Protocol for the isolation and single-nuclei multiomic analyses of the human fetal epicardium

Summary

The characterization of cell populations that reside in the outer layer of the heart has been hindered by difficulties in their isolation. Here, we present a protocol for isolation and single-nuclei multiomic analyses of the human fetal epicardium. We describe steps for microdissection, isolation, and enrichment of epicardial cells by mechanical dissociations and direct lysis. We then detail procedures for integrating transcriptome and chromatin accessibility datasets. This approach allows the analysis of diverse cell populations, marked by unique cis-regulatory elements.

For complete details on the use and execution of this protocol, please refer to Travisano et al.¹

Graphical abstract

Highlights

• •Microdissection of the outer layer of the human fetal hearts
• •Library preparation and quality control for the epicardial snMultiome
• •Integrative analyses of snRNA-seq and snATAC-seq from the same cells
• •Identification of cis-elements and binding motif from the human fetal epicardium

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

The characterization of cell populations that reside in the outer layer of the heart has been hindered by difficulties in their isolation. Here, we present a protocol for isolation and single-nuclei multiomic analyses of the human fetal epicardium. We describe steps for microdissection, isolation, and enrichment of epicardial cells by mechanical dissociations and direct lysis. We then detail procedures for integrating transcriptome and chromatin accessibility datasets. This approach allows the analysis of diverse cell populations, marked by unique cis-regulatory elements.

Before you begin

The epicardium and sub-epicardium contain different cell types and progenitors, including blood and lymphatic endothelium, interstitial cells, and inflammatory cells in the fetal hearts.² Sample preparations for single cell sequencing datasets used enzymatic digestions that require cell dissociation to be performed at least partially at 37°C. This process might affect scRNA-seq data quality in terms of cellular yield and viability compared to mechanical dissociations at 4°C.³ Micro-dissection of the surface layer of the heart tissue is a useful method to enrich the epicardial cells. This procedure helps to decrease expression of dissociation-induced genes and reduces the amount of input material compared to methods using cells dissociated from whole hearts. A similar strategy can also be used for other organs. We provide this mechanical dissociation method as an alternative to the enzymatic digestion and show that our method improves the data quality of Single Cell/Nuclei Multiome ATAC + Gene Expression profiling. The R Studio packages Seurat,⁴ for gene expression and Signac⁵ for DNA accessibility combined together are powerful and essential tools to analyze the datasets. To decrease potential sampling error, it is recommended to process ≥2 biological replicates and to capture at least 10K nuclei.

Institutional permissions

This study complies with the regulatory compliance for the use of human material. Samples were collected with written informed consent under the supervision of an appropriate Institutional Review Board (IRB) Protocol. Human samples were de-identified and assigned bar code labels linking to metadata such as subject number, blood draw type etc.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Tris-HCl (1 M)#	Sigma	T2194-100ML
NaCl (5 M)#	Sigma	59222C
MgCl2 (1 M)#	Sigma	M1028
MACS BSA stock solution#	Miltenyi Biotec	130-091-376
Tween-20#	Bio-Rad	1662404
Nonidet P 40#	Sigma	74385
Nuclease-free water	Thermo Fisher Scientific	AM9932
Digitonin (5%)#	Thermo Fisher Scientific	BN2006
DTT (dithiothreitol)#	Sigma	646563
Protector RNase inhibitor (2,000 U) #	Roche/Sigma	3335399001
Nuclei buffer∗ (20X)#	10× Genomics	2000153/2000207
RNase AWAY	Thermo Fisher Scientific	7002
Ethanol, absolute (200 proof)#	Sigma	E7023-500ML
Other
Vortex mixer	VWR	10153-838
Divided polystyrene reservoirs	VWR	41428-958
DNA LoBind tubes, 1.5 mL	Eppendorf	022431021
Mini centrifuge	Thermo Fisher Scientific	C1012
Eppendorf ThermoMixer C	Eppendorf	5382000023/2231000574
BD Falcon cell strainers (40 μm)#	Thermo Fisher Scientific	352340
4200 TapeStation∗	Agilent	G2991AA
High sensitivity D5000 ScreenTape	Agilent	5067-5592
High sensitivity D5000 reagents	Agilent	5067-5593
NovaSeq 6000#	Illumina	N/A
Countess 3 FL automated cell counter#	Thermo Fisher Scientific	AMQAF2000
DAPI#	Thermo Fisher Scientific	R37606
Countess cell counting chamber slides#	Thermo Fisher Scientific	C10283/C10228
Hemocytometer Neubauer	N/A	N/A
Critical commercial assays
Chromium Next GEM Single Cell Multiome ATAC + gene expression reagent bundle, 4 rxns	10× Genomics	Cat. #1000285
Chromium Next GEM Chip J single cell, 16 rxns	10× Genomics	Cat. #1000230
Dual index kit TT set A, 96 rxns	10× Genomics	Cat. #1000215
Deposited data
snRNA-seq 10w (2 biological replicates)	GEO	GEO ID: GSE241128
snRNA-seq 11w (3 biological replicates)	GEO	GEO ID: GSE241128
Software and algorithms
Cell Ranger ARC	10× Genomics	https://support.10xgenomics.com/single-cell-multiome-atac-gex/software/pipelines/latest/what-is-cell-ranger-arc
R version 4.1.1	R Project	https://www.r-project.org
BioRender	Science Suite Inc.	https://biorender.com/
ggplot2	Hadley Wickham	v3.3.6
MACS2	(Zhang et al.6)	https://pypi.org/project/MACS2/
chromVAR	(Schep et al.7)	https://greenleaflab.github.io/chromVAR/index.html
Seurat	(Hao et al.4)	https://satijalab.org/seurat/
Signac	(Stuart et al.5)	https://stuartlab.org/signac/
JASPAR 2020 v 0.99.10	(Fornes et al.8)	https://bioconductor.org/packages/release/data/annotation/html/JASPAR2020.html
TFBSTools v 1.36.0	(Tan and Lenhard9)	https://bioconductor.org/packages/release/bioc/html/TFBSTools.html

# In the section of “Chemicals, peptides, and recombinant proteins” we used the reagents suggested from 10× Genomics when available. Although alternatives are available from other companies and we expect compatible reagents to work similarly, we do not recommend using substitutions not present in this list or that have not been tested by 10× Genomics. Please refer to Chromium Single Cell Multiome ATAC + Gene Expression solution for more information.

Materials and equipment

Wash buffer∗

Component (storage)	Stock	Final	2 mL for 1 sample
Tris-HCl pH 7.4 (RT)	1 M	10 mM	20 μL
NaCl (RT)	5 M	10 mM	4 μL
MgCl2 (RT)	1 M	3 mM	6 μL
BSA (4°C)	10%	1%	200 μL
Tween-20 (RT)	10%	0.10%	20 μL
DTT	1000 mM	1 mM	2 μL
RNase Inhibitor	40 U/μL	1 U/μL	50 μL
Nuc Free water (RT)	–		1.7 mL

Prepare the same day and store on wet ice or at 4°C before use.

1× lysis buffer∗

Component (storage)	Stock	Final	∼100 μL is used for 2 samples)
Tris-HCl pH 7.4 (RT)	1 M	10 mM	1 μL
NaCl (RT)	5 M	10 mM	0.2 μL
MgCl2 (RT)	1 M	3 mM	0.3 μL
Tween-20 (RT)	10%	0.10%	1 μL
Nonidet P40 Substitute (RT)	10%	0.10%	1 μL
Digitonin (Incubate at 65°C to dissolve precipitate before use)	5%	0.01%	0.2 μL
BSA (4°C)	10%	1%	10 μL
DTT	1000 mM	1 mM	0.1 μL
RNAse Inhibitor	40 U/μL	1 U/μL	2.5 μL
Nuc Free water (RT)	–		83.7 μL

Prepare the same day and store on wet ice or at 4°C before use.

Lysis dilution buffer∗

Component (storage)	Stock	Final	1 mL for 2 samples
Tris-HCl pH 7.4 (RT)	1 M	10 mM	10 μL
NaCl (RT)	5 M	10 mM	2 μL
MgCl2 (RT)	1 M	3 mM	3 μL
BSA (4°C)	10%	1%	100 μL
DTT	1000 mM	1 mM	1 μL
RNAse Inhibitor	40 U/μL	1 U/μL	25 μL
Nuclease Free water (RT)	–		859 μL

Prepare the same day and store on wet ice or at 4°C before use.

0.1× lysis buffer∗

Component (storage)	Stock	Final	1 mL for 2 sample
1× Lysis Buffer (ice)	1×	0.1×	100 μL
Lysis Dilution Buffer (ice)			900 μL

Prepare the same day and store on wet ice or at 4°C before use.

Diluted Nuclei Buffer∗

Component (storage)	Stock	Final	1 mL for 1 sample
Nuclei Buffer 20× (−20°C; provided by 10× Genomics kit)	20×	1×	50 μL
DTT	1000 mM	1 mM	1 μL
RNAse Inhibitor	40 U/μL	1 U/μL	25 μL
Nuclease Free water (RT)			924 μL

Prepare the same day and store on wet ice or at 4°C before use.

∗All wash, lysis and nuclei buffers were prepared following the guidelines from the 10× Genomics protocol. Please refer to the Demonstrated protocol for more information at the following link: https://cdn.10xgenomics.com/image/upload/v1660261285/support-documents/CG000365_DemonstratedProtocol_NucleiIsolation_ATAC_GEX_Sequencing_RevC.pdf.

Materials needed for human epicardial dissection: 70% ethanol (prepare a fresh aliquot), Kimwipes, forceps, spatula, 10 cm petri-dishes, ice, nitrile gloves, nuclease-free and DNA 1.5 mL tubes (pre-labeled).

Buffers for nuclei digestion/isolation

CRITICAL: Please prepare fresh buffers before use and keep the buffers on wet ice before adding to the tissue/nuclei. Do not store the buffers more than 6 h on ice.

Note: Prepare ∼2 mL Wash Buffer/sample.

Note: Required amount of 0.1× Lysis Buffer is 1 mL for 2 samples.

Note: Required amount of Diluted Nuclei Buffer can be prepared after nuclei counting and depends on the targeted nuclei capture.

Hardware preparation

We have conducted the analysis on a MacOS computer. Windows is not recommended as the MACS2⁶ (Model-based Analysis for ChIP-seq) algorithm has only been tested on Linux or Mac OS. The RAM requirement depends on the number of cells to be analyzed. 16 GB RAM should be sufficient for an initial analysis.

Data analysis section

• •
  snRNA-seq and snATAC-seq Fastq data processing. The required package can be installed on your local computer or High-Performance Computing Center. Cell Ranger Arc pipelines were used to process the Chromium Single Cell Multiome ATAC + Gene Expression sequencing. 2.0.0 was used against the GRCh38-2020-A.

  • ○
    Cell Ranger ARC (v2.0.0.).
• •R software and required packages. Seurat and Signac are essential packages for analyzing, interpreting, and exploring single-nucleus Multiomic datasets. For this protocol, we used R version 4.1.1.

Step-by-step method details

Part 1: Experimental lab protocol

Preparation of equipment and solutions

Timing: 30–60 min

• 1.All dissection tools and pipettes should be autoclaved or sterilized (and cleaned with RNase AWAY). Proper PPE should be used throughout the protocol.
• 2.The following solutions should be fresh, prepared on the day of the experiment and kept on ice throughout the procedure: Lysis Buffer (LB) and Wash Buffer (WB).

Preparation of the micro-dissociation in DMEM media

Timing: 45 min

CRITICAL: To successfully microdissect the samples for a good epicardial preparation, the procedure should be performed on fresh and unfixed samples under a stereoscope.

• 3.
  Dissect fetal hearts out in a Petri dish with sterile DMEM.

  • a.
    Gently remove atria.
  • b.
    Gently position the ventricle (including the aorta and outflow tract (OFT) at the bottom of the dish with the forceps.
  • c.
    In the meantime, use micro-forceps to dissect the epicardium from the ventricle.
  • d.
    Gently peel the epithelium from the peritruncal regions.

CRITICAL: Because the tissue is very soft, we strongly recommend not to use any motorized tissue grinder or manual pestle.

• 4.Transfer the transparent epicardial/sub-epicardial cell preparation using a P200 Eppendorf pipette into a Petri dish containing 200 μL of cold 0.1X LB (Lysis Buffer).
• 5.Transfer sample to a 1.5 mL Eppendorf containing a total of 400 μL of 0.1X LB and digested by pipetting for 15–20 min directly without prior single cell enzymatic dissociation of the tissue.
• 6.Incubate lysate for 5 min on ice and centrifuge at 500 rcf for 5 min.
• 7.Discard supernatant and resuspend the pellet in 400 μL of WB and centrifuged at 500 rcf for 5 min (Step repeated 2 times).

CRITICAL: Incomplete digestion of the tissue and subsequent lysis of the nuclei might result in multi nuclear aggregation/clumps in the preparation.

Note: Presence of excessive debris and irregular nuclear shape combined with nuclei clumping, a known consequence of the chromatin properties to stick together, could be a sign of overlysis but could also be due to the excess of reagents such as the Digitonin used in the preparation of the lysis buffer.

CRITICAL: Over digestion of the tissue and subsequent lysis of the nuclei might result in damaging of the nuclear membrane and excessive debris.

• 8.Use a 40 um strainer to directly filter the nuclei preparation to a new Eppendorf and to remove the cell debris.

Note: Excess of RNase inhibitor may impact the nuclei digestion.

• 9.Resuspend Pellets in 100 μL of 1X Nuclei Buffer.

Nuclei counting using fluorescent cell counter

Timing: 30 min

Note: Counting and evaluating the right lysis condition of the nuclei will ensure a correct loading onto the 10× Genomics chips.

• 10.Gently pipette the samples for few seconds prior to counting to ensure a more accurate estimate of the nuclei concentration.
• 11.Proceed to count nuclei in at least 2–3 replicate counts. It will ensure higher accuracy.
• 12.Prepare counting slides and tubes.
• 13.Prepare the Neubauer chamber (or alternative counter slides).
• 14.Add 5 μL of nuclei suspension to 4 μL of PBS 0.04% BSA buffer in a 1.5 mL Eppendorf tube.
• 15.Add 1 μL of DAPI staining solution to tube, pipette to mix.
• 16.Load the 10 μL onto the hemacytometer.
• 17.Count the nuclei stained with DAPI under a fluorescence microscope to assess nuclei concentration.
• 18.Observe the hemacytometer under a bright-field and fluorescent microscope and adjust the focus per manufacturer’s recommendations.
• 19.Assess nuclei integrity using a 40x magnification objective.

Note: 40x or higher objective magnifications help to display the nuclear shape and ensure the integrity of the nuclear membrane.

• 20.Pipette gently to resuspend or filter the nuclei suspension to ensure having a preparation without clumps. Excess of debris might indicate an over lysis of the nuclei. Refer to Figure 1 for a proper nuclear lysis.

Note: Countess cell counter and Countess 3 automated cell counter (Invitrogen) can be used to estimate nuclei numbers. However, we advise to manually calculate the nuclei using the Neubauer chamber.

CRITICAL: Do not use trypan blue staining to count nuclei as it can lead to overestimated nuclei counts.

Note: After the isolation and the assessment of the nuclei concentration and viability, bring the input nuclei suspension in the Nuclei Buffer 1X at the optimal concentration following the 10× Genomics User Guide on nuclei recovery target.

• 21.The optimal input nuclei concentration for 10000 nuclei isolation using the Chromium 10× Genomics is 700–1,200 nuclei/μL. Refer to the relevant User Guide for specific information.

Note: If users perform sn multiome for both gene expression and chromatin accessibility from the same nuclei, they should use the 10× Genomics Chromium Next GEM Single Cell Multiome ATAC + Gene Expression protocol: https://cdn.10xgenomics.com/image/upload/v1666737555/support-documents/CG000338_ChromiumNextGEM_Multiome_ATAC_GEX_User_Guide_RevF.pdf

CRITICAL: Do not freeze or cryopreserve the nuclei suspension. The freezing and thawing process might damage the nuclear membrane.

Figure 1.

Quality assessment of Nuclei isolation

(A and B) Neubauer chamber images taken after loading 10 μL of diluted nuclei lysate on the hemocytometer. Bright-field picture (A) and fluorescent DAPI staining (B) are showing amount of nuclear lysate with relative amount of debris.

(C and D) Higher magnification of the bright field (C), which allows to check the integrity of the nuclear membrane, and of the relative DAPI staining (D).

snMultiome QC and library preparation

Timing: 2–3 day

Alternatives: A Bioanalyzer system could be used instead of the TapeStation.

Pause point: There is a designated stopping point after the library preparation (Figure 2). Libraries remain stable at −20°C for several months before sequencing.

Figure 2.

Examples of library pool QC traces

High Sensitivity D5000 ScreenTape QC analysis performed using the TapeStation Analysis Software 4.1.1.

(A and B) Shown are the representative traces of successful Bioanalyzer QC analyses for snRNAseq, at the cDNA amplification step (A), and snATACseq trace (B).

Sequencing quality control

Timing: 1–3 days

• 22.Use Cell Ranger ARC pipeline for multiome samples after sequencing to process the sequenced data.
• 23.Save metrics from the summary with CSV file through the Cell Ranger pipeline (summary.csv for multiome). Table 1 summarize the metrics for a correct QC and the ones obtain from Travisano et al. 2023.

snRNA-seq	Basic	Travisano et al.1
Estimated_Number_of_Cells	(500,1000)	10652
Mean_Reads_per_Cell	≥2,000	70026.34
Median_Genes_per_Cell	≥500	3032
Median_UMI_Counts_per_Cell	(1000, 1500)	6832
Reads_Mapped_Confidently_to_Intronic_Regions	(20%)	56.8%
Reads_Mapped_Confidently_to_Exonic_Regions	(25%)	26%
Q30 Bases in RNA Read	(>65%)	90.8%
Fraction of transcriptomic reads in cells	(>70%)	89.70%

The quality of the sequencing data is estimated for each dataset. We use the metrics from the CellRanger Summary output files as well as additional metrics (see Tables 1 and 2 below).

snATAC-seq	Basic	Travisano et al.1
Fraction of transposition events in peaks in cells	(0.05,0.2)	>0.2
Confidently mapped read pairs	(0.5, 0.75)	0.94
Q30 bases in read 1	(>65%)	88.10%
Q30 bases in read 2	(>65%)	91.30%
Median high-quality fragments per cell	≥500	21623
TSS enrichment score	(2,4)	8.6

Part 2: Single cell RNA-seq analysis

Timing: 1–2 h

CRITICAL: Before starting the analysis, confirm that you have a working R environment with all the required packages installed. Please refer to Seurat and Signac for the essential packages and related packages installation.

Note: Although possible, we have not launched RStudio analysis in a Docker.

Here are the essential steps to analyze the scRNA-seq datasets.

• 24.Prepare R libraries that will be used:

>library(sctransform)

>library(Seurat)

>library(SeuratData)

• 25.Load the datasets using the Seurat Package.

Note: Walkthroughs Seurat tutorials that help users to get started with the analysis are available at the: https://satijalab.org/seurat/articles/pbmc3k_tutorial.html or at https://satijalab.org/seurat/articles/sctransform_vignette

• 26.Read the datasets, create Seurat object, pre-process data, and select features for the analysis together with the normalization and stabilization of variance using the SCtransform package (using regularized negative binomial regression) are the first steps for the analysis.
• 27.Perform Principal Component Analysis (PCA) and Reduction of the dimensions by UMAP to establish dataset dimensionality that will allow Clustering and visualization of cells.

Note: The analysis of the snRNAseq performed using the SCtranform package will allow normalization and variance stabilization using regularized negative binomial regression.

Note: “barcodes.tsv.gz”, “features.tsv.gz”, and “matrix.mtx.gz” files should be placed in snRNA folder.

>snRNA.data <- Read10X(data.dir = "//Users/snRNA")

>snRNA <- CreateSeuratObject(counts = snRNA.data$`Gene Expression`, project = "snRNA", min.cells = 3, min.features = 200)

>snRNA

>snRNA [["percent.mt"]] <- PercentageFeatureSet(snRNA, pattern = "ˆMT-")

>snRNA <- SCTransform(snRNA, vars.to.regress = "percent.mt", verbose = FALSE)

>snRNA <- RunPCA(snRNA, verbose = FALSE)

>snRNA <- RunUMAP(snRNA, dims = 1:30, verbose = FALSE)

>snRNA <- FindNeighbors(snRNA, dims = 1:30, verbose = FALSE)

>snRNA <- FindClusters(snRNA, verbose = FALSE)

>DimPlot(snRNA, label = TRUE, repel= TRUE, reduction = “UMAP”) + NoLegend() >DefaultAssay(snRNA) <- "SCT"

• 28.Estimate the number of cells for each cluster and calculate the percentage of each cell type from the datasets by running the following code.

> cell.num <- table(Idents(MULTI))

• 29.Identify and label each cluster with specific markers using the FindAllMarkers function which identify markers for every cluster compared to all remaining cells.

>snRNAMarkers <- FindAllMarkers(snRNA)

>snRNAMarkers <- FindMarkers(object = snRNA, ident.1 = 14)

• 30.Prepare the library for the GO analysis.

Note: The GO analysis of specific population can be performed manually on the website (https://maayanlab.cloud/Enrichr/).

> library(enrichR)

>DEenrichRPlot(

object = snRNA,

ident.1 = c("14"),

balanced = TRUE,

logfc.threshold = 0.25,

assay = NULL,

max.genes=500,

test.use = "wilcox",

p.val.cutoff = 0.05,

cols = NULL,

enrich.database = "GO_Molecular_Function_2021",

num.pathway = 10)

> DEenrichRPlot(

object = snRNA,

ident.1 = c("14"),

balanced = TRUE,

logfc.threshold = 0.25,

assay = NULL,

max.genes=500,

test.use = "wilcox",

p.val.cutoff = 0.05,

cols = NULL,

enrich.database = "TF_Perturbations_Followed_by_Expression", num.pathway = 10)

Note: A network connection is required to connect the RStudio enrichR package to the specific enrich database (Refer to Figure 3 for the positive markers of the GO_Molecular_Function_2021).

Figure 3.

GO analysis of the Molecular Function of the cluster 14

Part 3: Single nuclei ATAC-seq analysis and integration with single cell RNA-seq

Timing: 3–4 h

To pre-process the snATAC-seq datasets, first prepare the required the R libraries.

• 31.Load datasets using Signac package and the Hsapiens genome reference:

>library(Signac)

>library(EnsDb.Hsapiens.v86)

>library(ggplot2)

>library(BSgenome.Hsapiens.UCSC.hg38)

Note: Walkthroughs Signac tutorials to help users to get started with the integration and analysis of the Multiome are available at the Stuart lab website: https://stuartlab.org/signac/articles/pbmc_multiomic. The authors followed the RStudio codes published from the vignette and used the Weighted Nearest Neighbor methods published in Seurat v4, to integrate and analyze the single-cell dataset measuring both DNA accessibility and gene expression.

>annotation <-

renameSeqlevels(annotation,mapSeqlevels(seqlevels(annotation),

"UCSC"))

>genome(annotation) <- "hg38"

Note: The author used the following RStudio script to get the annotation for the hg38.

• 32.Read the Gene Expression and Fragments respectively for RNA and ATAC assay and get gene annotations for hg38.
• 33.Pre-process data and select features for the analysis together with the normalization and stabilization of variance using the SCtransform package (using regularized negative binomial regression) are the first steps for the analysis.
• 34.Perform quality control metrics as DNA accessibility to assess low quality cells.

>MULTI <- NucleosomeSignal(MULTI)

>MULTI <- TSSEnrichment(MULTI)

> VlnPlot(object = MULTI,

features = c("nCount_RNA", "nCount_ATAC", "TSS.enrichment", "nucleosome_signal"),

ncol = 4, pt.size = 0)

• 35.Evaluate the QC metrics and set filters to remove low quality nuclei from the analysis (Refer to Figure 4).
• 36.Call peaks using MACS2 to identify peaks from all the cells of the datasets.

Note: It is possible to plot the UMAP of the snRNA-seq., the UMAP of the snATAC-seq and the UMAP obtained from their integration at this point as shown in the Signac tutorial webpage (https://stuartlab.org/signac/articles/pbmc_multiomic).

Optional: It is possible to use a reference analyzed datasets and transfer cell labels from the reference Seurat file at this point

• 37.Use the WNN (weighted-nearest neighbor) to perform integration of RNA+ATAC.
• 38.Use link peaks to genes function in order to identify potential cis-elements acting on the promoter of a certain gene.

Note: The authors did not label the clusters, which are identified by numbers. It is advised to assign a name for each cell cluster using the function “RenameIdents” and change the cluster number in idents.plot function with the corresponding assigned cell name.

> MULTI <- LinkPeaks( object = MULTI, peak.assay = "peaks", expression.assay = "SCT",

genes.use = c("PROX1", "VEGFC"))

>idents.plot <- c("0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15", "16", "17","18", "19", "20", "21", "22", "23", "24","25")

>p1 <- CoveragePlot( object = MULTI, region = "PROX1", features = "PROX1",

expression.assay = "SCT", idents = idents.plot,

extend.upstream = 10000, extend.downstream = 10000)

>p2 <- CoveragePlot( object = MULTI, region = "VEGFC", features = "VEGFC",

expression.assay = "SCT", idents = idents.plot,

extend.upstream = 10000, extend.downstream = 10000)

>p1 + p2

Note: To correctly compute the correlation between gene expression and the accessibility at nearby peaks, as shown in the Signac tutorial webpage (https://stuartlab.org/signac/articles/pbmc_multiomic) is strongly recommended to adjust the upstream and downstream regions to its locus.

• 39.Prepare the R libraries needed for the motif analysis.

>library(JASPAR2020)

>library(TFBSTools) >library(BSgenome.Hsapiens.UCSC.hg38) >library(patchwork)

• 40.Add motif information to the object:

>pfm <- getMatrixSet(x = JASPAR2020,opts = list(collection = "CORE", tax_group = 'vertebrates', all_versions = FALSE))

>MULTI <- AddMotifs(object = MULTI, genome = BSgenome.Hsapiens.UCSC.hg38,pfm = pfm)

• 41.Find DNA binding motifs overrepresented in a set of peaks from the cluster 14, cluster that has been identified as lymphatic endothelium.

> peaks.markers <- FindMarkers(object = MULTI, ident.1 = '14', only.pos = TRUE,

test.use = 'LR', min.pct = 0.05, latent.vars = 'nCount_peaks')

>top.da.peak <- rownames(peaks.markers[peaks.markers$p_val < 0.005, ])

>enriched.motifs <- FindMotifs(object = MULTI, features = top.da.peak)

>MotifPlot(object = MULTI, motifs = head(rownames(enriched.motifs)))

Note: The authors followed the instructions for the post-integration analysis of the overrepresented motifs in a set of differentially accessible peaks and of the Motif Footprint, respectively at the https://stuartlab.org/signac/articles/motif_vignette/ and at the https://stuartlab.org/signac/articles/footprint/ websites.

• 42.Load and run the ChromVAR R package⁷ to compute and visualize the per cell motif activities that are associated with variability in chromatin accessibility between cells.

>library(chromVAR)

> MULTI <- RunChromVAR(object = MULTI, genome = BSgenome.Hsapiens.UCSC.hg38) >DefaultAssay(MULTI) <- 'chromvar'

>FeaturePlot(object = MULTI, features = "MA0762.1", min.cutoff = 'q10', max.cutoff = 'q90', pt.size = 0.1)

• 43.Use the PlotFootprint function to plot the Motif footprinting and get positional information for our motif across the genome.

Note: The authors subset 3 clusters to ensure an easier visualization between different cell types.

>MULTI <- subset(MULTI, idents = c(‘14’,’1’,’7’))

>MULTI <- Footprint(object = MULTI, motif.name = "ETV2", genome = BSgenome.Hsapiens.UCSC.hg38)

>p2 <- PlotFootprint(MULTI, features = "ETV2")

>p2 + patchwork::plot_layout(ncol = 1)

• 44.Build a matrix of values which includes the normalized data from RNA or from the motifs in open chromatin.

Note: The authors highlighted the correlation of the PROX1 RNA levels and PROX1 Motif computed on all the clusters of the datasets in their averaged values. The correlation can be computed also on single cells or on a specific tissue/cell type using the subset function.

>PROX1rna <- AverageExpression(MULTI, assays = "SCT", features = "PROX1",group.by= "ident", add.ident = NULL, slot = "data", verbose = TRUE)

>PROX1 <- write.table(PROX1rna[["SCT"]], file='Rnaexp.tsv', quote=FALSE, sep='\t', col.names = TRUE)

>DefaultAssay(MULTI) <- 'chromvar'

>PROX1motif <- AverageExpression(MULTI, assays = "chromvar", features = "MA0794.1",group.by= "ident", add.ident = NULL, slot = "data", verbose = TRUE) >PROX1M <- write.table(PROX1motif[["chromvar"]], file='BindingMotif.tsv', quote=FALSE, sep='\t', col.names = TRUE)

>library(corrplot)

>library(RColorBrewer)

>Matrix <-cor(PROX1mat,method = "spearman") corrplot(Matrix, method = "color",type="upper", order="hclust", col=brewer.pal(n=8, name="RdYlBu"))

Figure 4.

Quality control metrics assessment of Multiomics data

Violin Plot showing per cell counts for either the RNA or ATAC assay and quality control metrics using the DNA accessibility before set filters (Top) and after filtering the low quality nuclei (Bottom).

Expected outcomes

The epicardial dissection protocol and the modified single nucleus isolation protocol are expected to provide enough intact nuclei to be processed for the generation of both RNA and ATAC libraries (Please refer to Figures 1A and 1B). The 10× Genomics kit-based library generation process is highly reproducible and provides good quality libraries (Please refer to Figures 2A and 2B) for sequencing, generating an exclusive resource that can be used to profile transcriptional and chromatin accessibility features of elusive cardiac cell populations. Figure 4 provides a violin plot visualization of the quality control metrics to apply before and after the removal of cells that are outliers for certain parameters (Refer to Step 34). Seurat and Signac tutorials provide all the instruction to obtain the UMAPs, the Pseudobulk of the genome accessibility or the TF footprint (Please refer to Figures 5, 6, and 7). Each cell-type/cluster is expected to be distinguished by specific markers using FeaturePlots function to visualize all known cell-type markers for identification and RenameIdents function to label each population.

Figure 5.

snRNA-seq, snATAC-seq and WNN UMAP clustering visualization

UMAPs generated from the snRNA-seq (Left), snATAC-seq (Middle) or snMultiome/ WNN integration (Right) assays.

Figure 6.

Visualization of genomic regions grouped by clusters

Pseudobulk of the genome accessibility for the PROX1 (left) and VEGFC (right) locus with the link peaks to gene function correlating open chromatin regions (peaks) with the expression of the nearby gene.

Figure 7.

Probability of Tn5 insertion across the genome centered around predicted ETV2 binding motif

The cluster 14 have higher enrichment compared with the mesothelial cells (groups 7 and 23).

Limitations

Part-one experimental

The authors recommend using the enlisted reagents or the alternatives suggested by 10× Genomics but not to use substitutions offered by other providers. The protocol was optimized for extremely low amount of tissue (epicardium from 2/3 biological replicates), and the main limitation of this procedure is the contamination of nuclei from the cortical cardiomyocytes. Another potential issue is over digestion of the samples resulting in the loss of nuclei.

Part-two analysis

The analyses made are based on the Seurat and Signac packages. Other packages such as Harmony and algorithms are available to the scientific community for integrative analysis. The authors recommend using the Seurat integration as linear-embedding model known to perform well for simple batch correction tasks and for integrating snRNA-seq + snATAC-seq datasets originating from the same cells.

Troubleshooting

Problem part-one/experimental

• •To decrease cardiomyocytes (CMs) contamination, the authors suggest practicing the dissection procedure before the processing of the samples with the kit few times.

Potential solution

The authors suggest that the epicardial preparation should not take more than 30 min to ensure high quality of the samples.

Problem part-one/experimental

• •Samples are over-digested, or the nuclei membrane look damaged (Under a 40x lens microscope) after the DAPI staining.

Potential solution

Prepare new buffers, lysis and wash and restart from the microdissection in Step 1.

Problem part-one/experimental

• •Incomplete digestion of the cells with consequent nuclei aggregation.

Potential solution

Evaluate the nuclei digestion by looking directly to the tube. Pipette tips have sticky property and tends to create cell aggregates on their inner surface. Check for cell aggregates at the pipette tips to ensure a proper digestion.

Problem part-two/analysis

• •The “library” command does not load R package, or any other command.

Potential solution

Check the installation status of each library and their dependencies and run the command to reinstall the missing R package.

>install.packages(“package_name”)

Problem part-two/analysis

• •The Uniform Manifold Approximation and Projection (UMAP) image does not match the one from the article (e.g., inverted x and y axis or different number of clusters).

Potential solution

The UMAP algorithm is a nonlinear dimensionality reduction method and the x/y axis of the plot generated might change across different computers/OS/R version used. The number of clusters might also change without affecting the analysis and can be adjusted changing the resolution (dims) used to run the RunUMAP. To generate the same clustering image the authors suggest to use the same R scripts, OS/Computer and to fix the seed in R.

>set.seed(123)).

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact and corresponding author, Dr. Stanislao Travisano (stravisano@chla.usc.edu).

Technical contact

Stanislao Igor Travisano (stravisano@chla.usc.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Raw RNA sequencing data have been deposited into the National Center for Biotechnology Information Gene Expression Omnibus: GSE241128. This paper does not report original code. Any additional information required to re-analyze the data reported in this paper is available from the lead contact upon request.