In fmicompbio/TabulaMurisSenisData: Bulk and single-cell RNA-seq data from the Tabula Muris Senis project

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(SingleCellExperiment)
library(TabulaMurisSenisData)
library(ggplot2)

Introduction

This package provides access to the processed bulk and single-cell RNA-seq data from the Tabula Muris Senis data set [@Schaum2019-nf;@Tabula_Muris_Consortium2020-um]. The processed bulk RNA-seq data was downloaded from GEO (accession number GSE132040). The single-cell data (droplet and FACS) was obtained from FigShare (see below for links). All data sets are provided as SingleCellExperiment objects for easy access and use within the Bioconductor ecosystem.

Bulk RNA-seq data

The bulk RNA-seq data can be accessed via the TabulaMurisSenisBulk() function. By setting the argument infoOnly to TRUE, we can get information about the size of the data set before downloading it.

tmp <- TabulaMurisSenisBulk(infoOnly = TRUE)
tms_bulk <- TabulaMurisSenisBulk()
tms_bulk

We list the available tissues.

table(colData(tms_bulk)$organ)

Single-cell RNA-seq data

Droplet data

The data files for the droplet single-cell RNA-seq data were downloaded from FigShare:

the full data (all tissues)
the data from individual tissues

We list the available tissues.

listTabulaMurisSenisTissues(dataset = "Droplet")

As for the bulk data, we can print the size of the data set before downloading it.

tmp <- TabulaMurisSenisDroplet(tissues = "All", infoOnly = TRUE)
tms_droplet <- TabulaMurisSenisDroplet(tissues = "All")
tms_droplet

We plot the UMAP of the entire data set and color by tissue, to re-create the plot from here.

# tissue colors
tissue_cols <- c(Pancreas = "#3182bd", Thymus = "#31a354", 
                 Trachea = "#636363", Bladder = "#637939", 
                 Lung = "#7b4173", Large_Intestine = "#843c39", 
                 Fat = "#969696", Tongue = "#a1d99b", 
                 Mammary_Gland = "#ce6dbd", Limb_Muscle = "#d6616b", 
                 Marrow = "#de9ed6", Skin = "#e6550d", 
                 Liver = "#e7969c", Heart_and_Aorta = "#e7ba52", 
                 Kidney = "#e7cb94", Spleen = "#fd8d3c")

# get dataset with all tissues
se <- tms_droplet$All
se 

# prepare data set for ggplot
ds <- as.data.frame(reducedDim(se, "UMAP"))
ds <- cbind(ds, tissue = colData(se)$tissue)
head(ds)

# plot
ggplot(ds, aes(x = UMAP1, y = UMAP2, color = tissue)) + 
  geom_point(size = 0.05) + 
  scale_color_manual(values = tissue_cols) + 
  theme_classic() + 
  guides(colour = guide_legend(override.aes = list(size = 5)))

FACS data

The data files for the FACS single-cell RNA-seq data were downloaded from FigShare:

the full data (all tissues)
the data from individual tissues

We list the available tissues.

listTabulaMurisSenisTissues(dataset = "FACS")

Also here, we can print the size of the data set before downloading it.

tmp <- TabulaMurisSenisFACS(tissues = "All", infoOnly = TRUE)
tms_facs <- TabulaMurisSenisFACS(tissues = "All")
tms_facs

We plot the UMAP of the entire data set and color by tissue, to re-create the plot from here.

# tissue colors
tissue_cols <- c(Skin = "#e6550d", Pancreas = "#3182bd", 
                 Limb_Muscle = "#d6616b", Heart = "#e7ba52", 
                 Spleen = "#fd8d3c", Diaphragm = "#8c6d31", 
                 Trachea = "#636363", Tongue = "#a1d99b", 
                 Thymus = "#31a354", `Brain_Non-Myeloid` = "#cedb9c", 
                 Brain_Myeloid = "#b5cf6b", Bladder = "#637939", 
                 Large_Intestine = "#843c39", BAT = "#9c9ede", 
                 GAT = "#bd9e39", MAT = "#a55194", SCAT = "#6baed6", 
                 Lung = "#7b4173", Liver = "#e7969c", 
                 Marrow = "#de9ed6", Kidney = "#e7cb94", 
                 Aorta = "#393b79", Mammary_Gland = "#ce6dbd")

# get dataset with all tissues
se <- tms_facs$All
se 

# prepare data set for ggplot
ds <- as.data.frame(reducedDim(se, "UMAP"))
ds <- cbind(ds, tissue = colData(se)$tissue)
head(ds)

# plot
ggplot(ds, aes(x = UMAP1, y = UMAP2, color = tissue)) + 
  geom_point(size = 0.05) + 
  scale_color_manual(values = tissue_cols) + 
  theme_classic() + 
  guides(colour = guide_legend(override.aes = list(size = 5)))

Explore data with `iSEE`

The Tabula Muris Senis datasets are provided in the form of SingleCellExperiment objects.
A natural companion to this data structure is the r Biocpkg("iSEE") package, which can be used for interactive and reproducible data exploration.

Any analysis steps should be performed in advance before calling iSEE, and since these datasets can be quite big, the operations can be time consuming, and/or require a considerable amount of resources.

sce_all_facs <- TabulaMurisSenisFACS(tissues = "All", processedCounts = TRUE)$All
sce_all_facs

We can see that the sce_all_facs contains both raw and processed counts (in the assays slots counts and logcounts), but also the PCA and UMAP embeddings as provided in the original publication.

assayNames(sce_all_facs)
reducedDimNames(sce_all_facs)

If desired, additional processing steps can be performed on the sce_all_facs - e.g., storing signature scores (computed via r Biocpkg("AUCell")) as colData elements.
Once these steps are completed, launching iSEE is as easy as

library(iSEE)

library(iSEE)
iSEE(sce_all_facs)

This will launch iSEE with a standard default set of panels.
Optionally, we can configure the initial set to start the app in the desired configuration - below, we show how to start iSEE with:

a ReducedDimensionPlot displaying the UMAP, colored by the expression of a feature;
a RowDataTable, for selecting a feature to be transmitted to the other panels (showing all genes whose name contains "Col1");
a FeatureAssayPlot, for showing the expression of the gene selected in the RowDataTable, split by tissue.

initial <- list()

################################################################################
# (Compact) Settings for Reduced dimension plot 1
################################################################################

initial[["ReducedDimensionPlot1"]] <- new(
    "ReducedDimensionPlot", 
    DataBoxOpen = TRUE, 
    Type = "UMAP", 
    VisualBoxOpen = TRUE, 
    ColorBy = "Feature name", 
    ColorByFeatureName = "Col1a1", 
    ColorByFeatureSource = "RowDataTable1", 
    ColorByFeatureDynamicSource = FALSE
)

################################################################################
# (Compact) Settings for Row data table 1
################################################################################

initial[["RowDataTable1"]] <- new(
    "RowDataTable", 
    Selected = "Col1a1", 
    Search = "Col1"
)

################################################################################
# (Compact) Settings for Feature assay plot 1
################################################################################

initial[["FeatureAssayPlot1"]] <- new(
    "FeatureAssayPlot", 
    DataBoxOpen = TRUE, 
    Assay = "logcounts", 
    XAxis = "Column data", 
    XAxisColumnData = "tissue", 
    YAxisFeatureName = "Col1a1", 
    YAxisFeatureSource = "RowDataTable1"
)

Click here to display the complete configuration chunk

initial <- list()

################################################################################
# Settings for Reduced dimension plot 1
################################################################################

initial[["ReducedDimensionPlot1"]] <- new("ReducedDimensionPlot", Type = "UMAP", XAxis = 1L, YAxis = 2L, 
    FacetRowByColData = "FACS.selection", FacetColumnByColData = "FACS.selection", 
    ColorByColumnData = "FACS.selection", ColorByFeatureNameAssay = "logcounts", 
    ColorBySampleNameColor = "#FF0000", ShapeByColumnData = "FACS.selection", 
    SizeByColumnData = "n_genes", FacetRowBy = "None", FacetColumnBy = "None", 
    ColorBy = "Feature name", ColorByDefaultColor = "#000000", 
    ColorByFeatureName = "Col1a1", ColorByFeatureSource = "RowDataTable1", 
    ColorByFeatureDynamicSource = FALSE, ColorBySampleName = "A10_B000497_B009023_S10.mm10-plus-0-0", 
    ColorBySampleSource = "---", ColorBySampleDynamicSource = FALSE, 
    ShapeBy = "None", SizeBy = "None", SelectionAlpha = 0.1, 
    ZoomData = numeric(0), BrushData = list(), VisualBoxOpen = TRUE, 
    VisualChoices = "Color", ContourAdd = FALSE, ContourColor = "#0000FF", 
    PointSize = 1, PointAlpha = 1, Downsample = FALSE, DownsampleResolution = 200, 
    CustomLabels = FALSE, CustomLabelsText = "A10_B000497_B009023_S10.mm10-plus-0-0", 
    FontSize = 1, LegendPointSize = 1, LegendPosition = "Bottom", 
    HoverInfo = TRUE, LabelCenters = FALSE, LabelCentersBy = "FACS.selection", 
    LabelCentersColor = "#000000", VersionInfo = list(iSEE = structure(list(
        c(2L, 5L, 1L)), class = c("package_version", "numeric_version"
    ))), PanelId = c(ReducedDimensionPlot = 1L), PanelHeight = 500L, 
    PanelWidth = 4L, SelectionBoxOpen = FALSE, RowSelectionSource = "---", 
    ColumnSelectionSource = "---", DataBoxOpen = FALSE, RowSelectionDynamicSource = FALSE, 
    ColumnSelectionDynamicSource = FALSE, RowSelectionRestrict = FALSE, 
    ColumnSelectionRestrict = FALSE, SelectionHistory = list())

################################################################################
# Settings for Row data table 1
################################################################################

initial[["RowDataTable1"]] <- new("RowDataTable", Selected = "Col1a1", Search = "Col1", SearchColumns = c("", 
"", "", "", ""), HiddenColumns = character(0), VersionInfo = list(
    iSEE = structure(list(c(2L, 5L, 1L)), class = c("package_version", 
    "numeric_version"))), PanelId = c(RowDataTable = 1L), PanelHeight = 500L, 
    PanelWidth = 4L, SelectionBoxOpen = FALSE, RowSelectionSource = "---", 
    ColumnSelectionSource = "---", DataBoxOpen = FALSE, RowSelectionDynamicSource = FALSE, 
    ColumnSelectionDynamicSource = FALSE, RowSelectionRestrict = FALSE, 
    ColumnSelectionRestrict = FALSE, SelectionHistory = list())

################################################################################
# Settings for Feature assay plot 1
################################################################################

initial[["FeatureAssayPlot1"]] <- new("FeatureAssayPlot", Assay = "logcounts", XAxis = "Column data", 
    XAxisColumnData = "tissue", XAxisFeatureName = "0610005C13Rik", 
    XAxisFeatureSource = "---", XAxisFeatureDynamicSource = FALSE, 
    YAxisFeatureName = "Col1a1", YAxisFeatureSource = "RowDataTable1", 
    YAxisFeatureDynamicSource = FALSE, FacetRowByColData = "FACS.selection", 
    FacetColumnByColData = "FACS.selection", ColorByColumnData = "age", 
    ColorByFeatureNameAssay = "logcounts", ColorBySampleNameColor = "#FF0000", 
    ShapeByColumnData = "FACS.selection", SizeByColumnData = "n_genes", 
    FacetRowBy = "None", FacetColumnBy = "None", ColorBy = "None", 
    ColorByDefaultColor = "#000000", ColorByFeatureName = "0610005C13Rik", 
    ColorByFeatureSource = "---", ColorByFeatureDynamicSource = FALSE, 
    ColorBySampleName = "A10_B000497_B009023_S10.mm10-plus-0-0", 
    ColorBySampleSource = "---", ColorBySampleDynamicSource = FALSE, 
    ShapeBy = "None", SizeBy = "None", SelectionAlpha = 0.1, 
    ZoomData = numeric(0), BrushData = list(), VisualBoxOpen = FALSE, 
    VisualChoices = "Color", ContourAdd = FALSE, ContourColor = "#0000FF", 
    PointSize = 1, PointAlpha = 1, Downsample = FALSE, DownsampleResolution = 200, 
    CustomLabels = FALSE, CustomLabelsText = "A10_B000497_B009023_S10.mm10-plus-0-0", 
    FontSize = 1, LegendPointSize = 1, LegendPosition = "Bottom", 
    HoverInfo = TRUE, LabelCenters = FALSE, LabelCentersBy = "FACS.selection", 
    LabelCentersColor = "#000000", VersionInfo = list(iSEE = structure(list(
        c(2L, 5L, 1L)), class = c("package_version", "numeric_version"
    ))), PanelId = c(FeatureAssayPlot = 1L), PanelHeight = 500L, 
    PanelWidth = 4L, SelectionBoxOpen = FALSE, RowSelectionSource = "---", 
    ColumnSelectionSource = "---", DataBoxOpen = FALSE, RowSelectionDynamicSource = FALSE, 
    ColumnSelectionDynamicSource = FALSE, RowSelectionRestrict = FALSE, 
    ColumnSelectionRestrict = FALSE, SelectionHistory = list())

initial

iSEE can then be launched with the following command:

iSEE(sce_all_facs, initial = initial)

knitr::include_graphics("ss_iSEE_facsdataset.jpg")

Note that these lengthy configuration options can be readily exported by clicking on the dedicated button "Display panel settings" in iSEE - we can do that anytime after we are done designing the configuration using the app interface.