surface markers, or reporter genes. Subsequently, the cells are automatically lysed and prepared for sequencing. This method combines the accurate manipulation of flow rates of individual cells or nuclei and chemical particles with a cell partitioning system, allowing single-cell capture in aqueous microdroplets (Figure 2b). For single-cell epigenetic studies such as scATAC-seq, transposition using a transposome complex containing transposase (Tn5) is performed on each cell in the IFC, where the transposase preferentially inserts the sequencing adapters (tagmentation) into nucleosome-free regions, followed by library construction23. Another popular microfluidic-based scRNA-seq/scATAC-seq workflow in the cardiovascular field is the Chromium system provided by 10X Genomics, which can process tens of thousands of cells or nuclei within a single day workflow at a lower cost of ~$0.04 per cell12.

Figure 2. Single-cell isolation and library construction methods.

Figure 2

a. Tens of thousands of tiny channels, reaction chambers, and control valves integrate into an Integrated Fluidic Circuit (IFC) chip. b. Gel beads loaded with primers and barcoded oligonucleotides are first mixed with cells and reagents and subsequently mixed with the oil-surfactant solution at a microfluidic junction. Next, single-cell GEMs are collected in the GEM outlet. RT takes place inside each GEM, after which cDNAs are pooled for amplification and library construction in bulk. c. The fixed and punched cells distribute randomly into each well in 96-well plates in each split-and-pool process. Finally, three rounds of barcoding are performed in 96-well plates, which can ultimately label 1 million cells with a unique barcode combination.

The microfluidic-based technique relies on the quality of the single-cell suspension24. Generation of the high-quality single-cell suspension remains challenging for heart tissue due to its resistance and vulnerability to dissociation of the tissue collagen matrix15. Therefore, many heart-related datasets used single-nuclei RNA-seq (snRNA-seq), which captured the nuclei of cell types that are difficult to recover from enzymatically digested tissue, such as cardiomyocytes, epicardium, and lymphatics12,25. scRNA-seq data, on the other hand, might provide additional insights into cell types that were biased against in nuclei isolation26. Koenig et al. observed a disparity in the representation of immune cells in both scRNA-seq and snRNA-seq datasets performed on the same hearts15. Monocytes and intermediate macrophage populations were preferentially captured in scRNA-seq, whereas resident macrophages were preferentially captured using snRNA-seq15. Therefore, supplementing snRNA-seq with scRNA-seq datasets will provide additional granularity into the landscape of inflammatory responses in cardiac pathology. While snRNA-seq presents comparable gene detection sensitivity and cell-type identification to scRNA-seq, it remains a concern that snRNA-seq might be biased against genes that have an uneven distribution between the nucleus and cytoplasm27.

Combinatorial Cellular Indexing

Another popular single-cell transcriptome method known as split-pool barcoding uses combinatorial cellular indexing to measure gene expression from thousands of single cells in a single experiment (Figure 2c)3,28. Cao et al. developed single-cell combinatorial indexing RNA sequencing (sci-RNA-seq)29. Cells are distributed evenly into each well in a 96-/384-well plate. Cellular barcode and well-specific poly-T primer containing unique molecular identifiers (UMI) are introduced to the transcript of the cells within each well via in situ reverse transcriptions. Next, the cells from all wells are pooled and redistributed to each well by fluorescence-activated cell sorting (FACS). For single-cell combinatorial ATAC-seq (sci-ATAC-seq), additional steps are required to produce pre-indexed (barcoded) nuclei using the Tn5 transposome complexes for in situ tagmentation before pooling and redistribution30. Improved versions of this method, sci-RNA-seq3 and SPLiT-seq, replaced FACS with dilution and filtration to prevent cell aggregation3,28. Increasing the number of barcodes with multiple rounds of indexing and limiting the number of cells distributed in each round of sorting can increase the complexity of the experiment. sci-ATAC-seq requires custom-loading of Tn5 with barcoded sequences, which are limited to 100,000 cells per experiment. An improved version, sci-ATAC-seq3, overcomes the limitation by performing the first two rounds of indexing by ligation to either end of the conventional, uniformly loaded Tn5 transposase complex, followed by a third round of indexing through PCR5.

Despite its ability to profile an exponentially higher number of cells than the microfluidic counterpart, the split-pool barcoding method suffers from reduced sensitivity in gene detection. Split-pool scRNA-seq can identify only approximately a quarter of genes typically detected with 10x Chromium31. While cell type identification is possible with just hundreds of highly expressed genes, the lack of granularity and depth in gene expression quantification might hamper the discovery of transcriptomics changes in diseases, involving genes at the lower spectrum of expression3. However, a recently developed method, dsciATAC-seq, combines droplet-based and combinatorial indexing approaches, has substantially improved the throughput, quality, and coverage of single-cell profiles32. Combinatorial barcoding experiments are also laborious, requiring multiple rounds of pipetting and sorting steps to ensure each cell receives unique combination of barcodes3,28,29. Lastly, cautions must be taken in experimental design, especially in estimating the number of final recovered cells28. The final recovered cells must be far lower than the total number of barcode combinations to avoid collisions, as explained by the Birthday Problem’s mathematics concept33. Nonetheless, a commercial solution offered by Parse Bioscience promises higher gene detection sensitivity with simplified protocol in the future with further protocol optimization34.

II. Application of scRNA-seq in Cardiovascular Research

Cell-type Clustering and Identification

The ability to define cell types using unsupervised learning emerges as a fundamental application of scRNA-seq (Figure 3). Unsupervised learning uses machine learning (ML) algorithms to uncover hidden patterns and classify unlabeled datasets without human intervention35. For example, Litvinukova et al. generated a massive cellular landscape of 14 adult human non-failing (NF) hearts, comprising 123,893 cells and 363,213 nuclei. By using an integrative approach of identifying differentially expressed gene markers and pathway analysis, the landscape has annotated major cell types, such as cardiomyocytes, fibroblasts, endothelial cells (EC), pericytes, smooth muscle cells (SMC), mesothelial cells (MC), neuronal cells (NC), and immune cells12. Using a similar cell type annotation approach, an independent study performed on seven donor adult hearts validated the presence of the same major cardiac cell types across 287,269 nuclear transcriptome profiles14. These single-cell landscapes of the human NF hearts serve as a valuable reference for disease comparisons15–17.

Figure 3. Summary of main cardiac cell types identified in human hearts.

Figure 3

The figure illustrates the major cell types and their respective markers in healthy and disease settings: cardiomyocytes, fibroblasts, endothelial cells, pericytes, smooth muscle cells, immune cells (myeloid and lymphoid), adipocytes, mesothelial cells and neuronal cells.

One of the key advantages of scRNA-seq is the ability to discern transcriptional heterogeneity within a population of cells, which is impossible with bulk tissue RNA-seq. Consistent with bulk tissue RNA-seq, scRNA-seq of human hearts showed high expression of genes that encode contractile proteins and calcium-mediated pathways such as TTN, RYR2 and SLC8A1. In addition, Litvinukova et al. further refined cardiomyocytes into 10 clusters of ventricular and atrial cardiomyocyte subtypes enriching in diverse cellular pathways and gene markers12. Tucker et al. also reported 5 subclusters of cardiomyocytes, consisting of both atrial and ventricle cardiomyocytes14. Despite the differences in the number of subclusters, both single-cell studies identified a subcluster of cardiomyocytes enriching for the expression of stress-response gene ANKRD1, which involves in cardiomyopathy-associated remodeling36. Besides, all single-cell studies of human hearts have detected rare populations of non-myocytes previously undescribed in bulk RNA-seq, including neuronal-like cells, adipocytes, mesothelial cells, and various immune cells12–14,16,17,19. However, the single-cell cluster annotations reported by most studies are often the ”best guesses,” often without orthogonal experimental validations such as RNAScope and SeqFISH37. Nonetheless, with the advent of spatial transcriptomics capabilities, cardiovascular research will likely integrate spatial gene activity and transcript regulatory networks to validate the presence and functional interactions of different cell types in space to coordinate tissue functions8,18.

Cell-type Specific Disease Gene Expression Perturbation

Understanding the precise cellular and molecular mechanisms underpinning heart failure will be the key to developing novel therapeutics. Nicin et al. generated snRNA-seq profiles from five patients with cardiac hypertrophy (HCM). They showed profound dysregulation of the gene expressed in cardiomyocytes compared with the healthy hearts16. Koenig et al. obtained 220,752 nuclei and 49,723 cells from the hearts of a large cohort of 28 non-diseased donors and 17 donors with non-ischemic dilated cardiomyopathy (DCM)15. Besides observing the upregulation of the cardiomyocyte stress markers, Koenig et al also reported DCM samples displaying a bias toward ADGRL3-expressing cardiomyocytes and a trend toward more NPPA/NPPB-expressing cardiomyocytes. Globally, a marked reduction in cardiomyocyte proportion in DCM observed by Koenig et al. was corroborated by a subsequent scRNA-seq of human DCM and HCM studies13.

The single-cell analysis also found notable transcriptional changes in the non-cardiomyocyte population in DCM hearts. Chaffin et al. showed a statistically credible increase of activated fibroblasts in both HCM and DCM hearts. Furthermore, the transcriptional differences observed in each cardiomyopathy group were contributed mainly by fibroblasts13. A population of activated fibroblast, regarded as exclusive in DCM patients, showed increased expression of known activated fibroblast markers such as POSTN, NOX4, and COL1A1, as well as genes previously not known to be associated with fibrosis such as FAM155A and TSHZ2. Also reported by Koenig et al, another activated fibroblast marker, THBS4, is up-regulated in matrifibrocytes and involved in myocardial remodeling, interstitial fibrosis and cardiomyopathy. In addition, the TF MEOX1, implicated in murine fibroblast activation, was dysregulated in DCM hearts38.

Intercellular Crosstalk Analysis

Cells interact with each other via ligand-receptor interactions to coordinate functions in a multicellular tissue. The expansion of protein-protein interaction data warehouses enables the prediction of intercellular signaling from single-cell transcriptome39. A series of bioinformatics tools have been widely adopted in single-cell analysis to delineate cell-cell communication, including CellChat40, CellPhoneDb41, and SingleCellSignalR42.

Using CellPhoneDB, Litvinukova et al. have elucidated cellular interactions among diverse cell types in healthy hearts and validated the interactions using orthogonal assays such as spatial transcriptomics12. For instance, ECs interacted with SMCs via Notch receptor-ligand interactions. Using Cellchat, Reichart et al. reported aberrant genotype-specific and chamber-specific intercellular signaling across disease in heart diseases, including dysregulation of fibrosis-related and vascular remodeling pathways19. In human HCM, Nicin et al. reported a significant reduction in cell-cell interactions in hypertrophied cardiomyocytes16. The analysis found significant repression of the gene encoding EpH receptor tyrosine kinase, EPHB1, and its ligand, ephrin (EFNB2; secreted by EC), in hypertrophied cardiomyocytes17. EFNB2 is indispensable in mouse embryonic ECs and endocardial cells for proper cardiovascular development43. However, the role of EPHB1 in cardiomyocytes is unknown. In vivo and in vitro experiments confirmed the cardioprotective role of ephrin EFNB2 in inhibiting the progression of cardiomyocyte hypertrophy16. The discovery of novel interaction between EPHB1 and EFNB2 showed that single-cell transcriptomic analysis could complement various orthogonal methods and imaging modalities to elucidate mechanisms of cellular crosstalk in cardiovascular diseases.

Trajectory Inference Analysis

Single-cell data provide unique opportunities for studying cellular dynamic processes such as cell cycle, differentiation, and disease activation. Such dynamic processes can be modeled computationally using trajectory inference methods which order cells along a trajectory or pseudotime based on congruence in their gene expression patterns. Currently, more than 70 trajectory inference tools have already been developed44.

Using Palantir, Koenig et al. observed considerable variation in how cardiac cell populations responded to heart failure15. Cardiomyocytes converged toward standard disease-associated cell states marked by high expression of ARGRL3 and NPPA/NPPB expression. In contrast, fibroblasts and myeloid cells underwent dramatic diversification, including acquiring disease-specific phenotypes15. Chaffin et al. applied cell-state trajectory analysis to dissect the expression gradient between quiescent and activated fibroblasts using Slingshot and scVelo13. The trajectory analyses unveiled several genes, such as PRELP, JAZF1, and COL22A1, whose expression changed along the fibroblast activation pseudotime, suggesting potential roles in cardiac fibrosis13. Despite having compelling in vitro experimental evidence that knockout of these novel genes impeded fibroblast activation, it is unclear if the myofibroblast transition from TGFβ1 in cultured cardiac fibroblasts reflects inferred trajectory in vivo. Nonetheless, the cardiac fibroblast activation trajectory analyses have expanded the current understanding of these cells in fibrotic disease of the heart and further inform the potential therapeutic targets for heart failure.

Challenges and Future Direction in Single-cell Transcriptomics

- Quality Control of Single-cell Dataset

Proper quality control (QC) steps are needed to ensure the robustness of the single-cell transcriptomics datasets. Despite advances in single-cell protocols, cells of poor quality can still be present in high-quality library preparation and should be omitted before downstream analysis. Therefore, several standard QC metrics are applicable in the data processing of single-cell datasets (Figure 4).

Figure 4. Illustration of computational analysis commonly applied to single-cell.

Figure 4

The scRNA-seq/snRNA-seq experimental design will determine how to parse the barcodes and UMIs from the sequencing reads. The preprocessing steps of the workflow include: (1) Generation of the gene expression count matrix; (2) Quality control of the raw gene expression count to exclude poor quality cells; (3) Normalization, scaling and feature selection; and (4) Unsupervised clustering of cells based on transcriptional similarities. Many of these steps have been implemented in open-source packages such as Seurat and Scanpy with detailed tutorials62,109.

Doublets or multiplets are also commonly encountered in the droplet-based method when two or more cells are partitioned into a single droplet during the cell sorting process, resulting in an artificial hybrid of expression profile. Several algorithms can detect potential doublets by combining expression profiles of randomly sorted cells and then scoring each cell against the in silico doublets45–47.

Another major challenge in scRNA-seq is the presence of ambient RNA in the cell suspension, which will be indistinguishable from a cell’s native RNA48. The problem is even more prevalent in snRNA-seq due to the lysis of cells, resulting in a substantial amount of cytoplasmic RNA contamination in ambient RNA. Algorithms such as decontX and SoupX can be applied to estimate contamination levels based on the counts derived from native RNA and contaminating ambient RNA48,49.

Perturbations during sample preparation can also lead to biological artifacts. For instance, damaged cells during the tissue dissociation step may express abnormally large proportions of mitochondrial genes50. These cells may appear as a unique cluster in the scRNA-seq data even though they did not represent a unique cell population in the original tissue sample. While using mitochondrial read percentage can serve as a plausible criterion to exclude damaged cells, caution is needed when selecting the threshold for metabolically active tissue such as the heart to avoid erroneous biological interpretations of scRNA-seq data51.

- Transcriptome Diversity

Current landscapes of single-cell studies in the cardiovascular field mainly explored transcriptome variation at the gene level. The genomes of all eukaryotes contain introns52. In humans, the average gene has about eight introns. Correct intron excision by the splicing machinery is a central, conserved step during gene expression in all eukaryotes. RNA-binding proteins and splicing factors regulate alternative splicing of the pre-mRNA. RBM20 is a splicing factor expressed highly in both atria and ventricles53. RBM20 regulates cardiac splicing, including the splicing of TTN54. Thus, the downstream consequences of RBM20 mutations may share similarities to those of TTN-truncating variants. Ensuring full-length coverage across mRNA is essential to analyze alternative splicing forms and identify single-nucleotide polymorphisms, regardless of the orientation of amplification55. Recently, long-read sequencing platforms such as Nanopore and Pacific Bioscience (PacBio) proposed solutions to integrate long-read sequencing with single-cell technology to capture the full diversity of RNA56. Future single-cell studies in the cardiovascular field should adopt long-read sequencing to uncover the full spectrum of RNA diversity, including poly-A-negative long non-coding RNA (lncRNA) and circular RNA in a cell-type-specific fashion.

III. Application of scATAC-seq in Cardiovascular Research

Cell Type-specific Transcription Factor (TF) Inference

One of the major applications of scATAC-seq analysis is the inference of TF activities responsible for cell type-specific gene regulation. Enhancers harbor the recognition sites for TFs for the regulation of gene expression. scATAC-seq has demonstrated the potential to uncover dynamic TF activities in enhancers, which are missed by scRNA-seq23.

The single-cell landscape of cardiac chromatin accessibility in healthy human hearts has uncovered >287,000 CREs, where most of the accessible regions are at least 2 kilobases away from the known gene promoters20. Cell-type-specific TF activities can be identified using TF motif analysis tools such as ChromVAR. For example, MEF2A/B and NKX2–5 motifs were significantly enriched in cardiomyocyte-specific enhancers, whereas TCF21 motif was enriched considerably in fibroblast-specific enhancers. These cell-type-specific TF motif enrichments were also observed in mouse heart scATAC-seq datasets, suggesting the existence of standard regulatory grammar across species in cardiac gene regulation57.

Similarly, in vascular tissue, TF motifs enrichment in cell-type-specific CREs also corroborated nicely with the gene expression in scRNA-seq, including EC-specific TF (SOX13), macrophage-specific TF (SPI1), and SMC-specific TF (TEAD3). In atherosclerosis, SMC elicits phenotypic switching by transdifferentiating into other cell types such as fibromyocytes, osteogenic cells, and immune-like cells. Motif analysis of the CREs suggested depletion of SMC-specific TFs such as SRF and NFY in fibromyocytes. In contrast, fibroblast-specific TFs such as AP-1, NFATC2 and STAT increased as the SMCs transdifferentiated into fibromyocytes7. A similar transdifferentiation trajectory from SMC to fibromyocytes was constructed using scRNA-seq in an independent study58. Turner et al. also performed chromVAR-based TF enrichment analysis and revealed highly enriched motifs for AP-1 family members and TCF21 in the fibromyocyte subpopulation6. Conversely, TF motifs such as MEF2A and CArG box motifs enriched in non-fibromyocyte SMC clusters. However, due to the lack of the chromatin accessibility dataset, the intricate orchestration of TF regulation along the transdifferentiation trajectory described above was not observed58.

Construction of Cell Type-specific cis-Regulatory Map

Another application of scATAC-seq is the inference of a cis-regulatory map that connects distal regulatory sequences such as enhancers to their target genes. One popular method, Cicero, uses sampling and aggregating groups of similar cells from scATAC-seq data to quantify the correlations between putative CREs59. Based on the correlations between CREs, Cicero predicts the regulatory links between CREs and their target genes using unsupervised ML.

In the human heart, distal CREs exhibit chamber-specific gene regulation20. For example, atrial cardiomyocytes-specific CREs linked with genes involved in atrial development (PITX2), and ion channel subunits (SCN5A), whereas ventricle-specific CREs linked to genes encoding sarcomeric functions such as MYL2. These genes are chamber-specific12,14. Disease-risk variants for complex diseases tend to enrich in distal regulatory elements60. For example, Hockers et al. showed cell-type-specific enrichment of disease-associated genetic variants using scATAC-seq. For instance, atrial fibrillation-associated genetic variants enrich in atrial and ventricle cardiomyocyte CREs20. CREs of SMCs and ECs in the arteries showed the highest enrichment of genetic variants for coronary artery diseases (CAD). Furthermore, most of the CREs harboring CAD genetic variants interacted with at least one gene, where almost half of the genes exhibited cell-type specificity.

Using scATAC-seq datasets, Turner et al. employed a multi-tiered approach to prioritize candidate functional CAD GWAS variants, and uncovered target cell types for CAD regulatory variants6. A handful of top candidate CAD variants were mapped to cell-type-specific regulatory regions. For example, a GWAS variant, rs9349379, located at the vicinity of EDN1 gene, was found to disrupt an MEF2 binding sites specific to SMC and macrophage. By using Peak2Gene function implemented in scATAC-seq analysis package, ArchR, a total of ~150,000 cis-regulatory links between co-regulated regions were proposed, many of which harbored CAD risk variants2. Turner et al. observed a strong correlation between the peaks that harbor dozens of tightly linked CAD variants at the BCAS3 gene loci, with the expression of a distal gene, TBX2. Similarly, another set of CAD GWAS variant was observed to correlate better with the expression of a distal gene, PRDM16, rather than a nearby gene, ACTRT2 gene. The prioritization of PRDM16 and TBX2 as target CAD genes by Peak2Gene analysis was supported by orthogonal assessment using Activity-By-Contact (ABC)61 enhancer-gene mapping and cis-eQTL analysis in arteries6.

IV. Single-Cell Multi-omics Profiling

We have entered the age of multi-omics profiling. An increasing number of techniques have been developed to capture more than one omics layer from the same cell, which allows us to paint a complete picture about each cell than using any single omics layer alone, making it possible to infer better the complex networks of interactions that are responsible for cellular phenome9. Computational methods have been developed to integrate multimodal single-cell omics experiments performed on separate samples of cells from the same tissue62. This section will review some popular single-cell multi-omics assays to adopt in cardiovascular research to delineate complex disease and development molecular mechanisms.

Transcriptome + Chromatin

Transcriptomic analyses of a single cell reveal its transcriptional state, whereas epigenome analyses uncover the associated chromatin regulatory landscape. While it is possible to profile these modalities separately, the computational integration of these separate assays might need to recapitulate the actual biological snapshot of the cell fully. Therefore, multi-omics joint profiling of the transcriptomics and epigenomics layers from the same cells allows the direct matching of transcriptional regulation to its output unambiguously, allowing for accurate reconstruction of the molecular trajectories underlying a cell’s phenome. Different techniques are available to perform joint gene expression profiling and chromatin states, including sciCAR, SNARE-seq, Paired-seq, SHARE-seq, and 10X Genomics GEX + ATAC Multiome. Like single omics technology, multi-omics profiling techniques are mainly microfluidic droplet-based technology or plate-based combinatorial indexing strategy. As an extension to sci-RNA-seq and sci-ATAC-seq, sciCAR uses a plate-based combinatorial indexing strategy to profile single-cell chromatin accessibility and gene expression jointly63. Similar to sciCAR, another method known as Paired-seq, also employs a combinatorial indexing strategy64. However, unlike sciCAR, Paired-seq includes tagmentation of the nuclei with barcoded Tn5 before cDNA generation via reverse transcription, which prevents the tagmentation of DNA/RNA heteroduplexes that are known to interfere with the second strand synthesis and thus avoids the unintended inclusion of cDNA molecules in the chromatin library. A more recent technique, SHARE-seq, also uses multiple rounds of barcoding to increase the scalability of cells profiled33. Moreover, SHARE-seq can sequence both ends of the ATAC-seq fragments, an advantage over Paired-seq, by extending the number of index cycles during the sequencing steps. Unlike combinatorial indexing methods, the 10x Genomics Multiome GEX + ATAC kit uses the standard 10x chromium controller to encapsulate cells in droplets62. Each droplet contains beads padded with a mixture of two oligos containing the same cell barcode but different capture sequences for mRNAs and chromatin. Usefully, this kit is commercially available and comparatively less laborious than the plate-based technique.

Transcriptome + Proteome

Alternative splicing increases protein diversity. In humans, up to 95% of genes with multiple exons undergo alternative splicing, expanding the repertoire of protein function diversity65. While transcriptomic readouts from scRNA-seq provide insights into the underlying mechanisms, proteins, but not transcripts, are usually the primary targets of drugs66. Furthermore, the protein and mRNA abundance might not be causally predictive66. In addition, proteins might undergo post-translational modifications, such as methylation, acetylation, ubiquitination, and SUMOylation, which are impossible to infer from transcription data alone. Therefore, an unbiased view of proteins is essential to model cellular dynamics and response to disease perturbations.

While it is possible to quantitatively record the levels of select cell surface proteins during the FACS-based isolation of single cells prior to scRNA-seq, this technique is low throughput in terms of the number of proteins that can be profiled due to spectral overlap. Proximity extension assays (PEA) measure the abundance of RNAs and proteins from the same cell using DNA polymerase67. PEA relies on binding two antibodies in proximity to generate a DNA reporter with low background noise, rendering both analytes stable, amplifiable and quantitatively detectable67. This technique, combined with microfluidic-based technologies such as the C1 system for single-cell library construction, allows the parallel measurement of 38 proteins and 96 transcripts.

Proximity ligation assay for RNA (PLAYR) uses mass cytometry to quantify the abundance of RNA and proteins by hybridizing transcripts with mass cytometry-compatible probes and simultaneously targeting specific proteins with antibodies conjugated with distinct metal isotopes68. While PLAYR allows the evaluation of RNA and proteins across thousands of single cells, it is limited by the availability of metal isotope tags and high-affinity antibodies.

Other techniques, including CITE-seq and REAP-seq, uses DNA barcode-conjugated antibodies to target hundreds of proteins in a single experiment69,70. After the incubation with antibody-oligo complexes, single cells are encapsulated into nanoliter-sized droplets in a microfluidic apparatus designed for 10x Genomics gel beads scRNA-seq library construction. However, these techniques are less suited for solid tissues that require cell dissociation, disrupting cell membranes’ integrity. Furthermore, the ubiquitous non-specific binding of DNA-conjugated antibodies results in inaccurate protein quantification71.

inCITE-seq jointly measures intranuclear proteins using DNA-conjugated antibodies and transcriptome on a droplet-based platform71. Unlike CITE-seq and REAP-seq, inCITE-seq allows the diffusion of antibodies across the lightly fixed nuclear membrane without cell dissociation. In addition, optimized blocking conditions in inCITE-seq minimize non-specific binding of DNA-conjugated antibodies in the nucleus71. Recently, Chen et al. developed NEAT-seq to quantify the abundance of nuclear proteins, chromatin accessibility, and transcriptome in a single cell72. Compared to inCITE-seq, NEAT-seq showed moderate improvements in the non-specific antibody binding by blocking the charges of the DNA oligo conjugated to the antibody72.

Spatial Transcriptomics

While single-cell technologies discover novel cell-type signatures, spatial relationships between cells are not captured. Techniques to study spatial transcriptomics have been invented to enable the reconstruction of spatially-resolved cellular transcriptomes73. For cardiovascular research, spatial transcriptomics technology is relevant to understanding how these different cell types interact in space to coordinate tissue functions, affecting developmental processes and pathophysiology8.

The imaging-based spatial transcriptomics method uses image processing to generate a gene-expression matrix. In situ sequencing (ISS) is an imaging-based approach where gene transcription is directly read within the tissue74. Target probes reverse genes’ transcription, followed by sequencing-by-ligation. Another microscopy-based approach is fluorescence in-situ hybridization (FISH), where a target sequence is detected by complementary fluorescent probe hybridization. FISH-based techniques such as MERFISH and seqFISH have detected approximately 10,000 genes at sub-cellular locations75,76. Unlike the imaging-based approach, NGS-based methods, such as 10x Visium, Slide-seq, DBiT-seq, smFISH and Seq-scope collect spatially-barcoded transcripts and then sequenced them77–80. The barcode of each sequencing read is used to map the spatial position, with the rest of the read being aligned to the transcriptome to generate a spatially-resolved gene expression map. Using spots with spatial barcodes, Asp et al. generated a transcriptional landscape of human embryonic hearts across three developmental stages8. While the spatial transcriptomics field is growing exponentially, the technique is challenged by limits to resolution and sensitivity. However, as spatial transcriptomics methods are compatible with paraffin-embedded tissues, retrospective analyses can be performed on samples collected decades ago in cardiovascular disease tissue biobank, enabling data exploration of spatial patterns and insights into cardiac disease biology.

By integrating spatial transcriptomics, snRNA-seq and scATAC, Kuppe et al. reported a comprehensive map delineating the spatiotemporal molecular changes at the early and late stages of myocardial infarction18. The landscape revealed spatial enrichment of distinct functional states of myogenic regions in control, remote, and border zones that were not revealed by examining cell-type compositions or histology alone. Cardiomyocyte-enriched molecular niches were depleted in ischemic-enriched samples, while fibrotic-and inflammatory-enriched molecular niches were depleted in myogenic-enriched samples. In contrast, molecular niches containing an abundance of fibroblasts and immune cells were more prevalent in remote and border regions than in control samples. Mapping of cell states revealed reproducible preferential distribution of distinct cardiomyocytes subtypes along distinct border zones: uninjured corner, transition zone from injured to remote myocardium, and transition zone within injured area. A pathway analysis of the spatial gene expression data revealed increased TGF signaling activity in the injured region, but a homogenous distribution of hypoxia pathway activity. The combination of spatial technologies and single-cell data presented an opportunity to investigate how the microenvironment of cardiac cells affects their states.

V. Future Applications of Single-cell Technologies in Cardiovascular Research

Single-cell QTL

GWAS revealed many genetic variants associated with complex diseases. Expression quantitative trait locus (eQTL) mapping determines the effects of genetic variants on gene expression levels. Molecular phenotyping such as eQTL mapping studies has enabled the prioritization of genetic variants within GWAS loci. These studies have shown that trait-associated genetic variants often function in a tissue- or cell-type-specific manner81. However, the average effect from bulk RNA sequencing has hampered our ability to study gene regulatory programs in distinct human cell types and interpret the non-coding DNA in the human genome.

eQTL studies using scRNA-seq have identified the cell-type-specific association between genetic variants and gene expression previously unobservable in bulk RNA-seq studies82. For example, Elorbany et al. used a scRNA-seq profile of 19 well-characterized Yoruba iPSC lines and uncovered dynamic eQTL related to gene expression changes over the cardiomyocyte differentiation time course83. Another single-cell eQTL analysis performed on fibroblast derived from 79 donors found a high degree of cell type-specific gene regulation missed by bulk RNA-seq84. In addition, the analyses identified novel cis-eQTLs for genes that characterize cell types during induced pluripotent stem cell (iPSC) reprogramming, located in functional genomic regions that reflect reprogramming efficiency based on epigenetic profiles. Those loci also demonstrated important cell type specificity that could impact iPSC reprogramming and pluripotency in a cell type-specific manner.

scRNA-seq of multiple samples in parallel at the scale required for an eQTL study remains challenging. Recently, a pooled experiment design has been developed for multiplexing multiple individuals in a single capture experiment, providing higher throughput and enabling the characterization of intrinsic line-to-line variation in differentiation efficiency in a controlled setting (Figure 5)85. The feasibility of pooled experiment design has been shown in multiple studies85,86. For example, Cuomo et al. performed pooled experiment design on 125 lines prior to endoderm differentiation and identified hundreds of genetic variants associated with temporal gene expression specific for mesendoderm and definite endoderm cell lineage specification86. Van der Wijst et al. generated ~25,000 single-cell expression profiles from the peripheral blood mononuclear cells (PBMCs) of 45 donors and identified genetic variants that significantly alter gene co-expression regulatory networks in a cell-type-specific manner82. More recently, using scRNA-seq data from 1.27 million PBMCs collected from 982 donors, there were able to map the genetic effects on gene expression in each immune cell type and identify 26,597 cell-specific eQTL, where about one-fifth of the eQTLs shared the same causal locus as a GWAS risk association87. The study uncovered the causal mechanism by which hundreds of loci contribute to autoimmune pathology through cell-type-specific gene expression changes using a Mendelian randomization approach. Notably, the study identified the causal route underlying 57 risk loci in multiple sclerosis. For example, the study showed that the loci at 3q12 causally act through changes in EAF2 expression, but only in immature and naïve B and memory B cells, even when the gene is ubiquitous across cell types.

Figure 5. Creation of Cell Village to minimize cost and batch effect in single-cell profiling.

Figure 5

Multiple cell lines or samples from genetically distinct donors can be pooled in a single dish and profiled using single-cell technologies. The identity of each cell can be deconvoluted based on the genotype of each donor. The feasibility of pooling samples has been demonstrated for human peripheral blood mononuclear cells (PBMCs) and cell lines. The pooling technique provides high

Careful consideration in the experiment design, such as the number of samples, number of cells per sample and reads per cell, is needed to produce the highest statistical power with the largest cost saving88. Given the same budget, the statistical power of cell-type-specific eQTL mapping can benefit from the low-coverage per-cell sequencing of more samples rather than the high-coverage sequencing of fewer samples88. Nonetheless, with the enhanced throughput offered by the single-cell pooled experiment design, we believe single-cell eQTL will provide us with the information required to reveal novel targets for personalized cardiovascular medicine.

Application of Machine Learning (ML) in Single-Cell Analysis

Single-cell sequencing datasets are high-dimensional, complex, and frequently noisy datasets. Therefore, analysis of these datasets faces several challenges: batch effect, limited and variable sensitivity, accurate representation, visualization, and data classification9. Here, we review how ML, particularly deep learning models, is being used to address these issues and apply to advances.

Critical to analyzing single-cell sequencing datasets is selecting the proper dimensionality reduction technique to transform raw high-dimensional data into a lower-dimensional representation that can be more easily visualized, classified, or used for downstream tasks89. Traditional dimensionality reduction has included Principal Component Analysis (PCA), t-Distributed Stochastic Neighbor Embedding (t-SNE), and Uniform Manifold Approximation and Projection for Dimension Reduction (UMAP), with PCA being a linear dimensionality reduction method and the basis of early scRNA-seq packages Seurat62.

Recently, an increasingly popular model for dimensionality reduction uses autoencoders90–92. Autoencoders are a class of deep learning models can learn to encode and decode a given input93. When applied to scRNA-seq datasets, autoencoders can learn a lower dimensionality representation of datasets while also capturing inherent structure within the dataset unsupervised. One of the main attributes of autoencoders is their adeptness at scaling with the growing size of scRNA-seq datasets. Lopez et al. developed a single workflow that can handle increasingly large scRNA seq datasets with greater efficiency and less memory94. Similarly, Ding et al. developed ‘scvis’ to handle increasingly larger datasets and demonstrated that ‘scvis’ has the added functionality of interpretability, allowing for uncertainty estimates.

Wang et al. tackled another critical problem in working with scRNA-seq datasets -- technical dropout events due to transcriptional burst effects and low amounts of RNA transcripts in single cells95. Similarly, Eraslan et al. developed a deep autoencoder network (DCA) for scalable denoising methods for increasingly large but sparse scRNA-seq datasets96.

Autoencoders have also been utilized to tackle one of the most pressing problems in normalizing batch effects. Two primary autoencoders strategies have been developed: learning a low dimensional representation over a common space invariant to technical confounders and developing an autoencoder to correct batch effect on the original count data. For example, Amodio et al. developed SAUCIE, which learns a representation that minimizes the discrepancy in probability distributions between the different batches of scRNA-seq97. Critically, SAUCE utilizes multitask learning to develop a framework optimized for clustering, visualization, imputation, and batch correction. Similarly, Wang et al. developed Batch Effect Removal Using Deep Autoencoders (BERMUDA), an autoencoder that removes batch effect by minimizing reconstruction loss and transfer loss between pairs of cell clusters during training98.

Finally, besides deep learning, other ML methods have also been applied successfully to single cell genomics data. For example, non-negative matrix factorization (NMF) is useful for integrating data from multiple modalities99. De-Convolution and Coupled-Clustering (DC3) performs joint analysis of bulk and single-cell data across multiple input data types from the heterogeneous cell population in scRNA-seq, scATAC-seq and single-cell chromatin contact assays100. Another method, Linked Inference of Genomic Experimental Relationships (LIGER), uses NMF to identify shared- and dataset-specific latent factors in defining cell types from multiple datasets99. Single-Cell Alignment Using Optimal Transport (SCOT) uses Optional Transport (OT) to align single-cell multi-omics modalities while preserving local geometry101. Besides multi-omics data integration, OT analysis is robust for time-course single-cell data102. Schiebinger et al. developed a conceptual framework, Waddington-OT, that uses the mathematical approach of optimal transport analysis to understand the cellular reprogramming trajectories102.

Single-Cell CRISPR screen

Pooled genetic screens based on CRISPR/Cas9 genome engineering system are widely used to study the functions of thousands of genes in a single experiment103. For instance, a FACS-based CRISPR screen for transcriptional regulators of CM maturation using a mouse Myh7-reporter line has recently uncovered several transcriptional regulators such as RNF20 and RNF40 implicated in congenital heart disease104. Different flavors of CRISPR screening methodology have been extensively reviewed by Nishiga et al105.

With the combinatorial approach of scRNA-seq, the CRISPR screen provides a powerful method to monitor transcriptome changes in response to perturbation at unprecedented resolution (Figure 1)106–108. These technologies, including Perturb-seq, CRISP-seq, Mosaic-seq, and CROP-seq, enabled a large-scale investigation of gene regulatory networks, genetic interactions, and enhancer-gene regulations in one experiment106–108. In general, a pooled CRISPR screen with scRNA-seq encodes the identity of the perturbation on an expressed single-guide RNAs (sgRNA) barcode. Depending on the context of the experiment, the cells are grown or challenged with stimuli, followed by scRNA-seq. The cellular barcode, unique to each cell, associates the cell’s transcriptional profile with the delivered genetic perturbations encoded by the sgRNAs. Therefore, single-cell CRISPR screening significantly eliminates the limitation of traditional screening experiments, where only one phenotype can be tested. However, successful CRISPR screen experiments require careful optimizations103.

Conclusion

Single-cell technology has made significant progress from single omics to analysis of multi-omics measurements from the same cell. In a short period, single-cell studies have expanded the horizon from establishing cell composition in healthy tissues to delineating gene regulatory network perturbations in cardiovascular diseases. The single-cell atlases generated from healthy tissues now serve as the baseline for comparison to unravel cell type-specific and chamber-specific changes in disease. Beyond transcriptomics, epigenomics, proteomics, and spatial transcriptomics are relevant to elucidate gene regulator mechanisms at single-cell resolution. With continuing innovation of experimental techniques and ML analytics, single-cell technology has the potential to transform biomedical research and advance precision medicine fundamentally.

Acknowledgment

We would like to acknowledge funding support from the American Heart Association grant 20POST35210896 (WLWT), F30 HL156478 (AZ), National Institutes of Health (NIH) R01 HG010359 (WW), R01 HL145676, R01 HL146690, R01 HL130020, R01 HL145676, and R01 HL141371 (JCW). The authors thank B.C. Wu (Stanford University, USA), D.T. Paik (Stanford University, USA), S. S. Hye (Stanford University, USA) and L.C.J. Mick (National University of Singapore, Singapore) for critical reading of the manuscript.

Footnotes

Competing Interests

W.J.G. is an equity holder of 10X Genomics and a co-founder of Protillion Biosciences and J.C.W. is a co-founder and scientific advisory board member for Greenstone Biosciences, but the work presented here is completely independent. The remaining authors declare no competing interests.

References

  • 1.Shapiro E, Biezuner T & Linnarsson S Single-cell sequencing-based technologies will revolutionize whole-organism science. Nat Rev Genet 14, 618–630 (2013). [DOI] [PubMed] [Google Scholar]
  • 2.Tang F et al. mRNA-Seq whole-transcriptome analysis of a single cell. Nat Methods 6, 377–382 (2009). [DOI] [PubMed] [Google Scholar]
  • 3.Cao J et al. The single-cell transcriptional landscape of mammalian organogenesis. Nature 566, 496–502 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Buettner F et al. Computational analysis of cell-to-cell heterogeneity in single-cell RNA-sequencing data reveals hidden subpopulations of cells. Nat Biotechnol 33, 155–160 (2015). [DOI] [PubMed] [Google Scholar]
  • 5.Domcke S et al. A human cell atlas of fetal chromatin accessibility HHS Public Access. Science (1979) 370, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Turner AW et al. Single-nucleus chromatin accessibility profiling highlights regulatory mechanisms of coronary artery disease risk. Nat Genet 54, 804–816 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Örd T et al. Single-Cell Epigenomics and Functional Fine-Mapping of Atherosclerosis GWAS Loci. Circ Res 240–258 (2021) doi: 10.1161/CIRCRESAHA.121.318971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Asp M et al. A Spatiotemporal Organ-Wide Gene Expression and Cell Atlas of the Developing Human Heart. Cell 179, 1647–1660.e19 (2019). Spatial-temporal single cell analyses of human embryo hearts, identifying distinctive gene profiles that correspond to discrete anatomical areas in each developmental stage, and providing a web-based resource of the developing human heart.
  • 9.Stuart T & Satija R Integrative single-cell analysis. Nature Reviews Genetics vol. 20 257–272 Preprint at 10.1038/s41576-019-0093-7 (2019). [DOI] [PubMed] [Google Scholar]
  • 10.Wang Z, Gerstein M & Snyder M RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet 10, 57–63 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Paik DT, Cho S, Tian L, Chang HY & Wu JC Single-cell RNA sequencing in cardiovascular development, disease and medicine. Nature Reviews Cardiology vol. 17 457–473 Preprint at 10.1038/s41569-020-0359-y (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Litviňuková M et al. Cells of the adult human heart. Nature 588, 466–472 (2020). Single cell RNA sequencing on an adult human heart reveals the distinct atrial and ventricular subsets of cells with different developmental origins and specialized properties, highlighting the cellular heterogeneity of cardiac cells and providing an online human cardiac cell atlas for a valuable reference.
  • 13.Chaffin M et al. Single-nucleus profiling of human dilated and hypertrophic cardiomyopathy. Nature 608, 174–180 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Tucker NR et al. Transcriptional and Cellular Diversity of the Human Heart. Circulation 142, 466–482 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Koenig AL et al. Single-cell transcriptomics reveals cell-type-specific diversification in human heart failure. Nature Cardiovascular Research 1, 263–280 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Nicin L et al. A human cell atlas of the pressure-induced hypertrophic heart. Nature Cardiovascular Research 1, 174–185 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Nicin L et al. Single Nuclei Sequencing Reveals Novel Insights into the Regulation of Cellular Signatures in Children with Dilated Cardiomyopathy. Circulation 1704–1719 (2021) doi: 10.1161/CIRCULATIONAHA.120.051391. [DOI] [PubMed] [Google Scholar]
  • 18. Kuppe C et al. Spatial multi-omic map of human myocardial infarction. Nature 608, 766–777 (2022). Spatial multi-omic analyses of human myocardial infarction tissue produced an integrative high-resolution map of cardiac cell-type compositions at enhanced resolution, revealing changes in the cardiac transcriptome and epigenome by identifying distinct cardiac tissue structures during injury, repair, and remodeling.
  • 19.Reichart D et al. Pathogenic variants damage cell composition and single cell transcription in cardiomyopathies. Science (1979) 377, eabo1984 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hocker JD et al. Cardiac cell type-specific gene regulatory programs and disease risk association. Sci. Adv vol. 7 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Svensson V, Vento-Tormo R & Teichmann SA Exponential scaling of single-cell RNA-seq in the past decade. Nature Protocols vol. 13 599–604 Preprint at 10.1038/nprot.2017.149 (2018). [DOI] [PubMed] [Google Scholar]
  • 22.Wu AR et al. Quantitative assessment of single-cell RNA-sequencing methods. Nat Methods 11, 41–46 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Buenrostro JD et al. Single-cell chromatin accessibility reveals principles of regulatory variation. Nature 523, 486–490 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Lombardo JA, Aliaghaei M, Nguyen QH, Kessenbrock K & Haun JB Microfluidic platform accelerates tissue processing into single cells for molecular analysis and primary culture models. Nat Commun 12, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.See K et al. Single cardiomyocyte nuclear transcriptomes reveal a lincRNA-regulated de-differentiation and cell cycle stress-response in vivo. Nat Commun 8, (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Denisenko E et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome Biol 21, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Bakken TE et al. Single-nucleus and single-cell transcriptomes compared in matched cortical cell types. PLoS One 13, (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Rosenberg AB et al. Single-cell profiling of the developing mouse brain and spinal cord with split-pool barcoding. Science (1979) 360, 176–182 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Cao J et al. Comprehensive single-cell transcriptional profiling of a multicellular organism. Science (1979) 357, 661–667 (2017). Sci-RNA-seq, a combinatorial indexing method for profiling the transcriptomes of single cells or nuclei, was developed and profiled over 50,000 cells from the L2 larval stage of C. Elegans, resulting in >50-fold “shotgun” coverage of its somatic cell composition.
  • 30.Cusanovich DA et al. Multiplex single-cell profiling of chromatin accessibility by combinatorial cellular indexing. Science (1979) 348, 910–914 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Li H & Humphreys BD Single Cell Technologies: Beyond Microfluidics. Kidney360 2, 1196–1204 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Lareau CA et al. Droplet-based combinatorial indexing for massive-scale single-cell chromatin accessibility. Nat Biotechnol 37, 916–924 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Ma S et al. Chromatin Potential Identified by Shared Single-Cell Profiling of RNA and Chromatin. Cell 183, 1103–1116.e20 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Parse Bioscience. Parse Bioscience. https://www.parsebiosciences.com.
  • 35.Greener JG, Kandathil SM, Moffat L & Jones DT A guide to machine learning for biologists. Nat Rev Mol Cell Biol 23, 40–55 (2022). [DOI] [PubMed] [Google Scholar]
  • 36.Moulik M et al. ANKRD1, the Gene Encoding Cardiac Ankyrin Repeat Protein, Is a Novel Dilated Cardiomyopathy Gene. J Am Coll Cardiol 54, 325–333 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Wang F et al. RNAscope: A Novel in Situ RNA Analysis Platform for Formalin-Fixed, Paraffin-Embedded Tissues. The Journal of Molecular Diagnostics 14, 22–29 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Alexanian M et al. A transcriptional switch governs fibroblast activation in heart disease. Nature 595, 438–443 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Orchard S et al. The MIntAct project--IntAct as a common curation platform for 11 molecular interaction databases. Nucleic Acids Res 42, D358–63 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Jin S et al. Inference and analysis of cell-cell communication using CellChat. Nat Commun 12, 1088 (2021). CellChat is a tool that can statistically predict and analyze intercellular communication networks using single-cell RNA-sequencing data, hence assisting in the discovery of novel intercellular communications in various tissues.
  • 41.Efremova M, Vento-Tormo M, Teichmann SA & Vento-Tormo R CellPhoneDB: inferring cell–cell communication from combined expression of multi-subunit ligand–receptor complexes. Nat Protoc 15, 1484–1506 (2020). [DOI] [PubMed] [Google Scholar]
  • 42.Cabello-Aguilar S et al. SingleCellSignalR: Inference of intercellular networks from single-cell transcriptomics. Nucleic Acids Res 48, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Gerety SS & Anderson DJ Cardiovascular ephrinB2 function is essential for embryonic angiogenesis. Development 129, 1397–1410 (2002). [DOI] [PubMed] [Google Scholar]
  • 44.Saelens W, Cannoodt R, Todorov H & Saeys Y A comparison of single-cell trajectory inference methods. Nat Biotechnol 37, 547–554 (2019). [DOI] [PubMed] [Google Scholar]
  • 45.Bais AS & Kostka D scds: computational annotation of doublets in single-cell RNA sequencing data. Bioinformatics 36, 1150–1158 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.DePasquale EAK et al. DoubletDecon: Deconvoluting Doublets from Single-Cell RNA-Sequencing Data. Cell Rep 29, 1718–1727.e8 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.McGinnis CS, Murrow LM & Gartner ZJ DoubletFinder: Doublet Detection in Single-Cell RNA Sequencing Data Using Artificial Nearest Neighbors. Cell Syst 8, 329–337.e4 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Young MD & Behjati S SoupX removes ambient RNA contamination from droplet-based single-cell RNA sequencing data. Gigascience 9, giaa151 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Yang S et al. Decontamination of ambient RNA in single-cell RNA-seq with DecontX. Genome Biol 21, 57 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Wang X, He Y, Zhang Q, Ren X & Zhang Z Direct Comparative Analyses of 10X Genomics Chromium and Smart-seq2. Genomics Proteomics Bioinformatics 19, 253–266 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Osorio D & Cai JJ Systematic determination of the mitochondrial proportion in human and mice tissues for single-cell RNA-sequencing data quality control. Bioinformatics 37, 963–967 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Maniatis T Mechanisms of Alternative Pre-mRNA Splicing. Science (1979) 251, 33–34 (1991). [DOI] [PubMed] [Google Scholar]
  • 53.Khan MAF et al. RBM20 Regulates Circular RNA Production From the Titin Gene. Circ Res 119, 996–1003 (2016). [DOI] [PubMed] [Google Scholar]
  • 54.van den Hoogenhof MMG et al. RBM20 Mutations Induce an Arrhythmogenic Dilated Cardiomyopathy Related to Disturbed Calcium Handling. Circulation 138, 1330–1342 (2018). [DOI] [PubMed] [Google Scholar]
  • 55.Harvey SE & Cheng C Methods for Characterization of Alternative RNA Splicing. in Long Non-Coding RNAs: Methods and Protocols (eds. Feng Y & Zhang L) 229–241 (Springer; New York, 2016). doi: 10.1007/978-1-4939-3378-5_18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Volden R & Vollmers C Single-cell isoform analysis in human immune cells. Genome Biol 23, 47 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Wong ES et al. Deep conservation of the enhancer regulatory code in animals. Science (1979) 370, (2020). [DOI] [PubMed] [Google Scholar]
  • 58.Wirka RC et al. Atheroprotective roles of smooth muscle cell phenotypic modulation and the TCF21 disease gene as revealed by single-cell analysis. Nat Med 25, 1280–1289 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Pliner HA et al. Cicero Predicts cis-Regulatory DNA Interactions from Single-Cell Chromatin Accessibility Data. Mol Cell 71, 858–871.e8 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Tan WLW et al. Epigenomes of human hearts reveal new genetic variants relevant for cardiac disease and phenotype. Circ Res 761–777 (2020) doi: 10.1161/CIRCRESAHA.120.317254. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Fulco CP et al. Activity-by-contact model of enhancer–promoter regulation from thousands of CRISPR perturbations. Nat Genet 51, 1664–1669 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62. Hao Y et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e29 (2021). This study used simultaneous assessment of various modalities to define cellular states based on multimodal data, which improves our ability to identify previously unreported subpopulations, such as lymphoid.
  • 63.Cao J et al. Joint profiling of chromatin accessibility and gene expression in thousands of single cells. Science (1979) 361, 1380–1385 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Zhu C et al. An ultra high-throughput method for single-cell joint analysis of open chromatin and transcriptome. Nat Struct Mol Biol 26, 1063–1070 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Pan Q, Shai O, Lee LJ, Frey BJ & Blencowe BJ Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing. Nat Genet 40, 1413–1415 (2008). [DOI] [PubMed] [Google Scholar]
  • 66.Emmerich CH et al. Improving target assessment in biomedical research: the GOT-IT recommendations. Nature Reviews Drug Discovery vol. 20 64–81 Preprint at 10.1038/s41573-020-0087-3 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67.Assarsson E et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One 9, (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Frei AP et al. Highly multiplexed simultaneous detection of RNAs and proteins in single cells. Nat Methods 13, 269–275 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69. Peterson VM et al. Multiplexed quantification of proteins and transcripts in single cells. Nat Biotechnol 35, 936–939 (2017). REAP-seq is a technique that uses DNA-labeled antibodies and droplet microfluidics to assess the levels of gene and protein expression in single cells in order to identify and describe an unidentified cell type.
  • 70.Stoeckius M et al. Simultaneous epitope and transcriptome measurement in single cells. Nat Methods 14, 865–868 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71.Chung H et al. Joint single-cell measurements of nuclear proteins and RNA in vivo. Nat Methods 18, 1204–1212 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 72.Chen AF et al. NEAT-seq: simultaneous profiling of intra-nuclear proteins, chromatin accessibility and gene expression in single cells. Nat Methods (2022) doi: 10.1038/s41592-022-01461-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73.Longo SK, Guo MG, Ji AL & Khavari PA Integrating single-cell and spatial transcriptomics to elucidate intercellular tissue dynamics. Nature Reviews Genetics vol. 22 627–644 Preprint at 10.1038/s41576-021-00370-8 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Crosetto N, Bienko M & van Oudenaarden A Spatially resolved transcriptomics and beyond. Nat Rev Genet 16, 57–66 (2015). [DOI] [PubMed] [Google Scholar]
  • 75.Chen KH, Boettiger AN, Moffitt JR, Wang S & Zhuang X Spatially resolved, highly multiplexed RNA profiling in single cells. Science (1979) 348, (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 76.Eng CHL et al. Transcriptome-scale super-resolved imaging in tissues by RNA seqFISH+. Nature 568, 235–239 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 77.Liu Y et al. High-Spatial-Resolution Multi-Omics Sequencing via Deterministic Barcoding in Tissue. Cell 183, 1665–1681.e18 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 78. Cho CS et al. Microscopic examination of spatial transcriptome using Seq-Scope. Cell 184, 3559–3572.e22 (2021). Seq-Scope, a spatial barcoding technology, outperforms existing methods by several orders of magnitude by enabling researchers to observe every gene expressed, as well as single cells and structures inside those cells, with an astonishingly high resolution of 0.6 micrometers.
  • 79.Rodriques SG et al. Slide-seq: A scalable technology for measuring genome-wide expression at high spatial resolution. https://www.science.org. [DOI] [PMC free article] [PubMed]
  • 80.Resolve Biosciences Inc. Key Findings. https://resolvebiosciences.com/.
  • 81.Võsa U et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat Genet 53, 1300–1310 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 82. van der Wijst MGP et al. Single-cell RNA sequencing identifies cell type-specific cis-eQTLs and co-expression QTLs. Nat Genet 50, 493–497 (2018). This study employed Single-cell eQTL analysis to uncover genetic variations that influence regulatory networks by producing expression profiles of 25,000 peripheral blood mononuclear cells from 45 donors and identifying previously described cis-eQTLs, as well as new cell type-specific cis-eQTLs.
  • 83.Elorbany R et al. Single-cell sequencing reveals lineage-specific dynamic genetic regulation of gene expression during human cardiomyocyte differentiation. PLoS Genet 18, (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 84.Neavin D et al. Single cell eQTL analysis identifies cell type-specific genetic control of gene expression in fibroblasts and reprogrammed induced pluripotent stem cells. Genome Biol 22, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 85.Kang HM et al. Multiplexed droplet single-cell RNA-sequencing using natural genetic variation. Nat Biotechnol 36, 89–94 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 86.Cuomo ASE et al. Single-cell RNA-sequencing of differentiating iPS cells reveals dynamic genetic effects on gene expression. Nat Commun 11, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 87.Yazar S et al. Single-cell eQTL mapping identifies cell type–specific genetic control of autoimmune disease. Science (1979) 376, (2022). [DOI] [PubMed] [Google Scholar]
  • 88.Mandric I et al. Optimized design of single-cell RNA sequencing experiments for cell-type-specific eQTL analysis. Nat Commun 11, 5504 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 89.Becht E et al. Dimensionality reduction for visualizing single-cell data using UMAP. Nat Biotechnol 37, 38–47 (2019). [DOI] [PubMed] [Google Scholar]
  • 90.Tran D et al. Fast and precise single-cell data analysis using a hierarchical autoencoder. Nat Commun 12, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 91.Svensson V, Gayoso A, Yosef N & Pachter L Interpretable factor models of single-cell RNA-seq via variational autoencoders. Bioinformatics 36, 3418–3421 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 92.Grønbech CH et al. ScVAE: Variational auto-encoders for single-cell gene expression data. Bioinformatics 36, 4415–4422 (2020). [DOI] [PubMed] [Google Scholar]
  • 93.Dong G, Liao G, Liu H & Kuang G A Review of the Autoencoder and Its Variants: A Comparative Perspective from Target Recognition in Synthetic-Aperture Radar Images. IEEE Geoscience and Remote Sensing Magazine vol. 6 44–68 Preprint at 10.1109/MGRS.2018.2853555 (2018). [DOI] [Google Scholar]
  • 94.Ding J, Condon A & Shah SP Interpretable dimensionality reduction of single cell transcriptome data with deep generative models. Nat Commun 9, (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 95.Wang D & Gu J VASC: Dimension Reduction and Visualization of Single-cell RNA-seq Data by Deep Variational Autoencoder. Genomics Proteomics Bioinformatics 16, 320–331 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 96.Eraslan G, Simon LM, Mircea M, Mueller NS & Theis FJ Single-cell RNA-seq denoising using a deep count autoencoder. Nat Commun 10, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 97.Amodio M et al. Exploring single-cell data with deep multitasking neural networks. Nat Methods 16, 1139–1145 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 98.Wang T et al. BERMUDA: A novel deep transfer learning method for single-cell RNA sequencing batch correction reveals hidden high-resolution cellular subtypes. Genome Biol 20, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 99.Liu J et al. Jointly defining cell types from multiple single-cell datasets using LIGER. Nat Protoc 15, 3632–3662 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 100.Zeng W et al. DC3 is a method for deconvolution and coupled clustering from bulk and single-cell genomics data. Nat Commun 10, (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 101.Demetci P, Santorella R, Sandstede B, Stafford Noble W & Singh R Gromov-Wasserstein Optimal Transport to Align Single-Cell Multi-Omics Data. (2020). [Google Scholar]
  • 102. Schiebinger G et al. Optimal-Transport Analysis of Single-Cell Gene Expression Identifies Developmental Trajectories in Reprogramming. Cell 176, 928–943.e22 (2019). This paper introduced Waddington-OT, a method for analyzing developmental time courses in order to model the regulatory systems that underpin them and deduce ancestor-descendant outcomes showing a broader variety of developmental programs than those previously documented, including cell adaptability in either a terminal stromal state or a mesenchymal-to-epithelial transition state.
  • 103.Bock C et al. High-content CRISPR screening. Nature Reviews Methods Primers 2, (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 104.VanDusen NJ et al. Massively parallel in vivo CRISPR screening identifies RNF20/40 as epigenetic regulators of cardiomyocyte maturation. Nat Commun 12, (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 105.Nishiga M, Liu C, Qi LS & Wu JC The use of new CRISPR tools in cardiovascular research and medicine. Nature Reviews Cardiology Preprint at 10.1038/s41569-021-00669-3 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 106.Schraivogel D et al. Targeted Perturb-seq enables genome-scale genetic screens in single cells. Nat Methods 17, 629–635 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 107.Dixit A et al. Perturb-Seq: Dissecting Molecular Circuits with Scalable Single-Cell RNA Profiling of Pooled Genetic Screens. Cell 167, 1853–1866.e17 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 108.Datlinger P et al. Pooled CRISPR screening with single-cell transcriptome readout. Nat Methods 14, 297–301 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 109.Wolf FA, Angerer P & Theis FJ SCANPY: large-scale single-cell gene expression data analysis. Genome Biol 19, 15 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 110. Cao J et al. A human cell atlas of fetal gene expression. Science (1979) 370, (2020). This study generated comprehensive atlases of early human development from two studies by Cao et al. and Domcke et al., using a pooled approach with three levels of combinatorial indexing, and examined the single-cell gene expression and chromatin landscapes of 15 organs in fetal samples.
  • 111.Wang L et al. Single-cell reconstruction of the adult human heart during heart failure and recovery reveals the cellular landscape underlying cardiac function. Nat Cell Biol 22, 108–119 (2020). [DOI] [PubMed] [Google Scholar]
  • 112.Barruet E et al. Functionally heterogeneous human satellite cells identified by single cell RNA sequencing. Elife 9, (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 113.Paik DT et al. Large-scale single-cell RNA-seq reveals molecular signatures of heterogeneous populations of human induced pluripotent stem cell-derived endothelial cells. Circ Res 123, 443–450 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 114.Li Y et al. Single-Cell Transcriptome Analysis Reveals Dynamic Cell Populations and Differential Gene Expression Patterns in Control and Aneurysmal Human Aortic Tissue. Circulation 1374–1388 (2020) doi: 10.1161/CIRCULATIONAHA.120.046528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 115.Pan H et al. Single-Cell Genomics Reveals a Novel Cell State During Smooth Muscle Cell Phenotypic Switching and Potential Therapeutic Targets for Atherosclerosis in Mouse and Human. Circulation 142, 2060–2075 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 116.Selewa A et al. Systematic Comparison of High-throughput Single-Cell and Single-Nucleus Transcriptomes during Cardiomyocyte Differentiation. Sci Rep 10, 1535 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 117.Depuydt MAC et al. Microanatomy of the Human Atherosclerotic Plaque by Single-Cell Transcriptomics. Circ Res 1437–1455 (2020) doi: 10.1161/CIRCRESAHA.120.316770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 118.Cui Y et al. Single-Cell Transcriptome Analysis Maps the Developmental Track of the Human Heart. Cell Rep 26, 1934–1950.e5 (2019). [DOI] [PubMed] [Google Scholar]
  • 119.Nomura S et al. Cardiomyocyte gene programs encoding morphological and functional signatures in cardiac hypertrophy and failure. Nat Commun 9, (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 120.Burridge PW et al. Chemically defined and small molecule-based generation of human cardiomyocytes. Nat Methods 11, 855–860 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES