spatialHeatmap: Visualizing Spatial Assays in Anatomical Images and Large-Scale Data Extensions
Last update: 30 August, 2023
Package
spatialHeatmap 2.6.2
1 Introduction
1.1 Motivation
The spatialHeatmap package provides functionalities for visualizing cell-, tissue- and organ-specific data of biological assays by coloring the corresponding spatial features defined in anatomical images according to quantitative abundance levels of measured biomolecules, such as transcripts, proteins or metabolites. A color key is used to represent the quantitative assay values and can be customized by the user. This core functionality of the package is called a spatial heatmap (SHM) plot. Additional important functionalities include Spatial Enrichment (SE), Spatial Co-Expression (SCE), and Single Cell to SHM Co-Visualization (SC2SHM-CoViz). These extra utilities are useful for identifying biomolecules with spatially selective abundance patterns (SE), groups of biomolecules with related abundance profiles using hierarchical clustering, K-means clustering, or network analysis (SCE), and to co-visualize single cell embedding results with SHMs (SCSHM-CoViz). The latter co-visualization functionality (Figure 1E) is described in a separate vignette called SCSHM-CoViz.
The functionalities of spatialHeatmap can be used either in a command-driven mode
from within R or a graphical user interface (GUI) provided by a Shiny App that
is part of this project. While the R-based mode provides flexibility to
customize and automate analysis routines, the Shiny App includes a variety of
convenience features that will appeal to experimentalists and users less
familiar with R. The Shiny App can be used on both local computers as
well as centralized server-based deployments (e.g. cloud-based or custom
servers). This way it can be used for both hosting community data on a
public server or running on a private system. The core functionalities of the
spatialHeatmap package are illustrated in Figure 1.
Figure 1: Overview of spatialHeatmap functionality
(A) The spatialHeatmap package plots numeric assay data onto spatially annotated images. The assay data can be provided as numeric vectors, tabular data, SummarizedExperiment, or SingleCellExperiment objects. The latter two are widely used data containers for organizing both assays as well as associated annotation data and experimental designs. (B) Anatomical and other spatial images need to be provided as annotated SVG (aSVG) files where the spatial features and the corresponding components of the assay data have matching labels (e.g. tissue labels). To work with SVG data efficiently, the SVG S4 class container has been developed. The assay data are used to color the matching spatial features in aSVG images according to a color key. (C)-(D) The result is called a spatial heatmap (SHM). (E) Large-scale data mining such as hierarchical clustering and network analysis can be integrated to facilitate the identification of biomolecules with similar abundance profiles. Moreover, (E) Single cell embedding results can be co-visualized with SHMs.
The package supports anatomical images from public repositories or those provided by users. In general any type of image can be used as long as it can be provided in SVG (Scalable Vector Graphics) format and the corresponding spatial features, such as organs, tissues, cellular compartments, are annotated (see aSVG below). The numeric values plotted onto an SHM are usually quantitative measurements from a wide range of profiling technologies, such as microarrays, next generation sequencing (e.g. RNA-Seq and scRNA-Seq), proteomics, metabolomics, or many other small- or large-scale experiments. For convenience, several preprocessing and normalization methods for the most common use cases are included that support raw and/or preprocessed data. Currently, the main application domains of the spatialHeatmap package are numeric data sets and spatially mapped images from biological, agricultural and biomedical areas. Moreover, the package has been designed to also work with many other spatial data types, such as population data plotted onto geographic maps. This high level of flexibility is one of the unique features of spatialHeatmap. Related software tools for biological applications in this field are largely based on pure web applications (Maag 2018; Lekschas et al. 2015; Papatheodorou et al. 2018; Winter et al. 2007; Waese et al. 2017) or local tools (Muschelli, Sweeney, and Crainiceanu 2014) that typically lack customization functionalities. These restrictions limit users to utilizing pre-existing expression data and/or fixed sets of anatomical image collections. To close this gap for biological use cases, we have developed spatialHeatmap as a generic R/Bioconductor package for plotting quantitative values onto any type of spatially mapped images in a programmable environment and/or in an intuitive to use GUI application.
1.2 Design
The core feature of spatialHeatmap is to map assay values (e.g.
gene expression data) of one or many biomolecules (e.g. genes) measured under
different conditions in form of numerically graded colors onto the
corresponding cell types or tissues represented in a chosen SVG image. In the
gene profiling field, this feature supports comparisons of the expression
values among multiple genes by plotting their SHMs next to each
other. Similarly, one can display the expression values of a single or multiple
genes across multiple conditions in the same plot (Figure 3). This level of flexibility is
very efficient for visualizing complicated expression patterns across genes,
cell types and conditions. In case of more complex anatomical images with
overlapping multiple layer tissues, it is important to visually expose the
tissue layer of interest in the plots. To address this, several default and
customizable layer viewing options are provided. They allow to hide features in
the top layers by making them transparent in order to expose features below
them. This transparency viewing feature is highlighted below in the mouse
example (Figure 4). Moreover, one can plot multiple distinct
aSVGs in a single SHM plot as shown in Figure 10. This is
particularly useful for displaying abundance trends across multiple development
stages, where each is represented by its own aSVG image. In addition to
static SHM representations, one can visualize them in form of interactive HTML files or videos.
To maximize reusability and extensibility, the package organizes large-scale
omics assay data along with the associated experimental design information in a
SummarizedExperiment object (Figure 1A; Morgan et al. 2018). The latter is one of the core S4 classes within
the Bioconductor ecosystem that has been widely adapted by many other software
packages dealing with gene-, protein- and metabolite-level profiling data.
In case of gene expression data, the assays slot of
the SummarizedExperiment container is populated with a gene expression
matrix, where the rows and columns represent the genes and tissue/conditions,
respectively. The colData slot contains experimental design definitions including
replicate and treatment information. The tissues and/or cell type information in the object maps via
colData to the corresponding features in the SVG images using unique
identifiers for the spatial features (e.g. tissues or cell types). This
allows to color the features of interest in an SVG image according to the
numeric data stored in a SummarizedExperiment object. For simplicity the
numeric data can also be provided as numeric vectors or data.frames. This
can be useful for testing purposes and/or the usage of simple data sets that
may not require the more advanced features of the SummarizedExperiment class,
such as measurements with only one or a few data points. The details about how to
access the SVG images and properly format the associated expression data are
provided in the Supplementary Section of this vignette.
1.3 Image Format: SVG
SHMs are images where colors encode numeric values in features of
any shape. For plotting SHMs, Scalable Vector Graphics (SVG) has
been chosen as image format since it is a flexible and widely adapted vector
graphics format that provides many advantages for computationally embedding
numerical and other information in images. SVG is based on XML formatted text
describing all components present in images, including lines, shapes and
colors. In case of biological images suitable for SHMs, the shapes
often represent anatomical or cell structures. To assign colors to specific
features in SHMs, annotated SVG (aSVG) files are used where the
shapes of interest are labeled according to certain conventions so that they
can be addressed and colored programmatically. One or multiple aSVG files can be parsed and stored in the SVG S4 container with utilities provided by the spatialHeatmap package (Figure 1B). The main slots of SVG include coordinate,
attribute, dimension, svg, and raster. They correspond to feature coordinates, styling attributes (color, line width, etc.), width and heigth, original aSVG instances,
and raster image paths, respectively. Raster images are required only when including photographic image components in SHMs (Figure 7), which is optional.
SVGs and aSVGs of anatomical structures can be downloaded from many sources including the repositories described below. Alternatively, users can generate them themselves with vector graphics software such as Inkscape. Typically, in aSVGs one or more shapes of a feature of interest, such as the cell shapes of an organ, are grouped together by a common feature identifier. Via these group identifiers one or many feature types can be colored simultaneously in an aSVG according to biological experiments assaying the corresponding feature types with the required spatial resolution. To color spatial features according to numeric assay values, common identifiers are required for spatial features between the assay data and aSVGs. The color gradient used to visually represent the numeric assay values is controlled by a color gradient parameter. To visually interpret the meaning of the colors, the corresponding color key is included in the SHM plots. Additional details for properly formatting and annotating both aSVG images and assay data are provided in the Supplementary Section section of this vignette.
1.4 Data Repositories
If not generated by the user, SHMs can be generated with data downloaded from
various public repositories. This includes gene, protein and metabolic
profiling data from databases, such as GEO,
BAR and Expression
Atlas from EMBL-EBI (Papatheodorou et al. 2018). A
particularly useful resource, when working with spatialHeatmap, is the EBI
Expression Atlas. This online service contains both assay and anatomical
images. Its assay data include mRNA and protein profiling experiments for
different species, tissues and conditions. The corresponding anatomical image
collections are also provided for a wide range of species including animals and
plants. In spatialHeatmap several import functions are provided to work with
the expression and aSVG repository from the Expression Atlas
directly. The aSVG images developed by the spatialHeatmap project are
available in its own repository called spatialHeatmap aSVG
Repository,
where users can contribute their aSVG images that are formatted according to
our guidlines.
1.5 Tutorial Overview
The following sections of this vignette showcase the most important
functionalities of the spatialHeatmap package using as initial example a simple
to understand testing data set, and then more complex mRNA profiling data from the
Expression Atlas and GEO databases. The co-visualization functionality is explained in a separate vignette (see SCSHM-CoViz).
First, SHM plots are generated for both the testing
and mRNA expression data. The latter include gene expression data sets from
RNA-Seq and microarray experiments of Human Brain, Mouse
Organs, Chicken Organs, and Arabidopsis Shoots. The
first three are RNA-Seq data from the Expression
Atlas, while the last one is a microarray data
set from GEO. Second, gene context
analysis tools are introduced, which facilitate the visualization of
gene modules sharing similar expression patterns. This includes the
visualization of hierarchical clustering results with traditional matrix
heatmaps (Matrix Heatmap) as well as co-expression network plots
(Network). Third, the Spatial Enrichment functionality is illustrated
with mouse RNA-seq data. Lastly, an overview of the corresponding Shiny App
is presented that provides access to the same functionalities as the R
functions, but executes them in an interactive GUI environment (Chang et al. 2021; Chang and Borges Ribeiro 2018).
2 Getting Started
2.1 Installation
The spatialHeatmap package should be installed from an R (version \(\ge\) 3.6)
session with the BiocManager::install command.
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("spatialHeatmap")2.2 Packages and Documentation
Next, the packages required for running the sample code in this vignette need to be loaded.
library(spatialHeatmap); library(SummarizedExperiment); library(ExpressionAtlas); library(GEOquery); library(igraph); library(BiocParallel); library(kableExtra); library(org.Hs.eg.db); library(org.Mm.eg.db); library(ggplot2)The following lists the vignette(s) of this package in an HTML browser. Clicking the corresponding name will open this vignette.
browseVignettes('spatialHeatmap')To reduce runtime, intermediate results can be cached under ~/.cache/shm.
cache.pa <- '~/.cache/shm' # Path of the cache directory.A temporary directory is created to save output files.
tmp.dir <- normalizePath(tempdir(check=TRUE), winslash="/", mustWork=FALSE)3 Spatial Heatmaps
3.1 Quick Start
Spatial Heatmaps (SHMs) are plotted with the shm function. To provide a quick
and intuitive overview how these plots are generated, the following uses a
generalized tesing example where a small vector of random numeric values is
generated that are used to color features in an aSVG image. The image chosen
for this example is an aSVG depicting the human brain. The corresponding image
file homo_sapiens.brain.svg is included in this package for testing purposes.
3.1.1 aSVG Image
After the full path to the chosen target aSVG image on a user’s system is
obtained with the system.file function, the function read_svg is used to import the aSVG information relevant for creating SHMs, which is stored in an SVG object
svg.hum.
svg.hum.pa <- system.file("extdata/shinyApp/data", 'homo_sapiens.brain.svg', package="spatialHeatmap")
svg.hum <- read_svg(svg.hum.pa)All features and their attributes can be accessed with the attribute
function, where fill and stroke are the two most important ones providing
color and line width information, respectively. The feature column includes group labels for sub-features in the sub.feature column. SHM plots are created by mapping assay data to labels in feature.
feature.hum <- attribute(svg.hum)[[1]]
tail(feature.hum[, 1:6], 3) # Partial features and respective attributes
## # A tibble: 3 × 6
## feature id fill stroke sub.feature index
## <chr> <chr> <chr> <dbl> <chr> <int>
## 1 cerebellum UBERON_0002037 none 0.016 cerebellum_2_2 218
## 2 nucleus.accumbens UBERON_0001882 none 0.016 nucleus.accumbens 248
## 3 telencephalic.ventricle UBERON_0002285 none 0.016 telencephalic.ventr… 249Feature coordinates can be accessed with the coordinate function.
coord.df <- coordinate(svg.hum)[[1]]
tail(coord.df, 3) # Partial features and respective coordinates
## # A tibble: 3 × 4
## x y feature index
## <dbl> <dbl> <chr> <int>
## 1 193. 321. telencephalic.ventricle 249
## 2 193. 321. telencephalic.ventricle 249
## 3 193. 321. telencephalic.ventricle 2493.1.2 Numeric Data
The following generates a small named vector for testing,
where the data slot contains four numbers, and the name slot is populated with
three feature names and one missing one (here ’notMapped"). The numbers
are mapped to features (feature.hum) via matching names among the numeric vector and the aSVG,
respectively. Accordingly, only numbers and features with matching name
counterparts can be colored in the aSVG image. In addition, a summary of the numeric assay to feature mappings is stored
in a data.frame returned by the shm function (see below).
set.seed(20) # To obtain reproducible results, a fixed seed is set.
unique(feature.hum$feature)[1:10]
## [1] "g4320" "locus.ceruleus" "diencephalon"
## [4] "medulla.oblongata" "middle.temporal.gyrus" "caudate.nucleus"
## [7] "middle.frontal.gyrus" "occipital.lobe" "parietal.lobe"
## [10] "pineal.gland"
my_vec <- setNames(sample(1:100, 4), c('substantia.nigra', 'putamen', 'prefrontal.cortex', 'notMapped'))
my_vec
## substantia.nigra putamen prefrontal.cortex notMapped
## 38 63 2 293.1.3 Plot SHM
Before plotting SHMs, the aSVG instance and numeric data are stored in an SPHM object for the sake of efficient data management and reusability.
dat.quick <- SPHM(svg=svg.hum, bulk=my_vec)Next, the SHM is plotted with the shm function (Figure
2). Internally, the numbers in my_vec are translated into
colors based on the color key and then
painted onto the corresponding features in the aSVG. In the given example
(Figure 2) only three features (‘substantia.nigra’, ‘putamen’, and ‘prefrontal.cortex’) in the aSVG have matching entries in the data my_vec.
shm.res <- shm(data=dat.quick, ID='testing', ncol=1, sub.title.size=20, legend.nrow=3)
Figure 2: SHM of human brain with testing data
The plots from the left to the right represent the color key for the numeric data, followed by four SHM plots and the legend of the spatial features. The numeric values provided in my_vec are used to color the corresponding features in the SHM plots according to the color key while the legend plot identifies the spatial regions.
The named numeric values in my_vec, that have name matches with the features in the
chosen aSVG, are stored in the mapped_feature slot.
# Mapped features
spatialHeatmap::output(shm.res)$mapped_feature
## ID feature value fill SVG
## 1 testing substantia.nigra 38 #FF8700 homo_sapiens.brain.svg
## 2 testing putamen 63 #FF0000 homo_sapiens.brain.svg
## 3 testing prefrontal.cortex 2 #FFFF00 homo_sapiens.brain.svg3.2 Human Brain
This subsection introduces how to query and download cell- and tissue-specific assay data in
the Expression Atlas database using the ExpressionAtlas package (Keays 2019). After the choosen data is downloaded directly into a user's R session, the
expression values for selected genes can be plotted onto a chosen aSVG image with
or without prior preprocessing steps (e.g. normalization).
3.2.1 Gene Expression Data
The following example searches the Expression Atlas for expression data derived from specific tissues and species of interest, here ‘cerebellum’ and ‘Homo sapiens’, respectively.
all.hum <- read_cache(cache.pa, 'all.hum') # Retrieve data from cache.
if (is.null(all.hum)) { # Save downloaded data to cache if it is not cached.
all.hum <- searchAtlasExperiments(properties="cerebellum", species="Homo sapiens")
save_cache(dir=cache.pa, overwrite=TRUE, all.hum)
}The search result contains 15 accessions. In the following code, the accession ‘E-GEOD-67196’ from Prudencio et al. (2015) has been chosen, which corresponds to an RNA-Seq profiling experiment of ‘cerebellum’ and ‘frontal cortex’ brain tissue from patients with amyotrophic lateral sclerosis (ALS).
all.hum[2, ]## DataFrame with 1 row and 4 columns
## Accession Species Type Title
## <character> <character> <character> <character>
## 1 E-GEOD-75852 Homo sapiens RNA-seq of coding RNA Human iPSC-Derived C..The getAtlasData function allows to download the chosen RNA-Seq experiment
from the Expression Atlas and import it into a RangedSummarizedExperiment
object.
rse.hum <- read_cache(cache.pa, 'rse.hum') # Read data from cache.
if (is.null(rse.hum)) { # Save downloaded data to cache if it is not cached.
rse.hum <- getAtlasData('E-GEOD-67196')[[1]][[1]]
save_cache(dir=cache.pa, overwrite=TRUE, rse.hum)
}The design of the downloaded RNA-Seq experiment is described in the colData slot of
rse.hum. The following returns only its first four rows and columns.
colData(rse.hum)[1:2, c(2, 4)]
## DataFrame with 2 rows and 2 columns
## organism organism_part
## <character> <character>
## SRR1927019 Homo sapiens cerebellum
## SRR1927020 Homo sapiens frontal cortex3.2.2 aSVG Image
The following example shows how to download from the above described SVG
repositories an aSVG image that matches the tissues and species
assayed in the gene expression data set downloaded above.
The return_feature function queries the repository for feature- and
species-related keywords, here c('frontal cortex', 'cerebellum') and c('homo sapiens', 'brain'), respectively.
The remote data are downloaded before calling return_feature.
# Remote aSVG repos.
data(aSVG.remote.repo)
tmp.dir.ebi <- file.path(tmp.dir, 'ebi.zip')
tmp.dir.shm <- file.path(tmp.dir, 'shm.zip')
# Download the remote aSVG repos as zip files.
download.file(aSVG.remote.repo$ebi, tmp.dir.ebi)
download.file(aSVG.remote.repo$shm, tmp.dir.shm)
remote <- list(tmp.dir.ebi, tmp.dir.shm)The downloaded aSVG repos are queried and returned aSVG files are saved in an empty directory (tmp.dir) to avoid overwriting of existing SVG files.
tmp.dir <- file.path(tempdir(check=TRUE), 'shm') # Empty directory. feature.df.hum <- return_feature(feature=c('frontal cortex', 'cerebellum'), species=c('homo sapiens', 'brain'), dir=tmp.dir, remote=remote) # Query aSVGs
feature.df.hum[1:8, ] # Return first 8 rows for checking
unique(feature.df.hum$SVG) # Return all matching aSVGsTo build this vignettes according to Bioconductor’s package requirements, the
following code section uses aSVG file instances included in the
spatialHeatmap package rather than the downloaded instances above.
svg.dir <- system.file("extdata/shinyApp/data", package="spatialHeatmap") # Directory of the aSVG collection in spatialHeatmap
feature.df.hum <- return_feature(feature=c('frontal cortex', 'cerebellum'), species=c('homo sapiens', 'brain'), keywords.any=TRUE, return.all=FALSE, dir=svg.dir, remote=NULL)Note, the target tissues frontal cortex and cerebellum are included in both
the experimental design slot of the downloaded expression data as well as the
annotations of the aSVG. This way these features can be colored in the downstream
SHM plots. If necessary users can also change from within R the feature identifiers in an aSVG (see Supplementary Section).
tail(feature.df.hum[, c('feature', 'stroke', 'SVG')], 3)
## # A tibble: 3 × 3
## feature stroke SVG
## <chr> <dbl> <chr>
## 1 cerebral.cortex 0.1 mus_musculus.brain_sp.svg
## 2 cerebellum 0.1 mus_musculus.brain_sp.svg
## 3 somatosensor.cortex 0.100 mus_musculus.brain_sp.svgAmong the returned aSVG files, homo_sapiens.brain.svg is chosen for creating SHMs. Since it is the same as the Quick Start, the aSVG stored in svg.hum is used in the downstream steps.
3.2.3 Experimental Design
To display ‘pretty’ sample names in columns and legends of downstream tables and plots respectively, the following example imports a ‘targets’ file that can be customized by users in a text program. The targets file content is used to replace the text in the
colData slot of the RangedSummarizedExperiment object with a version containing
shorter sample names for plotting purposes.
The custom targets file is imported and then loaded into colData slot of rse.hum. A slice of the simplified colData object is shown below.
hum.tar <- system.file('extdata/shinyApp/data/target_human.txt', package='spatialHeatmap')
target.hum <- read.table(hum.tar, header=TRUE, row.names=1, sep='\t') # Importing
colData(rse.hum) <- DataFrame(target.hum) # Loading to "colData"
colData(rse.hum)[c(1:2, 41:42), 4:5]
## DataFrame with 4 rows and 2 columns
## organism_part disease
## <character> <character>
## SRR1927019 cerebellum ALS
## SRR1927020 frontal cortex ALS
## SRR1927059 cerebellum normal
## SRR1927060 frontal cortex normal3.2.4 Preprocess Assay Data
The raw count gene expression data is stored
in the assay slot of rse.hum. The following shows how to apply basic preprocessing routines on the count data, such as normalizing, aggregating replicates, and removing genes with unreliable expression responses, which are optional for plotting SHMs.
The norm_data function is developed to normalize RNA-seq raw count data. The following example uses the ESF normalization option due to its good time performance, which is estimateSizeFactors from DESeq2 (Love, Huber, and Anders 2014).
se.nor.hum <- norm_data(data=rse.hum, norm.fun='ESF', log2.trans=TRUE)Replicates are aggregated with the summary
statistics chosen under the aggr argument (e.g. aggr='mean'). The
columns specifying replicates can be assigned to the sam.factor and
con.factor arguments corresponding to samples and conditions, respectively.
For tracking, the corresponding sample/condition labels are used as column
titles in the aggregated assay instance, where they are concatenated with a
double underscore as separator (Table 1).
se.aggr.hum <- aggr_rep(data=se.nor.hum, sam.factor='organism_part', con.factor='disease', aggr='mean')
assay(se.aggr.hum)[c(120, 49939, 49977), ]| cerebellum__ALS | frontal.cortex__ALS | cerebellum__normal | frontal.cortex__normal | |
|---|---|---|---|---|
| ENSG00000006047 | 1.134172 | 5.2629629 | 0.5377534 | 5.3588310 |
| ENSG00000268433 | 5.324064 | 0.3419665 | 3.4780744 | 0.1340332 |
| ENSG00000268555 | 5.954572 | 2.6148548 | 4.9349736 | 2.0351776 |
The filtering example below retains genes with expression values
larger than 5 (log2 space) in at least 1% of all samples (pOA=c(0.01, 5)), and
a coefficient of variance (CV) between 0.30 and 100 (CV=c(0.30, 100)). After that, the Ensembl gene ids are converted to UniProt ids with the function cvt_id.
se.fil.hum <- filter_data(data=se.aggr.hum, sam.factor='organism_part', con.factor='disease', pOA=c(0.01, 5), CV=c(0.3, 100))
se.fil.hum <- cvt_id(db='org.Hs.eg.db', data=se.fil.hum, from.id='ENSEMBL', to.id='SYMBOL')3.2.5 SHM: Multiple Genes
Spatial features of interest can be subsetted with the function sub_sf by assigning their indexes (see below) to the argument show. In the following, ‘brain outline’, ‘prefrontal.cortex’, ‘frontal.cortex’, and ‘cerebellum’ are subsetted.
Next, for efficient data management and reusability the subset aSVG and assay data are stored in an SPHM object.
# Subsetting aSVG features.
svg.hum.sub <- sub_sf(svg.hum, show=c(64:132, 162:163, 164, 190:218))
tail(attribute(svg.hum.sub)[[1]][, 1:6], 3)
## # A tibble: 3 × 6
## feature id fill stroke sub.feature index
## <chr> <chr> <chr> <dbl> <chr> <int>
## 1 cerebellum UBERON_0002037 none 0.016 cerebellum_2_4 216
## 2 cerebellum UBERON_0002037 none 0.016 cerebellum_2_3 217
## 3 cerebellum UBERON_0002037 none 0.016 cerebellum_2_2 218
# Storing assay data and subsetted aSVG in an 'SPHM' object.
dat.hum <- SPHM(svg=svg.hum.sub, bulk=se.fil.hum)SHMs for multiple genes can be plotted by providing the
corresponding gene IDs under the ID argument as a character vector. The
shm function will then sequentially arrange the SHMs for
each gene in a single composite plot. To facilitate comparisons among expression
values across genes and/or conditions, the lay.shm parameter can be assigned
gene or con, respectively. For instance, in Figure 3 the
SHMs of the genes 'OLFM4 and LOC440742 are organized
by condition in a horizontal view. This functionality is particularly useful when comparing associated genes such as gene families.
res.hum <- shm(data=dat.hum, ID=c('OLFM4', 'LOC440742'), lay.shm='con', legend.r=1.5, legend.nrow=3, h=0.6)
Figure 3: SHMs of two genes
The subplots are organized by “condition”. Only cerebellum and frontal cortex are colored, because they are present in both the aSVG and the expression data.
In the above example, the normalized expression values of chosen genes are used to color the frontal cortex and cerebellum, where the different conditions, here normal and ALS, are given in separate SHMs. The color and feature mappings are defined by the corresponding color key and legend plot on the left and right, respectively.
By default, spatial features in assay data are mapped to their counterparts in aSVG according to the same identifiers on a one-to-one basis. However, the mapping can be customized, such as mapping a spatial feature in the data to a different or multiple counterparts in the aSVG. This advanced functionality is demonstrated in the Supplementary Section.
3.2.6 SHM: Other Selected Features
SHMs can be saved to interactive HTML files as well as video files. Each HTML file
contains an interactive SHM with zooming in and out functionality. Hovering over
graphics features will display data, gene, condition and other information. The
video will play the SHM subplots in the order specified under the lay.shm
argument.
The following example saves the interactive HTML and video files under the directory tmp.dir.
shm(data=dat.hum, ID=c('OLFM4', 'LOC440742'), lay.shm='con', legend.r=1.5, legend.nrow=3, h=0.6, aspr=2.3, animation.scale=0.7, bar.width=0.1, bar.value.size=4, out.dir=tmp.dir)The following code saves individual SHMs into the same SVG file shm_hum.svg with the color scale and legend plot included.
res <- shm(data=dat.hum, ID=c('OLFM4', 'LOC440742'), lay.shm='con', legend.r=1.5, legend.nrow=3, h=0.5, aspr=2.3, animation.scale=0.7, bar.width=0.08, bar.value.size=12)
ggsave(file="./shm_hum.svg", plot=output(res)$spatial_heatmap, width=10, height=8) The following code exports each SHM (associated with a specific gene and condition) as separate SVG files in tmp.dir. In contrast to the original aSVG file, spatial features in the output SVG files are assigned heat colors.
write_svg(input=res, out.dir=tmp.dir)A meta function plot_meta is developed as a wraper of individual steps necessary for plotting SHMs. The benefit of this function is creating SHMs with the Linux command line as shown below.
Rscript -e "spatialHeatmap::plot_meta(svg.path=system.file('extdata/shinyApp/data', 'mus_musculus.brain.svg', package='spatialHeatmap'), bulk=system.file('extdata/shinyApp/data', 'bulk_mouse_cocluster.rds', package='spatialHeatmap'), sam.factor='tissue', aggr='mean', ID=c('AI593442', 'Adora1'), ncol=1, bar.width=0.1, legend.nrow=5, h=0.6)"3.2.7 SHM: Customization
To provide a high level of flexibility, many arguments are developed for shm.
An overview of important arguments and their utility is provided in Table 2.
| argument | description | |
|---|---|---|
| 1 | data | An SPHM object containing assay data and aSVG (s). |
| 2 | sam.factor | Applies to SummarizedExperiment (SE). Column name of sample replicates in colData slot. Default is NULL |
| 3 | con.factor | Applies to SE. Column name of condition replicates in colData slot. Default is NULL |
| 4 | ID | A character vector of row items for plotting spatial heatmaps |
| 5 | col.com | A character vector of color components for building colour scale. Default is c(‘yellow’, ‘orange’,‘red’) |
| 6 | col.bar | ‘selected’ or ‘all’, the former means use values of ID to build the colour scale while the latter use all values in data. Default is ‘selected’. |
| 7 | bar.width | A numeric of colour bar width. Default is 0.7 |
| 8 | trans.scale | One of ‘log2’, ‘exp2’, ‘row’, ‘column’, or NULL, which means transform the data by ‘log2’ or ‘2-base expoent’, scale by ‘row’ or ‘column’, or no manipuation respectively. |
| 9 | ft.trans | A vector of aSVG features to be transparent. Default is NULL. |
| 10 | legend.r | A numeric to adjust the dimension of the legend plot. Default is 1. The larger, the higher ratio of width to height. |
| 11 | sub.title.size | The title size of each spatial heatmap subplot. Default is 11. |
| 12 | lay.shm | ‘gen’ or ‘con’, applies to multiple genes or conditions respectively. ‘gen’ means spatial heatmaps are organised by genes while ‘con’ organised by conditions. Default is ‘gen’ |
| 13 | ncol | The total column number of spatial heatmaps, not including legend plot. Default is 2. |
| 14 | ft.legend | ‘identical’, ‘all’, or a vector of samples/features in aSVG to show in legend plot. ‘identical’ only shows matching features while ‘all’ shows all features. |
| 15 | legend.ncol, legend.nrow | Two numbers of columns and rows of legend keys respectively. Default is NULL, NULL, since they are automatically set. |
| 16 | legend.position | the position of legend keys (‘none’, ‘left’, ‘right’,‘bottom’, ‘top’), or two-element numeric vector. Default is ‘bottom’. |
| 17 | legend.key.size, legend.text.size | The size of legend keys and labels respectively. Default is 0.5 and 8 respectively. |
| 18 | line.size, line.color | The size and colour of all plogyon outlines respectively. Default is 0.2 and ‘grey70’ respectively. |
| 19 | verbose | TRUE or FALSE. Default is TRUE and the aSVG features not mapped are printed to R console. |
| 20 | out.dir | The directory to save HTML and video files of spatial heatmaps. Default is NULL. |
3.3 Mouse Organs
This section generates an SHM plot for mouse data from the Expression Atlas. The code components are very similar to the previous Human Brain example. For brevity, the corresponding text explaining the code has been reduced to a minimum.
3.3.1 Gene Expression Data
The chosen mouse RNA-Seq data compares tissue level gene expression across mammalian species (Merkin et al. 2012). The following searches the Expression Atlas for expression data from ‘kidney’ and ‘Mus musculus’.
all.mus <- read_cache(cache.pa, 'all.mus') # Retrieve data from cache.
if (is.null(all.mus)) { # Save downloaded data to cache if it is not cached.
all.mus <- searchAtlasExperiments(properties="kidney", species="Mus musculus")
save_cache(dir=cache.pa, overwrite=TRUE, all.mus)
}Among the many matching entries, accession ‘E-MTAB-2801’ will be downloaded.
all.mus[7, ]
rse.mus <- read_cache(cache.pa, 'rse.mus') # Read data from cache.
if (is.null(rse.mus)) { # Save downloaded data to cache if it is not cached.
rse.mus <- getAtlasData('E-MTAB-2801')[[1]][[1]]
save_cache(dir=cache.pa, overwrite=TRUE, rse.mus)
}The design of the downloaded RNA-Seq experiment is described in the colData slot of
rse.mus. The following returns only its first three rows.
colData(rse.mus)[1:3, ]
## DataFrame with 3 rows and 4 columns
## AtlasAssayGroup organism organism_part strain
## <character> <character> <character> <character>
## SRR594393 g7 Mus musculus brain DBA/2J
## SRR594394 g21 Mus musculus colon DBA/2J
## SRR594395 g13 Mus musculus heart DBA/2J3.3.2 aSVG Image
The following example shows how to retrieve from the remote SVG
repositories an aSVG image that matches the tissues and species
assayed in the downloaded data above. The sample data from Human Brain are used such as remote.
feature.df.mus <- return_feature(feature=c('heart', 'kidney'), species=c('Mus musculus'), dir=tmp.dir, remote=remote)To meet the R/Bioconductor package requirements, the following uses aSVG file instances included in the
spatialHeatmap package rather than the downloaded instances.
feature.df.mus <- return_feature(feature=c('heart', 'kidney'), species=NULL, dir=svg.dir, remote=NULL) Return the names of the matching aSVG files.
unique(feature.df.mus$SVG)
## [1] "gallus_gallus.svg" "mus_musculus.male.svg"The mus_musculus.male.svg instance is selected and imported.
svg.mus.pa <- system.file("extdata/shinyApp/data", "mus_musculus.male.svg", package="spatialHeatmap")
svg.mus <- read_svg(svg.mus.pa)3.3.3 Experimental Design
A sample target file that is included in this package is imported and then loaded to the colData slot of rse.mus. To inspect its content, the first three rows are shown.
mus.tar <- system.file('extdata/shinyApp/data/target_mouse.txt', package='spatialHeatmap')
target.mus <- read.table(mus.tar, header=TRUE, row.names=1, sep='\t') # Importing
colData(rse.mus) <- DataFrame(target.mus) # Loading
target.mus[1:3, ]
## AtlasAssayGroup organism organism_part strain
## SRR594393 g7 Mus musculus brain DBA.2J
## SRR594394 g21 Mus musculus colon DBA.2J
## SRR594395 g13 Mus musculus heart DBA.2J3.3.4 Preprocess Assay Data
The raw RNA-Seq counts are preprocessed with the following steps: (1) normalization, (2) aggregation of replicates, and (3) filtering of un-reliable expression data. The details of these steps are explained in the sub-section of the Human Brain example.
rse.mus <- cvt_id(db='org.Mm.eg.db', data=rse.mus, from.id='ENSEMBL', to.id='SYMBOL', desc=TRUE) # Convert Ensembl ids to UniProt ids.
se.nor.mus <- norm_data(data=rse.mus, norm.fun='CNF', log2.trans=TRUE) # Normalization
se.aggr.mus <- aggr_rep(data=se.nor.mus, sam.factor='organism_part', con.factor='strain', aggr='mean') # Aggregation of replicates
se.fil.mus <- filter_data(data=se.aggr.mus, sam.factor='organism_part', con.factor='strain', pOA=c(0.01, 5), CV=c(0.6, 100)) # Filtering of genes with low counts and variance 3.3.5 SHM: Transparency
The pre-processed expression data for gene Scml2 is plotted in form
of an SHM. In this case the plot includes expression data for 8 tissues across 3
mouse strains.
dat.mus <- SPHM(svg=svg.mus, bulk=se.fil.mus)
shm(data=dat.mus, ID=c('H19'), legend.width=0.7, legend.text.size=10, sub.title.size=9, ncol=3, ft.trans=c('skeletal muscle'), legend.ncol=2, line.size=0.2, line.color='grey70')
Figure 4: SHM of mouse organs
This is a multiple-layer image where the shapes of the ‘skeletal muscle’ is set transparent to expose ‘lung’ and ‘heart’.
The SHM plots in Figures 4 and below demonstrate
the usage of the transparency feature via the ft.trans parameter. The
corresponding mouse organ aSVG image includes overlapping tissue layers. In
this case the skelectal muscle layer partially overlaps with lung and heart
tissues. To view lung and heart in Figure 4, the skelectal
muscle tissue is set transparent with ft.trans=c('skeletal muscle').
To fine control the visual effects in feature rich aSVGs, the line.size and
line.color parameters are useful. This way one can adjust the thickness and
color of complex structures.
gg <- shm(data=dat.mus, ID=c('H19'), legend.text.size=10, sub.title.size=9, ncol=3, legend.ncol=2, line.size=0.1, line.color='grey70')
Figure 5: SHM of mouse organs
This is a multiple-layer image where the view onto ‘lung’ and ‘heart’ is obstructed by displaying the ‘skeletal muscle’ tissue.
A third example on real data from Expression Atlas is SHMs of time series across chicken organs. Since the procedures are the same with the examples above, this example is illustrated in the Supplementary Section.
3.4 Arabidopsis Shoot
This section generates an SHM for Arabidopsis thaliana tissues with gene expression
data from the Affymetrix microarray technology. The chosen experiment used
ribosome-associated mRNAs from several cell populations of shoots and roots that were
exposed to hypoxia stress (Mustroph et al. 2009). In this case the expression data
will be downloaded from GEO with utilites
from the GEOquery package (Davis and Meltzer 2007). The data preprocessing routines are
specific to the Affymetrix technology. The remaining code components for
generating SHMs are very similar to the previous examples. For brevity, the
text in this section explains mainly the steps that are specific to this data
set.
3.4.1 Gene Expression Data
The GSE14502 data set is downloaded with the getGEO function from the GEOquery package. Intermediately, the expression data is stored in an
ExpressionSet container (Huber et al. 2015), and then converted to a
SummarizedExperiment object.
gset <- read_cache(cache.pa, 'gset') # Retrieve data from cache.
if (is.null(gset)) { # Save downloaded data to cache if it is not cached.
gset <- getGEO("GSE14502", GSEMatrix=TRUE, getGPL=TRUE)[[1]]
save_cache(dir=cache.pa, overwrite=TRUE, gset)
}
se.sh <- as(gset, "SummarizedExperiment")The gene symbol identifiers are extracted from the rowData component to be used
as row names. Similarly, one can work with AGI identifiers by providing below AGI
under Gene.Symbol.
rownames(se.sh) <- make.names(rowData(se.sh)[, 'Gene.Symbol'])A slice of the experimental design stored in the
colData slot is returned. Both the samples and treatments are contained in the title column.
The samples are indicated by corresponding promoters (pGL2, pCO2, pSCR, pWOL, p35S) and treatments include control and hypoxia.
colData(se.sh)[60:63, 1:2]
## DataFrame with 4 rows and 2 columns
## title geo_accession
## <character> <character>
## GSM362227 shoot_hypoxia_pGL2_r.. GSM362227
## GSM362228 shoot_hypoxia_pGL2_r.. GSM362228
## GSM362229 shoot_control_pRBCS_.. GSM362229
## GSM362230 shoot_control_pRBCS_.. GSM3622303.4.2 aSVG Image
In this example, the aSVG image has been generated in Inkscape from
the corresponding figure in Mustroph et al. (2009). Detailed instructions for generating custom aSVG images are provided in the
SVG tutorial. The resulting custom aSVG file ‘arabidopsis.thaliana_shoot_shm.svg’ is included in the spatialHeatmap package and imported as below.
svg.sh.pa <- system.file("extdata/shinyApp/data", "arabidopsis.thaliana_shoot_shm.svg", package="spatialHeatmap")
svg.sh <- read_svg(svg.sh.pa)3.4.3 Experimental Design
A sample target file that is included in this package is imported and then loaded to the colData slot of se.sh. To inspect its content, four selected rows are returned.
sh.tar <- system.file('extdata/shinyApp/data/target_arab.txt', package='spatialHeatmap')
target.sh <- read.table(sh.tar, header=TRUE, row.names=1, sep='\t') # Importing
colData(se.sh) <- DataFrame(target.sh) # Loading
target.sh[60:63, ]
## col.name samples conditions
## shoot_hypoxia_pGL2_rep1 GSM362227 shoot_pGL2 hypoxia
## shoot_hypoxia_pGL2_rep2 GSM362228 shoot_pGL2 hypoxia
## shoot_control_pRBCS_rep1 GSM362229 shoot_pRBCS control
## shoot_control_pRBCS_rep2 GSM362230 shoot_pRBCS control3.4.4 Preprocess Assay Data
The downloaded GSE14502 data set has already been normalized with the RMA algorithm (Gautier et al. 2004). Thus, the pre-processing steps can be restricted to replicate aggregation and filtering.
se.aggr.sh <- aggr_rep(data=se.sh, sam.factor='samples', con.factor='conditions', aggr='mean') # Replicate agggregation using mean
se.fil.arab <- filter_data(data=se.aggr.sh, sam.factor='samples', con.factor='conditions', pOA=c(0.03, 6), CV=c(0.30, 100)) # Filtering of genes with low intensities and variance3.4.5 SHM: Microarray
The expression profile for the HRE2 gene is plotted for the control and the hypoxia treatment across six cell types (Figure 6).
dat.sh <- SPHM(svg=svg.sh, bulk=se.fil.arab)
shm(data=dat.sh, ID=c("HRE2"), legend.ncol=2, legend.text.size=10, legend.key.size=0.02)
Figure 6: SHM of Arabidopsis shoots
The expression profile of the HRE2 gene is plotted for control and hypoxia treatment across six cell types.
3.5 Superimposing raster and vector graphics
spatialHeatmap allows to superimpose raster images with vector-based SHMs. This
way one can generate SHMs that resemble photographic representations
of tissues, organs or entire organisms. For this to work the shapes represented in the
vector-graphics need to be an aligned carbon copy of the raster image.
Supported file formats for the raster image are JPG/JPEG and PNG, and for the
vector image it is SVG. Matching raster and vector graphics are indicated by
identical base names in their file names (e.g. imageA.png and imageA.svg).
The layout order in SHMs composed of multiple independent images can be
controlled by numbering the corresponding file pairs accordingly such as
imageA_1.png and imageA_1.svg, imageA_2.png and imageA_2.svg, etc.
In the following example, the required image pairs have been pre-generated from a study on abaxial bundle sheath (ABS) cells in maize leaves (Bezrutczyk et al. 2021). Their file names are labeled 1 and 2 to indicate two developmental stages.
Import paths of first png/svg image pair:
raster.pa1 <- system.file('extdata/shinyApp/data/maize_leaf_shm1.png', package='spatialHeatmap')
svg.pa1 <- system.file('extdata/shinyApp/data/maize_leaf_shm1.svg', package='spatialHeatmap')Import paths of second png/svg image pair:
raster.pa2 <- system.file('extdata/shinyApp/data/maize_leaf_shm2.png', package='spatialHeatmap')
svg.pa2 <- system.file('extdata/shinyApp/data/maize_leaf_shm2.svg', package='spatialHeatmap')The two pairs of png/svg images are imported in the SVG container svg.overlay.
svg.overlay <- read_svg(svg.path=c(svg.pa1, svg.pa2), raster.path=c(raster.pa1, raster.pa2))A slice of attributes in the first aSVG instance is shown.
attribute(svg.overlay)[[1]][1:3, ]
## # A tibble: 3 × 10
## feature id fill stroke sub.feature index element parent index.all
## <chr> <chr> <chr> <dbl> <chr> <int> <chr> <chr> <int>
## 1 rect817 rect817 none 0.0843 rect817 1 rect container 1
## 2 cell1 cell1 #98f0aa 0 path819 2 g container 2
## 3 cell1 cell1 #98f0aa 0 path821 3 g container 2
## # ℹ 1 more variable: index.sub <int>Create random quantitative assay data.
df.ovl <- data.frame(matrix(runif(6, min=0, max=5), nrow=3))
colnames(df.ovl) <- c('cell1', 'cell2') # Assign column names.
rownames(df.ovl) <- paste0('gene', seq_len(3)) # Assign row names
df.ovl[1:2, ]
## cell1 cell2
## gene1 1.637970 3.788981771
## gene2 1.850373 0.009639693To minimize masking of the features in the SHMs by dense regions in the raster images,
the alpha.overlay argument allows to adjust the transparency level. In Figure 7,
the spatial features of interest are superimposed onto the raster image.
dat.over <- SPHM(svg=svg.overlay, bulk=df.ovl)
shm(data=dat.over, charcoal=FALSE, ID=c('gene1'), alpha.overlay=0.5, bar.width=0.09, sub.title.vjust=4, legend.r=0.2)
Figure 7: Superimposing raster images with SHMs (colorful backaground)
The expression profiles of gene1 are plotted on ABS cells.
Another option for reducing masking effects is to display the raster image in black and white by setting charcoal=TRUE (Figure 8).
shm(data=dat.over, charcoal=TRUE, ID=c('gene1'), alpha.overlay=0.5, bar.width=0.09, sub.title.vjust=4, legend.r=0.2)
Figure 8: Superimposing raster images with SHMs (black and white background)
The expression profiles of gene1 are plotted on ABS cells.
3.6 SHMs of Multiple Variables
The SHM plots shown so far are restricted to two variables, here spatial features
(e.g. tissues) and treatments. In theory, the complexity of experimental designs scales
to any number of variables in spatialHeatmap. This section
extends to experiments with three or more variables, such as experiments with spatiotemporal resolution and geographical locations, genotypes, treatments, etc.
To visualize multi-variable assay data, the variables are reduced to two by keeping the spatial feature unchanged and combining all other variables into a composite one. For instance, the following example contains four variables including spatial features, time points, drug treatments and injury conditions. The latter three are combined to produce a composite variable.
3.6.1 Gene Expression Data
The following uses RNA-seq data assayed from hippocampus and hypothalamus in mouse brain, and the experimental variables include traumatic brain injury (TBI), 3 or 29 days post injury (DPI), candesartan or vehicle treatment (Attilio et al. 2021). The original data are modified for demonstration purpose and included in spatialHeatmap as a SummarizedExperiment object, which is imported below.
se.mus.vars <- readRDS(system.file('extdata/shinyApp/data/mus_brain_vars_se.rds', package='spatialHeatmap'))The experiment design is stored in colData slot, where ‘Veh’, ‘Drug’, ‘TBI’, and ‘NoTBI’ refer to ‘vehicle’, ‘candesartan’, ‘traumatic brain injury’, and ‘sham injury’ respectively. The time, treatment and injury variables are combined into a composite one comVar.
colData(se.mus.vars)[1:3, ]
## DataFrame with 3 rows and 5 columns
## tissue time treatment injury
## <character> <character> <character> <character>
## hippocampus__3DPI.Veh.NoTBI hippocampus 3DPI Veh NoTBI
## hippocampus__3DPI.Veh.NoTBI hippocampus 3DPI Veh NoTBI
## hippocampus__3DPI.Veh.TBI hippocampus 3DPI Veh TBI
## comVar
## <character>
## hippocampus__3DPI.Veh.NoTBI 3DPI.Veh.NoTBI
## hippocampus__3DPI.Veh.NoTBI 3DPI.Veh.NoTBI
## hippocampus__3DPI.Veh.TBI 3DPI.Veh.TBI
unique(colData(se.mus.vars)$comVar)
## [1] "3DPI.Veh.NoTBI" "3DPI.Veh.TBI" "3DPI.Drug.NoTBI" "3DPI.Drug.TBI"
## [5] "29DPI.Veh.NoTBI" "29DPI.Veh.TBI" "29DPI.Drug.NoTBI" "29DPI.Drug.TBI"Since this example data are small, the pre-processing only involves normalization without the filtering step. The expression values are aggregated across replicates in tissues (tissue) and replicates in the composite variable (comVar) with the summary statistic of mean, which is similar with the Human Brain exmple.
se.mus.vars.nor <- norm_data(data=se.mus.vars, norm.fun='ESF', log2.trans=TRUE) # Normalization.
## Normalising: ESF
## type
## "ratio"
se.mus.vars.aggr <- aggr_rep(data=se.mus.vars.nor, sam.factor='tissue', con.factor='comVar', aggr='mean') # Aggregate replicates.
assay(se.mus.vars.aggr)[1:3, 1:3]
## hippocampus__3DPI.Veh.NoTBI hippocampus__3DPI.Veh.TBI
## Zscan10 2.815413 0.09246384
## Zbp1 1.757163 1.87394752
## Xlr4a 1.412010 0.98498887
## hippocampus__3DPI.Drug.NoTBI
## Zscan10 1.089357
## Zbp1 1.434792
## Xlr4a 1.4608833.6.2 aSVG Image
The aSVG image of mouse brain is downloaded from the [EBI anatomogram] repository (https://github.com/ebi-gene-expression-group/anatomogram/tree/master/src/svg){target="_blank"} and included in spatialHeatmap, which is imported as below.
svg.mus.brain.pa <- system.file("extdata/shinyApp/data", "mus_musculus.brain.svg", package="spatialHeatmap")
svg.mus.brain <- read_svg(svg.mus.brain.pa)The aSVG features and attributes are partially shown.
tail(attribute(svg.mus.brain)[[1]], 3)
## # A tibble: 3 × 10
## feature id fill stroke sub.feature index element parent index.all
## <chr> <chr> <chr> <dbl> <chr> <int> <chr> <chr> <int>
## 1 hypothalamus UBER… none 0.05 hypothalam… 22 path LAYER… 15
## 2 nose UBER… none 0.05 nose 23 path LAYER… 16
## 3 corpora.quadrig… UBER… none 0.05 corpora.qu… 24 path LAYER… 17
## # ℹ 1 more variable: index.sub <int>3.6.3 SHM: Multiple Variables
The expression values of gene Acnat1 in hippocampus and hypothalamus across the composite variable are mapped to matching aSVG features. The output SHM plots of each composite variable are organized in a composite plot in Figure 9.
dat.mul.dim <- SPHM(svg=svg.mus.brain, bulk=se.mus.vars.aggr)
shm(data=dat.mul.dim, ID=c('Acnat1'), legend.r=1.5, legend.key.size=0.02, legend.text.size=12, legend.nrow=3)
Figure 9: SHM plots of multiple variable
Gene expression values of Acnat1 in hippocampus and hypothalamus under composite variables are mapped to corresponding aSVG features.
3.7 Multiple aSVGs
In a spatiotemporal application, different development stages may need to be represented
in separate aSVG images. In such a case, the shm function is able to arrange
multiple aSVGs in a single SHM plot. To organize the subplots, the names
of the separate aSVG files are expected to include the following suffixes: *_shm1.svg,
*_shm2.svg, etc.
As a simple testing example, the following stores random numbers as expression
values in a data.frame.
df.random <- data.frame(matrix(runif(50, min=0, max=10), nrow=10))
colnames(df.random) <- c('shoot_totalA__treatment1', 'shoot_totalA__treatment2', 'shoot_totalB__treatment1', 'shoot_totalB__treatment2', 'notMapped') # Assign column names
rownames(df.random) <- paste0('gene', 1:10) # Assign row names
df.random[1:2, ]
## shoot_totalA__treatment1 shoot_totalA__treatment2
## gene1 1.924185 2.6496046
## gene2 4.520996 0.6959067
## shoot_totalB__treatment1 shoot_totalB__treatment2 notMapped
## gene1 5.112577 7.649234 1.257528
## gene2 4.749874 4.356737 5.105836The paths to the aSVG files are obtained, here for younger and older plants
using *_shm1 and *_shm1, respectively, which are generated from
Mustroph et al. (2009). Subsequently, the two aSVG files are loaded with the read_svg
function.
svg.sh1 <- system.file("extdata/shinyApp/data", "arabidopsis.thaliana_organ_shm1.svg", package="spatialHeatmap")
svg.sh2 <- system.file("extdata/shinyApp/data", "arabidopsis.thaliana_organ_shm2.svg", package="spatialHeatmap")
svg.sh.mul <- read_svg(c(svg.sh1, svg.sh2))The following generates the corresponding SHMs plot for gene2. The orginal
image dimensions can be preserved by assigning TRUE to the preserve.scale argument.
dat.mul.svg <- SPHM(svg=svg.sh.mul, bulk=df.random)
shm(data=dat.mul.svg, ID=c('gene2'), width=0.7, legend.r=0.2, legend.width=1, preserve.scale=TRUE, bar.width=0.09, line.color='grey50') 
Figure 10: Spatial heatmap of Arabidopsis at two growth stages
The expression profile of “gene2” under condition1 and condition2 is plotted for two growth stages (top and bottom row).
Note in Figure 10 shoots are drawn with thicker outlines than roots.
This is another useful feature of shm, i.e. preserving the outline
thicknesses defined in aSVG files. This feature is particularly useful in cellular SHMs
where different cell types may have different cell-wall thicknesses. The outline
widths can be updated with update_feature programatically, or within Inkscape
manually. The former is illustrated in the Supplementary Section.
4 Extended functionalities
4.1 Spatial Enrichment
The Spatial Enrichment (SpEn) functionality is an extension of SHMs for identifying groups of biomolecules (e.g. RNAs, proteins, metabolites) that are particularly abundant or enriched in certain spatial regions, such as tissue-specific transcripts. Given a group of spatial features, SpEn identifies biomolecules significantly up- or down-regulated in each spatial feature relative to all other features (references). These biomolecules are classified as spatially enriched or depleted respectively. Then by querying a feature in the enrichment results, the corresponding enriched and depleted biomolecules will be returned, and their abundance values are subsequently visualized as enrichment SHMs. Similarly, biomolecules enriched or depleted in one experimental variable relative to reference variables can be detected and visualized as well. SpEn utilizes differential expression (DE) analysis methods to detect enriched or depleted biomolecules, including edgeR (McCarthy et al. 2012), limma (Ritchie et al. 2015), DESeq2 (Love, Huber, and Anders 2014), distinct (Tiberi and Robinson. 2020).
The application of SpEn is illustrated with the
above mouse organ data. The function sf_var is used to subset spatial features and experiment variables of interest in the assay data. In the following, five features (‘brain’, ‘liver’, ‘lung’, ‘colon’, and ‘kidney’) and three experiment variables (mouse strains ‘DBA.2J’, ‘C57BL.6’, and ‘CD1’) are subsetted. The com.by argument specifies whether the enrichment will be performed on spatial features (ft) or variables (var). In the following, com.by is set ft, so SpEn is performed for spatial features and the variables under each spatial feature are treated as replicates.
sub.mus <- sf_var(data=rse.mus, feature='organism_part', ft.sel=c('brain', 'lung', 'colon', 'kidney', 'liver'), variable='strain', var.sel=c('DBA.2J', 'C57BL.6', 'CD1'), com.by='ft')
colData(sub.mus)[1:3, c('organism_part', 'strain')]
## DataFrame with 3 rows and 2 columns
## organism_part strain
## <character> <character>
## brain__DBA.2J brain DBA.2J
## colon__DBA.2J colon DBA.2J
## kidney__DBA.2J kidney DBA.2JThe subsetted data are filtered. Details about this step are given under the human brain section.
sub.mus.fil <- filter_data(data=sub.mus, pOA=c(0.5, 15), CV=c(0.8, 100), verbose=FALSE)The SpEn is implemented in the function spatial_enrich. In the following, the method edgeR is chosen and the count data are internally normalized by the TMM method from edgeR. The enrichment results are selected by log2 fold change (log2.fc) and FDR (fdr). The outliers argument specifies a number of outliers allowed in references.
enr.res <- spatial_enrich(sub.mus.fil, method=c('edgeR'), norm='TMM', log2.fc=1, fdr=0.05, outliers=1)The overlap of enriched biomolecules (type='up') across spatial features are presented in an UpSet plot (plot='upset'). Assigning matrix or venn to the plot argument will present the overlaps in form of a matrix plot or Venn diagram respectively.
ovl_enrich(enr.res, type='up', plot='upset')
Figure 11: Overlap of enriched biomolecules across spatial features
The enriched and depleted genes in brain are queried with the function query_enrich. In the type column, ‘up’ and ‘down’ refer to ‘enriched’ and ‘depleted’ respectively, while the total column shows the total reference features excluding outliers.
en.brain <- query_enrich(enr.res, 'brain')
## Done!
up.brain <- subset(rowData(en.brain), type=='up' & total==4)
up.brain[1:2, 1:3] # Enriched.
## DataFrame with 2 rows and 3 columns
## type total method
## <character> <numeric> <character>
## Kif5c up 4 edgeR
## Mapk10 up 4 edgeR
dn.brain <- subset(rowData(en.brain), type=='down' & total==4)
dn.brain[1:2, 1:3] # Depleted.
## DataFrame with 2 rows and 3 columns
## type total method
## <character> <numeric> <character>
## Cdhr5 down 4 edgeR
## Slc22a18 down 4 edgeROne enriched (Kif5c) and one depleted
(Cdhr5) gene in mouse brain are chosen to plot SHMs. The resulting SHMs are termed as SHMs of spatially-enriched/depleted biomolecules (enrichment SHMs) respectively.
dat.enrich <- SPHM(svg=svg.mus, bulk=en.brain)
shm(data=dat.enrich, ID=c('Kif5c', 'Cdhr5'), legend.r=1, legend.nrow=7, sub.title.size=15, ncol=3, bar.width=0.09, lay.shm='gene')## ggplots/grobs: mus_musculus.male.svg ...
## ggplot: Kif5c, DBA.2J C57BL.6 CD1
## ggplot: Cdhr5, DBA.2J C57BL.6 CD1
## Legend plot ...
## CPU cores: 1
## Converting "ggplot" to "grob" ...
## Kif5c_DBA.2J_1 Kif5c_C57BL.6_1 Kif5c_CD1_1 Cdhr5_DBA.2J_1 Cdhr5_C57BL.6_1 Cdhr5_CD1_1
## Converting "ggplot" to "grob" ...
## 
Figure 12: Enrichment SHMs
Top row: SHMs of spatially-enriched gene. Bottom row: SHMs of spatially-depleted gene.
The expression profiles of the two chosen genes (Figure 12) are also presented in line graphs.
graph_line(assay(en.brain)[c('Kif5c', 'Cdhr5'), ], lgd.pos='right')
Figure 13: Line graph of gene expression profiles
4.2 Hierarchical Clustering
SHMs are suitable for comparing assay profiles among small number of biomolecules
(e.g. few genes or proteins) across cell types and conditions. To also
support analysis routines of larger number of biomolecules, spatialHeatmap integrates
functionalities for identifying groups of biomolecules with similar and/or dissimilar
assay profiles, and subsequently analyzing the results with data mining
methods that scale to larger numbers of biomolecules than SHMs, such as hierarchical
clustering and network analysis methods.
To identify similar biomolecules, the submatrix function can be used. It identifies biomolecules sharing similar profiles with one or more query biomolecules of
interest. The given example uses correlation coefficients as similarity metric.
The p argument allows to restrict the number of
similar biomolecules to return based on a percentage cutoff relative to the number of
biomolecules in the assay data set (e.g. 1% of the top most similar biomolecules). If several
query biomolecules are provided then the function returns the similar genes for each
query, while assuring uniqueness among biomolecules in the result.
In a typical scenario, a spatial feature of interest is chosen in the first place, then a query biomolecule is chosen for the chosen feature, such as a tissue-specific gene. The following introduces the large-scale data mining using as sample data the preprocessed gene expression data (se.fil.mus) from the Mouse Organs section. The brain is selected as the query feature and ‘Kif5c’ is selected as the query gene, which is spatially enriched in brain and known to play important roles in motor neurons (Kanai et al. 2000).
sub.mat <- submatrix(data=se.fil.mus, ID='Kif5c', p=0.15)The result from the previous step is the assay matrix subsetted to the genes sharing similar assay profiles with the query gene ‘Kif5c’.
assay(sub.mat)[1:2, 1:3] # Subsetted assay matrix## brain__DBA.2J colon__DBA.2J heart__DBA.2J
## Akap6 8.551867 2.331506 7.042128
## Adcyap1r1 7.962620 1.204208 3.775402
Subsequently, hierarchical clustering is applied to the subsetted assay matrix
containing only the genes that share profile similarities with the query gene ‘Kif5c’. The clustering result is displayed as a matrix heatmap where
the rows and columns are sorted by the corresponding hierarchical clustering
dendrograms (Figure 14). The position of the query gene (‘Kif5c’) is indicated in the heatmap by black lines. Setting static=FALSE
will launch the interactive mode, where users can zoom into the heatmap by
selecting subsections in the image or zoom out by double clicking.
res.hc <- matrix_hm(ID=c('Kif5c'), data=sub.mat, angleCol=60, angleRow=60, cexRow=0.8, cexCol=0.8, margin=c(10, 6), static=TRUE, arg.lis1=list(offsetRow=0.01, offsetCol=0.01))
Figure 14: Matrix Heatmap
Rows are genes and columns are samples. The query gene is tagged by black lines.
The most important information of Figure 14 is returned in res.hc. The row dendrogram is saved in res.hc$rowDendrogram. By using the function cut_dendro, it can be cut at a certain height (here h=15) to obtain the cluster containing the query gene.
cut_dendro(res.hc$rowDendrogram, h=15, 'Kif5c')## [1] "Aplp1" "Fbxl16" "Kif1a" "Kif5a" "Kif5c" "Nptxr" "Nrxn1" "Ntm"
## [9] "Stx1b"4.3 Network Graphs
4.3.1 Module Identification
Network analysis is performed with the WGCNA algorithm (Langfelder and Horvath 2008; Ravasz et al. 2002) using as input the subsetted assay matrix generated in the
previous section. The objective is to identify network modules that can be
visualized in the following step in form of network graphs. Applied to the gene
expression sample data used here, these network modules represent groups of
genes sharing highly similar expression profiles. Internally, the network
module identification includes five major steps. First, a correlation matrix
(Pearson or Spearman) is computed for each pair of biomolecules. Second, the
obtained correlation matrix is transformed into an adjacency matrix that
approximates the underlying global network to scale-free topology (Ravasz et al. 2002).
Third, the adjacency matrix is used to calculate a topological overlap matrix (TOM)
where shared neighborhood information among biomolecules is used to preserve robust
connections, while removing spurious connections. Fourth, the distance transformed
TOM is used for hierarchical clustering. To maximize time performance, the
hierarchical clustering is performed with the flashClust package (Langfelder and Horvath 2012).
Fifth, network modules are identified with the dynamicTreeCut package (Langfelder, Zhang, and Steve Horvath 2016). Its ds
(deepSplit) argument can be assigned integer values from 0 to 3, where
higher values increase the stringency of the module identification process. To
simplify the network module identification process with WGCNA, the individual
steps can be executed with a single function called adj_mod. The result is a
list containing the adjacency matrix and the final module assignments stored in
a data.frame. Since the interactive network feature used in the
visualization step below performs best on smaller modules, only modules are
returned that were obtained with stringent ds settings (here ds=2 and ds=3).
adj.mod <- adj_mod(data=sub.mat)A slice of the adjacency matrix is shown below.
adj.mod[['adj']][1:3, 1:3]## Akap6 Adcyap1r1 Erich6
## Akap6 1.0000000 0.49348330 0.10614679
## Adcyap1r1 0.4934833 1.00000000 0.06216853
## Erich6 0.1061468 0.06216853 1.00000000The module assignments are stored in a data frame. Its columns contain the results
for the ds=2 and ds=3 settings. Integer values \(>0\) are the module labels, while \(0\)
indicates unassigned biomolecules. The following returns the first three rows of the module
assignment result.
adj.mod[['mod']][1:3, ] ## 0 1 2 3
## Akap6 0 30 30 23
## Adcyap1r1 0 0 0 44
## Erich6 7 5 4 304.3.2 Module Visualization
Network modules can be visualized with the network function. To plot a module
containing a biomolecule (gene) of interest, its ID needs to be
provided under the corresponding argument. Typically, this could be one of the
biomolecules chosen for the above SHM plots. There are two modes to visualize the
selected module: static or interactive. Figure 15 was generated
with static=TRUE. Setting static=FALSE will generate the interactive
version. In the network plot shown below the nodes and edges represent genes
and their interactions, respectively. The thickness of the edges denotes the
adjacency levels, while the size of the nodes indicates the degree of
connectivity of each biomolecule in the network. The adjacency and connectivity levels
are also indicated by colors (Figure 15). The gene of interest assigned
under ID is labeled in the plot with the suffix tag: *_target.
network(ID="Kif5c", data=sub.mat, adj.mod=adj.mod, adj.min=0, vertex.label.cex=1.2, vertex.cex=3, static=TRUE)
Figure 15: Static network
Node size denotes gene connectivity while edge thickness stands for co-expression similarity.
Setting static=FALSE launches the interactive network. In this mode there
is an interactive color bar that denotes the gene connectivity. To modify it,
the color labels need to be provided in a comma separated format, e.g.:
yellow, orange, red. The latter would indicate that the gene connectivity
increases from yellow to red.
If the expression matrix contains gene/protein annotation information in the rowData slot and specified through desc, then it will be shown when moving the cursor over a network
node.
network(ID="Kif5c", data=sub.mat, adj.mod=adj.mod, desc='desc', static=FALSE)5 Shiny App
In additon to running spatialHeatmap from R, the package includes a Shiny
App that provides access to the same
functionalities from an intuitive-to-use web browser interface. Apart from
being very user-friendly, this App conveniently organizes the results of the
entire visualization workflow in a single browser window with options to adjust
the parameters of the individual components interactively. For instance, genes
can be selected and replotted in SHMs simply by clicking the corresponding
rows in the expression table included in the same window.
This representation is very efficient in guiding the interpretation of the results
in a visual and user-friendly manner. For testing purposes, the spatialHeatmap
Shiny App also includes ready-to-use sample expression data and aSVG images
along with embedded user instructions.
5.1 Local System
The Shiny App of spatialHeatmap can be launched from an R session with the following function call.
shiny_shm()The main dashboard panels of the Shiny App are organized as follows:
- Datasets: options for uploading custom datasets, selecting default datsets, or downloading example datasets.
- Spatial Heatmap: interactive tables of assay data/metadata, plotting SHMs with single or multiple genes, settings to customize the SHMs.
- Spatial Enrichment: overlap plots of enrichment results across spatial features, data table of enriched and depleted biomolecules.
- Data Mining: hierarchical clustering, K-means clustering, network analysis.
- Co-visualization: co-visualizing single-cell and bulk data in embedding plots (PCA, UMAP, or TSNE) and SHMs repectively. This functionality is described in a separate vignette.
A screenshot is shown below depicting SHM plots generated with the spatialHeatmap Shiny App (Figure 16).
Figure 16: Screenshot of spatialHeatmap’s Shiny App
The assay data along with metatdata are uploaded to the Shiny App as tabular (e.g. in CSV or TSV format) or ‘.rds’ files. The latter is a SummarizedExperiment object saved with the saveRDS function. In addition, the images are uploaded as aSVG files. The filter_data function can be used to export assay data from SummarizedExperiment to tabular files (TSV format). In the following example, the file argument specifies the the output file name, while the sam.factor and con.factor specifies spatial features and experiment variables in the colData slot respectively, which will be retained in the column names of the output tabular file. An example of the output format is shown in Table 1. In addition, the desc argument can be optionally used to specify gene annotations in the rowData slot, which will be appended to the output tabular file. More details of assay data formats are provided in the Supplement.
se.fil.arab <- filter_data(data=se.aggr.sh, desc="Target.Description", sam.factor='samples', con.factor='conditions', pOA=c(0.03, 6), CV=c(0.30, 100), file='./filtered_data.txt')5.2 Server Deployment
As most Shiny Apps, spatialHeatmap can be deployed as a centralized web
service. A major advantage of a web server deployment is that the
functionalities can be accessed remotely by anyone on the internet without the
need to use R on the user system. For deployment one can use custom web
servers or cloud services, such as AWS, GCP or
shinysapps.io. An example web instance for testing
spatialHeatmap online is available
here.
5.3 Custom Shiny App
The spatialHeatmap package also allows users to create custom Shiny App
instances using the custom_shiny function. This function provides options to include
custom assay and image data, and define default settings (e.g. color schemes). For details users want
to consult the help file. To maximize flexibility, the default settings are stored in a yaml file in the App. This makes it easy to refine and optimize default settings simply by changing this yaml file.
5.4 Database Backend
To maintain scalability, the Shiny App is designed to work with backend databases that comprise HDF5-based SummarizedExperiment objects (Fischer, Smith, and Pau 2020; Pagès 2020), aSVG files, and the pairing between assay data and aSVGs. This allows users to manage large amounts of assay data and aSVGs in a batch. Both data.frame or SummarizedExperiment formats are supported data structures to create the backend database. Each assay dataset is saved as an independent HDF5-based SummarizedExperiment object in a specified directory (Fischer, Smith, and Pau 2020; Pagès 2020). Simultaneously, the corresponding aSVG files and the pairing information (stored as a nested list) between the assay data and aSVGs are also saved in the same directory. Finally, the assay data, aSVG files, and pairing list are compressed into a ‘.tar’ file to form the database. More details are referred to the help file of write_hdf5.
6 Supplementary Section
The advantages of integrating the features of spatialHeatmap are showcased in a discovery workflow.
6.1 Numeric Data
The numceric data used to color the features in aSVG images can be provided as
three different object types including vector, data.frame and
SummerizedExperiment. When working with complex omics-based assay data then
the latter provides the most flexibility, and thus should be the preferred
container class for managing numeric data in spatialHeatmap. Both
data.frame and SummarizedExperiment can hold data from many measured biomolecules,
such as many genes or proteins. In contrast to this, the
vector class is only suitable for data from single biomolecules. Due to its
simplicity this less complex container is often useful for testing or when
dealing with simple data sets.
6.1.1 Object Types
6.1.1.1 vector
When using numeric vectors as input to shm, then their name slot needs
to be populated with strings matching the feature names in the corresponding aSVG.
To also specify experiment variables, their labels need to be appended to the feature names
with double underscores as separator, i.e. ’spFfeature__variable’.
The following example replots the testing example for two spatial features (‘putamen’ and ‘prefrontal.cortex’) and two experiment variables (‘1’ and ‘2’).
vec <- sample(x=1:100, size=5) # Random numeric values
names(vec) <- c('putamen__variable1', 'putamen__variable2', 'prefrontal.cortex__variable1', 'prefrontal.cortex__variable2', 'notMapped') # Assign unique names to random values
vec## putamen__variable1 putamen__variable2
## 47 38
## prefrontal.cortex__variable1 prefrontal.cortex__variable2
## 62 70
## notMapped
## 10With this configuration the resulting plot contains two SHMs
for the human brain, corresponding to ‘variable1’ and ‘variable2’ respectively. To keep the build time of this package to a minimum, the shm function call in the code block below is not evaluated, and thus the corresponding SHMs are not shown.
dat.vec <- SPHM(svg=svg.hum, bulk=vec)
shm(data=dat.vec, ID='testing', ncol=1, legend.r=1.2, sub.title.size=14, ft.trans='g4320', legend.nrow=3)6.1.1.2 data.frame
The data.frame stores assay data in a table, where columns and rows are spatial features/variables and biomolecules respectively. The naming of spatial features and variables in the column names follows the same convention as the above vector example. The following illustrates the data.frame container with random numbers.
df.test <- data.frame(matrix(sample(x=1:1000, size=100), nrow=20)) # Create numeric data.frame
colnames(df.test) <- names(vec) # Assign column names
rownames(df.test) <- paste0('gene', 1:20) # Assign row names
df.test[1:3, ]## putamen__variable1 putamen__variable2 prefrontal.cortex__variable1
## gene1 902 924 816
## gene2 564 931 123
## gene3 602 143 883
## prefrontal.cortex__variable2 notMapped
## gene1 116 58
## gene2 673 571
## gene3 380 331With the resulting data.frame, SHMs can be plotted for one or multiple genes (ID=c('gene1')).
dat.df <- SPHM(svg=svg.hum, bulk=df.test)
shm(data=dat.df, ID=c('gene1'), ncol=1, legend.r=1.2, sub.title.size=14, legend.nrow=3)Additional gene annotation information can be appended to the data.frame. This information can then be displayed
interactively in the network plots of the Shiny App by placing the cursor over network nodes.
df.test$ann <- paste0('ann', 1:20)
df.test[1:3, ]## putamen__variable1 putamen__variable2 prefrontal.cortex__variable1
## gene1 902 924 816
## gene2 564 931 123
## gene3 602 143 883
## prefrontal.cortex__variable2 notMapped ann
## gene1 116 58 ann1
## gene2 673 571 ann2
## gene3 380 331 ann36.1.1.3 SummarizedExperiment
The SummarizedExperiment class is a much more extensible and flexible container for providing metadata for both rows and columns of numeric data.
To import experimental design information (e.g. replicates, treatments) from tabular files, users can provide
a target file that will be stored in the colData slot of the
SummarizedExperiment (SE) object. Usually,
the target file contains at least two columns: one for spatial features and
one for experimental variables, where replicates are indicated by identical entries. The actual numeric matrix representing the assay data is stored in
the assay slot, where the rows correspond to biomolecules (e.g. genes). Optionally, additional annotation information for the rows (e.g. gene
descriptions) can be stored in the rowData slot.
For constructing a valid SummarizedExperiment object that can be used by
the shm function, the target file should meet the following requirements.
It should contain at least two columns. One column represents spatial features (
spFeature) and the other one experimental variables (variable) such as treatments. The rows in the target file correspond to the columns of the numeric data stored in theassayslot.To be colored in SHMs, the spatial features must have common identifiers between the assay data and aSVG. Note, the double underscore is a special string reserved for specific purposes in spatialHeatmap, and thus should be avoided for naming spatial features and variables.
The following example illustrates the design of a valid SummarizedExperiment object for generating SHMs. In this example, the ‘putamen’ tissue has 2 variables and each has 2 replicates. Thus, there are 4 assays for putamen. The same design applies to the prefrontal.cortex tissue.
spft <- c(rep('putamen', 4), rep('prefrontal.cortex', 4))
vars <- rep(c('variable1', 'variable1', 'variable2', 'variable2'), 2)
target.test <- data.frame(spFeature=spft, variable=vars, row.names=paste0('assay', 1:8))
target.test## spFeature variable
## assay1 putamen variable1
## assay2 putamen variable1
## assay3 putamen variable2
## assay4 putamen variable2
## assay5 prefrontal.cortex variable1
## assay6 prefrontal.cortex variable1
## assay7 prefrontal.cortex variable2
## assay8 prefrontal.cortex variable2The assay slot is populated with a data.frame containing random
numbers. Each column corresponds to an assay in the target file (here imported
into colData), while each row corresponds to a gene.
df.se <- data.frame(matrix(sample(x=1:1000, size=160), nrow=20))
rownames(df.se) <- paste0('gene', 1:20)
colnames(df.se) <- row.names(target.test)
df.se[1:3, ]## assay1 assay2 assay3 assay4 assay5 assay6 assay7 assay8
## gene1 213 794 297 129 269 149 521 617
## gene2 961 331 673 636 598 980 839 84
## gene3 133 559 290 317 372 177 844 916Next, the final SummarizedExperiment object is constructed by providing the
numeric and target data under the assays and colData arguments,
respectively.
se <- SummarizedExperiment(assays=df.se, colData=target.test)
se## class: SummarizedExperiment
## dim: 20 8
## metadata(0):
## assays(1): ''
## rownames(20): gene1 gene2 ... gene19 gene20
## rowData names(0):
## colnames(8): assay1 assay2 ... assay7 assay8
## colData names(2): spFeature variableIn addition, row-wise annotation information (e.g. for genes) can be included in the rowData slot.
rowData(se) <- df.test['ann']The replicates are aggregated by means.
se.aggr <- aggr_rep(data=se, sam.factor='spFeature', con.factor='variable', aggr='mean')
assay(se.aggr)[1:2, ]## putamen__variable1 putamen__variable2 prefrontal.cortex__variable1
## gene1 503.5 213.0 209
## gene2 646.0 654.5 789
## prefrontal.cortex__variable2
## gene1 569.0
## gene2 461.5With the fully configured SummarizedExperiment object, a similar SHM is plotted as in the previous examples.
dat.se <- SPHM(svg=svg.hum, bulk=se.aggr)
shm(data=dat.se, ID=c('gene1'), ncol=1, legend.r=1.2, sub.title.size=14, ft.trans=c('g4320'), legend.nrow=3)6.2 aSVG Files
6.2.1 aSVG repository
A public aSVG repository, that can be used by spatialHeatmap directly, is the EBI anatomogram.
It contains annatomical aSVG images from different species. The same aSVG
images are also used by the web service of the Expression Atlas project. In addition, the
spatialHeatmap has its own repository called spatialHeatmap aSVG
Repository,
that stores custom aSVG files developed for this project (e.g. Figure
6).
If users cannot find a suitable aSVG image in these two repositories, we have developed a step-by-step SVG tutorial for creating custom aSVGs. For example, by following this tutorial, various anatomical images in the BAR eFP browser can be used as templates to create custom aSVGs. In the future, we will expand our repository with more aSVGs, and users are encouraged to contribute their own aSVGs to this repository, allowing them to share their creations with the community.
6.2.2 Update aSVG features
To edit spatial feature identifiers in aSVGs, the update_feature function can be used. The demonstration below first creates an empty folder tmp.dir1 and copies
into it the aSVG homo_sapiens.brain.svg.
tmp.dir1 <- file.path(tempdir(check=TRUE), 'shm1')
if (!dir.exists(tmp.dir1)) dir.create(tmp.dir1)
svg.hum.pa <- system.file("extdata/shinyApp/data", 'homo_sapiens.brain.svg', package="spatialHeatmap")
file.copy(from=svg.hum.pa, to=tmp.dir1, overwrite=TRUE) # Copy "homo_sapiens.brain.svg" file into 'tmp.dir1'The folder tmp.dir1 is queried with feature and species keywords, and matches are returned in a data.frame.
feature.df <- return_feature(feature=c('frontal cortex', 'prefrontal cortex'), species=c('homo sapiens', 'brain'), dir=tmp.dir1, remote=NULL, keywords.any=FALSE)
feature.dfNew feature identifiers are stored in a vector, corresponding to each of the returned features (here ‘prefrontal.cortex’ and ‘frontal.cortex’).
f.new <- c('prefrontalCortex', 'frontalCortex')To also update strokes (thickness of spatial feature outlines) and colors, store new strokes and colors in separate vectors in a similar way as above.
s.new <- c('0.05', '0.1') # New strokes.
c.new <- c('red', 'green') # New colors.Next, new features, strokes, and colors are stored as three respective columns in a data.frame. The column names featureNew, strokeNew, and colorNew are internally recognized by the update_feature function when updating the aSVG.
feature.df.new <- cbind(featureNew=f.new, strokeNew=s.new, colorNew=c.new, feature.df)
feature.df.newFinally, update_feature is used to update spatial features, strokes, and colors in the aSVG stored in the tmp.dir1 folder.
update_feature(df.new=feature.df.new, dir=tmp.dir1)6.3 Advanced Functionalities
6.3.1 SHM: Re-matching
spatialHeatmap supports re-matching one spatial feature in the assay data to one or multiple counterparts in the aSVG. The re-matching is defined in a named list. In the list, a name slot refers to a spatial feature in assay data and the corresponding list elements represent aSVG features for re-matching.
The example below takes assay data from the Human Brain section, such as svg.hum.sub and se.fil.hum. The spatial feature frontal.cortex in assay data is re-matched to frontal.cortex and prefrontal.cortex in the aSVG through a list (remat.hum). Figure 17 is the SHM before re-matching while Figure 18 is the SHM after re-matching.
remat.hum <- list(frontal.cortex=c('frontal.cortex', 'prefrontal.cortex'))dat.no.match <- SPHM(svg=svg.hum.sub, bulk=se.fil.hum)
shm(data=dat.no.match, ID=c('OLFM4'), lay.shm='con', ncol=1, legend.r=0.8, legend.nrow=2, h=0.6)
Figure 17: SHMs before re-matching
dat.rematch <- SPHM(svg=svg.hum.sub, bulk=se.fil.hum, match=remat.hum)
shm(data=dat.rematch, ID=c('OLFM4'), lay.shm='con', ncol=1, legend.r=0.8, legend.nrow=2, h=0.6)
Figure 18: SHMs after re-matching
The spatial feature frontal.cortex in assay data is re-matched to aSVG features frontal.cortex and prefrontal.cortex.
Re-matching has various applications, such as mapping gene expression profiles from one species to a closely related species in the same anatomical region. It enables leveraging existing data when direct measurements are challenging. Additionally, re-matching can be used to map gene assay profiles of a sub-tissue to the whole tissue, when assaying the entire tissue is difficult. It allows extrapolating and analyzing gene expression patterns of the entire tissue based on data from the sub-tissues.
6.4 SHMs of Time Series
This section generates a SHM plot for chicken data from the Expression Atlas. The code components are very similar to the Human Brain example. For brevity, the corresponding text explaining the code has been reduced to a minimum.
6.4.1 Gene Expression Data
The following searches the Expression Atlas for expression data from ‘heart’ and ‘gallus’.
all.chk <- read_cache(cache.pa, 'all.chk') # Retrieve data from cache.
if (is.null(all.chk)) { # Save downloaded data to cache if it is not cached.
all.chk <- searchAtlasExperiments(properties="heart", species="gallus")
save_cache(dir=cache.pa, overwrite=TRUE, all.chk)
}Among the matching entries, accession ‘E-MTAB-6769’ will be downloaded, which is an RNA-Seq experiment comparing the developmental changes across nine time points of seven organs (Cardoso-Moreira et al. 2019).
all.chk[3, ]
rse.chk <- read_cache(cache.pa, 'rse.chk') # Read data from cache.
if (is.null(rse.chk)) { # Save downloaded data to cache if it is not cached.
rse.chk <- getAtlasData('E-MTAB-6769')[[1]][[1]]
save_cache(dir=cache.pa, overwrite=TRUE, rse.chk)
}The first three rows of the design in the downloaded experiment are shown.
colData(rse.chk)[1:3, ]## DataFrame with 3 rows and 8 columns
## AtlasAssayGroup organism strain genotype
## <character> <character> <character> <character>
## ERR2576379 g1 Gallus gallus Red Junglefowl wild type genotype
## ERR2576380 g1 Gallus gallus Red Junglefowl wild type genotype
## ERR2576381 g2 Gallus gallus Red Junglefowl wild type genotype
## developmental_stage age sex organism_part
## <character> <character> <character> <character>
## ERR2576379 embryo 10 day female brain
## ERR2576380 embryo 10 day female brain
## ERR2576381 embryo 10 day female cerebellum6.4.2 aSVG Image
The following example shows how to download from the above SVG
repositories an aSVG image that matches the tissues and species
assayed in the downloaded data. As before downloaded images are saved to the directory tmp.dir.
tmp.dir <- file.path(tempdir(check=TRUE), 'shm')
# Query aSVGs.
feature.df <- return_feature(feature=c('heart', 'kidney'), species=c('gallus'), dir=tmp.dir, remote=remote)To meet the R/Bioconductor package requirements, the following uses aSVG file instances included in the spatialHeatmap package rather than the downloaded instances.
feature.df <- return_feature(feature=c('heart', 'kidney'), species=c('gallus'), dir=svg.dir, remote=NULL)
feature.df[1:2, c('feature', 'stroke', 'SVG')] # A slice of the search result.The target aSVG instance gallus_gallus.svg is imported.
svg.chk.pa <- system.file("extdata/shinyApp/data", "gallus_gallus.svg", package="spatialHeatmap")
svg.chk <- read_svg(svg.chk.pa)6.4.3 Experimental Design
A sample target file that is included in this package is imported and then loaded to the colData slot of rse.chk, the first three rows of which are displayed.
chk.tar <- system.file('extdata/shinyApp/data/target_chicken.txt', package='spatialHeatmap')
target.chk <- read.table(chk.tar, header=TRUE, row.names=1, sep='\t') # Importing
colData(rse.chk) <- DataFrame(target.chk) # Loading
target.chk[1:3, ]
## AtlasAssayGroup organism strain genotype
## ERR2576379 g1 Gallus gallus Red Junglefowl wild type genotype
## ERR2576380 g1 Gallus gallus Red Junglefowl wild type genotype
## ERR2576381 g2 Gallus gallus Red Junglefowl wild type genotype
## developmental_stage age sex organism_part
## ERR2576379 embryo day10 female brain
## ERR2576380 embryo day10 female brain
## ERR2576381 embryo day10 female cerebellum6.4.4 Preprocess Assay Data
The raw RNA-Seq count are preprocessed with the following steps: (1) normalization, (2) aggregation of replicates, and (3) filtering of reliable expression data, details of which are seen in the Human Brain example.
se.nor.chk <- norm_data(data=rse.chk, norm.fun='ESF', log2.trans=TRUE) # Normalization
se.aggr.chk <- aggr_rep(data=se.nor.chk, sam.factor='organism_part', con.factor='age', aggr='mean') # Replicate agggregation using mean
se.fil.chk <- filter_data(data=se.aggr.chk, sam.factor='organism_part', con.factor='age', pOA=c(0.01, 5), CV=c(0.6, 100)) # Filtering of genes with low counts and varince6.4.5 SHM: Time Series
The expression profile for gene ENSGALG00000006346 is plotted across nine time
points in four organs in form of a composite SHM with 9 panels. Their layout in
three columns is controlled with the argument setting ncol=3. In the legend plot, spatial features are labeled by label=TRUE.
dat.chk <- SPHM(svg=svg.chk, bulk=se.fil.chk)
shm(data=dat.chk, ID='ENSGALG00000006346', width=0.9, legend.width=0.9, legend.r=1.5, sub.title.size=9, ncol=3, legend.nrow=3, label=TRUE, verbose=FALSE)
Figure 19: Time course of chicken organs
The SHM shows the expression profile of a single gene across nine time points and four organs.
7 Version Informaion
sessionInfo()## R version 4.3.1 (2023-06-16)
## Platform: x86_64-pc-linux-gnu (64-bit)
## Running under: Ubuntu 22.04.3 LTS
##
## Matrix products: default
## BLAS: /home/biocbuild/bbs-3.17-bioc/R/lib/libRblas.so
## LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.10.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_GB LC_COLLATE=C
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: America/New_York
## tzcode source: system (glibc)
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] org.Mm.eg.db_3.17.0 org.Hs.eg.db_3.17.0
## [3] AnnotationDbi_1.62.2 BiocParallel_1.34.2
## [5] igraph_1.5.1 GEOquery_2.68.0
## [7] ExpressionAtlas_1.28.0 jsonlite_1.8.7
## [9] RCurl_1.98-1.12 xml2_1.3.5
## [11] limma_3.56.2 SeuratObject_4.1.3
## [13] Seurat_4.3.0.1 kableExtra_1.3.4
## [15] SingleCellExperiment_1.22.0 ggplot2_3.4.3
## [17] SummarizedExperiment_1.30.2 Biobase_2.60.0
## [19] GenomicRanges_1.52.0 GenomeInfoDb_1.36.2
## [21] IRanges_2.34.1 S4Vectors_0.38.1
## [23] BiocGenerics_0.46.0 MatrixGenerics_1.12.3
## [25] matrixStats_1.0.0 spatialHeatmap_2.6.2
## [27] knitr_1.43 BiocStyle_2.28.0
##
## loaded via a namespace (and not attached):
## [1] spatstat.sparse_3.0-2 bitops_1.0-7
## [3] doParallel_1.0.17 httr_1.4.7
## [5] webshot_0.5.5 RColorBrewer_1.1-3
## [7] dynamicTreeCut_1.63-1 backports_1.4.1
## [9] tools_4.3.1 sctransform_0.3.5
## [11] utf8_1.2.3 R6_2.5.1
## [13] lazyeval_0.2.2 uwot_0.1.16
## [15] withr_2.5.0 sp_2.0-0
## [17] gridExtra_2.3 preprocessCore_1.62.1
## [19] progressr_0.14.0 WGCNA_1.72-1
## [21] cli_3.6.1 spatstat.explore_3.2-1
## [23] flashClust_1.01-2 grImport_0.9-7
## [25] labeling_0.4.3 sass_0.4.7
## [27] spatstat.data_3.0-1 readr_2.1.4
## [29] genefilter_1.82.1 ggridges_0.5.4
## [31] pbapply_1.7-2 systemfonts_1.0.4
## [33] yulab.utils_0.0.8 foreign_0.8-84
## [35] svglite_2.1.1 scater_1.28.0
## [37] parallelly_1.36.0 impute_1.74.1
## [39] rstudioapi_0.15.0 RSQLite_2.3.1
## [41] FNN_1.1.3.2 generics_0.1.3
## [43] gridGraphics_0.5-1 gtools_3.9.4
## [45] ica_1.0-3 spatstat.random_3.1-5
## [47] dendextend_1.17.1 dplyr_1.1.2
## [49] GO.db_3.17.0 Matrix_1.6-1
## [51] ggbeeswarm_0.7.2 fansi_1.0.4
## [53] abind_1.4-5 lifecycle_1.0.3
## [55] yaml_2.3.7 edgeR_3.42.4
## [57] shinytoastr_2.1.1 BiocFileCache_2.8.0
## [59] gplots_3.1.3 Rtsne_0.16
## [61] grid_4.3.1 blob_1.2.4
## [63] promises_1.2.1 dqrng_0.3.0
## [65] crayon_1.5.2 shinydashboard_0.7.2
## [67] miniUI_0.1.1.1 lattice_0.21-8
## [69] beachmat_2.16.0 cowplot_1.1.1
## [71] annotate_1.78.0 KEGGREST_1.40.0
## [73] magick_2.7.5 pillar_1.9.0
## [75] metapod_1.8.0 future.apply_1.11.0
## [77] codetools_0.2-19 leiden_0.4.3
## [79] glue_1.6.2 data.table_1.14.8
## [81] vctrs_0.6.3 png_0.1-8
## [83] gtable_0.3.4 assertthat_0.2.1
## [85] cachem_1.0.8 xfun_0.40
## [87] S4Arrays_1.0.6 mime_0.12
## [89] survival_3.5-7 iterators_1.0.14
## [91] statmod_1.5.0 bluster_1.10.0
## [93] ellipsis_0.3.2 fitdistrplus_1.1-11
## [95] ROCR_1.0-11 nlme_3.1-163
## [97] bit64_4.0.5 filelock_1.0.2
## [99] RcppAnnoy_0.0.21 UpSetR_1.4.0
## [101] bslib_0.5.1 irlba_2.3.5.1
## [103] rpart_4.1.19 vipor_0.4.5
## [105] KernSmooth_2.23-22 Hmisc_5.1-0
## [107] spsComps_0.3.3.0 colorspace_2.1-0
## [109] DBI_1.1.3 nnet_7.3-19
## [111] DESeq2_1.40.2 tidyselect_1.2.0
## [113] curl_5.0.2 bit_4.0.5
## [115] compiler_4.3.1 rvest_1.0.3
## [117] htmlTable_2.4.1 BiocNeighbors_1.18.0
## [119] DelayedArray_0.26.7 plotly_4.10.2
## [121] bookdown_0.35 checkmate_2.2.0
## [123] scales_1.2.1 caTools_1.18.2
## [125] lmtest_0.9-40 rappdirs_0.3.3
## [127] stringr_1.5.0 digest_0.6.33
## [129] goftest_1.2-3 spatstat.utils_3.0-3
## [131] rmarkdown_2.24 XVector_0.40.0
## [133] base64enc_0.1-3 htmltools_0.5.6
## [135] pkgconfig_2.0.3 sparseMatrixStats_1.12.2
## [137] dbplyr_2.3.3 highr_0.10
## [139] fastmap_1.1.1 rlang_1.1.1
## [141] htmlwidgets_1.6.2 shiny_1.7.5
## [143] DelayedMatrixStats_1.22.6 farver_2.1.1
## [145] jquerylib_0.1.4 zoo_1.8-12
## [147] BiocSingular_1.16.0 magrittr_2.0.3
## [149] Formula_1.2-5 scuttle_1.10.2
## [151] GenomeInfoDbData_1.2.10 ggplotify_0.1.2
## [153] patchwork_1.1.3 munsell_0.5.0
## [155] Rcpp_1.0.11 viridis_0.6.4
## [157] reticulate_1.31 stringi_1.7.12
## [159] zlibbioc_1.46.0 MASS_7.3-60
## [161] plyr_1.8.8 parallel_4.3.1
## [163] listenv_0.9.0 ggrepel_0.9.3
## [165] deldir_1.0-9 Biostrings_2.68.1
## [167] splines_4.3.1 tensor_1.5
## [169] hms_1.1.3 locfit_1.5-9.8
## [171] fastcluster_1.2.3 spatstat.geom_3.2-4
## [173] reshape2_1.4.4 ScaledMatrix_1.8.1
## [175] XML_3.99-0.14 evaluate_0.21
## [177] scran_1.28.2 BiocManager_1.30.22
## [179] foreach_1.5.2 tzdb_0.4.0
## [181] httpuv_1.6.11 RANN_2.6.1
## [183] tidyr_1.3.0 purrr_1.0.2
## [185] polyclip_1.10-4 future_1.33.0
## [187] scattermore_1.2 rsvd_1.0.5
## [189] xtable_1.8-4 rsvg_2.4.0
## [191] later_1.3.1 viridisLite_0.4.2
## [193] tibble_3.2.1 memoise_2.0.1
## [195] beeswarm_0.4.0 cluster_2.1.4
## [197] globals_0.16.2 shinyAce_0.4.28 Funding
This project has been funded by NSF awards: PGRP-1546879, PGRP-1810468, PGRP-1936492.
9 References
Attilio, Peter J, Dustin M Snapper, Milan Rusnak, Akira Isaac, Anthony R Soltis, Matthew D Wilkerson, Clifton L Dalgard, and Aviva J Symes. 2021. “Transcriptomic Analysis of Mouse Brain After Traumatic Brain Injury Reveals That the Angiotensin Receptor Blocker Candesartan Acts Through Novel Pathways.” Front. Neurosci. 15 (March): 636259.
Bezrutczyk, Margaret, Nora R Zöllner, Colin P S Kruse, Thomas Hartwig, Tobias Lautwein, Karl Köhrer, Wolf B Frommer, and Ji-Yun Kim. 2021. “Evidence for Phloem Loading via the Abaxial Bundle Sheath Cells in Maize Leaves.” Plant Cell 33 (3): 531–47.
Cardoso-Moreira, Margarida, Jean Halbert, Delphine Valloton, Britta Velten, Chunyan Chen, Yi Shao, Angélica Liechti, et al. 2019. “Gene Expression Across Mammalian Organ Development.” Nature 571 (7766): 505–9.
Chang, Winston, and Barbara Borges Ribeiro. 2018. Shinydashboard: Create Dashboards with ’Shiny’. https://CRAN.R-project.org/package=shinydashboard.
Chang, Winston, Joe Cheng, JJ Allaire, Cars on Sievert, Barret Schloerke, Yihui Xie, Jeff Allen, Jonathan McPherson, Alan Dipert, and Barbara Borges. 2021. Shiny: Web Application Framework for R. https://CRAN.R-project.org/package=shiny.
Davis, Sean, and Paul Meltzer. 2007. “GEOquery: A Bridge Between the Gene Expression Omnibus (Geo) and Bioconductor.” Bioinformatics 14: 1846–7.
Fischer, Bernd, Mike Smith, and Gregoire Pau. 2020. Rhdf5: R Interface to Hdf5. https://github.com/grimbough/rhdf5.
Gautier, Laurent, Leslie Cope, Benjamin M. Bolstad, and Rafael A. Irizarry. 2004. “Affy—Analysis of Affymetrix Genechip Data at the Probe Level.” Bioinformatics 20 (3): 307–15. https://doi.org/10.1093/bioinformatics/btg405.
Huber, W., V. J. Carey, R. Gentleman, S. An ders, M. Carlson, B. S. Carvalho, H. C. Bravo, et al. 2015. “Orchestrating High-Throughput Genomic Analysis Wit H Bioconductor.” Nature Methods 12 (2): 115–21. http://www.nature.com/nmeth/journal/v12/n2/full/nmeth.3252.html.
Kanai, Y, Y Okada, Y Tanaka, A Harada, S Terada, and N Hirokawa. 2000. “KIF5C, a Novel Neuronal Kinesin Enriched in Motor Neurons.” J. Neurosci. 20 (17): 6374–84.
Keays, Maria. 2019. ExpressionAtlas: Download Datasets from Embl-Ebi Expression Atlas.
Langfelder, Peter, and Steve Horvath. 2008. “WGCNA: An R Package for Weighted Correlation Network Analysis.” BMC Bioinformatics 9 (December): 559.
Langfelder, Peter, Bin Zhang, and with contributions from Steve Horvath. 2016. DynamicTreeCut: Methods for Detection of Clusters in Hierarchical Clustering Dendrograms. https://CRAN.R-project.org/package=dynamicTreeCut.
Langfelder, P., and S. Horvath. 2012. “Fast R Functions for Robust Correlations and Hierarchical Clustering.” J. Stat. Softw. 46 (11). https://www.ncbi.nlm.nih.gov/pubmed/23050260.
Lekschas, Fritz, Harald Stachelscheid, Stefanie Seltmann, and Andreas Kurtz. 2015. “Semantic Body Browser: Graphical Exploration of an Organism and Spatially Resolved Expression Data Visualization.” Bioinformatics 31 (5): 794–96.
Love, Michael I., Wolfgang Huber, and Simon Anders. 2014. “Moderated Estimation of Fold Change and Dispersion for Rna-Seq Data with Deseq2.” Genome Biology 15 (12): 550. https://doi.org/10.1186/s13059-014-0550-8.
Maag, Jesper L V. 2018. “Gganatogram: An R Package for Modular Visualisation of Anatograms and Tissues Based on Ggplot2.” F1000Res. 7 (September): 1576.
McCarthy, Davis J., Chen, Yunshun, Smyth, and Gordon K. 2012. “Differential Expression Analysis of Multifactor Rna-Seq Experiments with Respect to Biological Variation.” Nucleic Acids Research 40 (10): 4288–97.
Merkin, Jason, Caitlin Russell, Ping Chen, and Christopher B Burge. 2012. “Evolutionary Dynamics of Gene and Isoform Regulation in Mammalian Tissues.” Science 338 (6114): 1593–9.
Morgan, Martin, Valerie Obenchain, Jim Hester, and Hervé Pagès. 2018. SummarizedExperiment: SummarizedExperiment Container.
Muschelli, John, Elizabeth Sweeney, and Ciprian Crainiceanu. 2014. “BrainR: Interactive 3 and 4D Images of High Resolution Neuroimage Data.” R J. 6 (1): 41–48.
Mustroph, Angelika, M Eugenia Zanetti, Charles J H Jang, Hans E Holtan, Peter P Repetti, David W Galbraith, Thomas Girke, and Julia Bailey-Serres. 2009. “Profiling Translatomes of Discrete Cell Populations Resolves Altered Cellular Priorities During Hypoxia in Arabidopsis.” Proc Natl Acad Sci U S A 106 (44): 18843–8.
Pagès, Hervé. 2020. HDF5Array: HDF5 Backend for Delayedarray Objects.
Papatheodorou, Irene, Nuno A Fonseca, Maria Keays, Y Amy Tang, Elisabet Barrera, Wojciech Bazant, Melissa Burke, et al. 2018. “Expression Atlas: gene and protein expression across multiple studies and organisms.” Nucleic Acids Res. 46 (D1): D246–D251. https://doi.org/10.1093/nar/gkx1158.
Prudencio, Mercedes, Veronique V Belzil, Ranjan Batra, Christian A Ross, Tania F Gendron, Luc J Pregent, Melissa E Murray, et al. 2015. “Distinct Brain Transcriptome Profiles in C9orf72-Associated and Sporadic ALS.” Nat. Neurosci. 18 (8): 1175–82.
Ravasz, E, A L Somera, D A Mongru, Z N Oltvai, and A L Barabási. 2002. “Hierarchical Organization of Modularity in Metabolic Networks.” Science 297 (5586): 1551–5.
Ritchie, Matthew E, Belinda Phipson, Di Wu, Yifang Hu, Charity W Law, Wei Shi, and Gordon K Smyth. 2015. “limma Powers Differential Expression Analyses for RNA-Sequencing and Microarray Studies.” Nucleic Acids Research 43 (7): e47. https://doi.org/10.1093/nar/gkv007.
Tiberi, Simone, and Mark D. Robinson. 2020. Distinct: Distinct: A Method for Differential Analyses via Hierarchical Permutation Tests. https://github.com/SimoneTiberi/distinct.
Waese, Jamie, Jim Fan, Asher Pasha, Hans Yu, Geoffrey Fucile, Ruian Shi, Matthew Cumming, et al. 2017. “EPlant: Visualizing and Exploring Multiple Levels of Data for Hypothesis Generation in Plant Biology.” Plant Cell 29 (8): 1806–21.
Winter, Debbie, Ben Vinegar, Hardeep Nahal, Ron Ammar, Greg V Wilson, and Nicholas J Provart. 2007. “An ‘Electronic Fluorescent Pictograph’ Browser for Exploring and Analyzing Large-Scale Biological Data Sets.” PLoS One 2 (8): e718.