network: Network graphs

In jianhaizhang/spatialHeatmap: spatialHeatmap

Network graphs

Description

This function presents a network module returned by adj_mod in form of a static or interactive network graph. In the network graph, nodes are biomolecules (genes, proteins, etc) and edges are adjacencies (coexpression similarities) between biomolecules. The thicker edge denotes higher adjacency between nodes while larger node indicates higher connectivity (sum of a node's adjacencies with all its direct neighbours).

In the interactive mode, there is an interactive color scale to denote node connectivity. The color ingredients can only be separated by comma, semicolon, single space, dot, hypen, or underscore. E.g. 'yellow,orange,red', which means node connectivity increases from yellow to red. If too many edges (e.g.: > 500) are displayed, the app may get crashed, depending on the computer RAM. So the "Adjacency threshold" option sets a threthold to filter out weak edges. Meanwhile, the "Maximun edges" limits the total edges to show. In case a very low adjacency threshold is chosen and introduces too many edges that exceed the "Maximun edges", the App will internally increase the adjacency threshold until the edge total is within the "Maximun edges". The adjacency threshold of 1 leads to no edges. In this case the App wil internally decrease this threshold until the number of edges reaches the "Maximun edges". If adjacency threshold of 0.998 is selected and no edge remains, this App will also internally update the edges to 1 or 2. To maintain acceptable performance, users are advised to choose a stringent threshold (e.g. 0.9) initially, then decrease the value gradually. The interactive feature allows users to zoom in and out, or drag a node around. All the node IDs in the network module are listed in "Select by id" in decreasing order according to node connectivity. The ID of a chosen target biomolecule is appended "_target" as a label. By clicking an ID in this list, users can identify the corresponding node in the network. If the input data matrix has annotations for biomolecules, then the annotation can be seen by hovering the cursor over a node.

Usage

network(
  ID = NULL,
  mod.lab = NULL,
  data,
  assay.na = NULL,
  adj.mod,
  ds = "3",
  adj.min = 0,
  con.min = 0,
  desc = NULL,
  node.col = c("turquoise", "violet"),
  edge.col = c("yellow", "blue"),
  vertex.label.cex = 1,
  vertex.cex = 3,
  edge.cex = 3,
  layout = "circle",
  mar.net = c(2, 1, 1, 1),
  mar.key = c(3, 10, 1, 10),
  key.lab.size = 1,
  key.text.size = 1,
  main = NULL,
  static = TRUE,
  dir = NULL,
  ...
)

Arguments

ID	A biomolecule of interest. The network module containing this ID will be visualized.
mod.lab	A module label (e.g. 1, 2, 3), the module of which will be visualized.
data	The subsetted data matrix returned by the function submatrix.
assay.na	Applicable when data is 'SummarizedExperiment' or 'SingleCellExperiment', where multiple assays could be stored. The name of target assay to use.
adj.mod	A two-component list returned by adj_mod, consisting of the adjacency matrix and module assignments.
ds	One of '0', '1', '2', or '3' (default), the module identification sensitivity level. The smaller 'ds' indicates larger but less modules while the larger 'ds' denotes smaller but more modules. See function adj_mod for details.
adj.min	A cutoff of adjacency between nodes. Edges with adjacency below this cutoff will not be shown. Default is 0. Applicable to static network.
con.min	A cutoff of node connectivity. Nodes with connectivity below which will not be shown. Default is 0. Applicable to static network.
desc	A column name in the rowData slot of SummarizedExperiment. If provided, when mousing over the nodes in the interactive network, the corresponding description will show up.
node.col	A vector of color ingredients for node color scale in the static image. The default is 'c("turquoise", "violet")', where node connectivity increases from "turquoise" to "violet".
edge.col	A vector of color ingredients for edge color scale in the static image. The default is 'c("yellow", "blue")', where edge adjacency increases from "yellow" to "blue".
vertex.label.cex	The size of node label in the static and interactive networks. The default is 1.
vertex.cex	The size of nodes in the static image. The default is 3.
edge.cex	The size of edges in the static image. The default is 10.
layout	The layout of the network in static images, either 'circle' (default) or 'fr'. The 'fr' stands for force-directed layout algorithm developed by Fruchterman and Reingold.
mar.net, mar.key	A four-component numeric vector corresponding to bottom, left, top, and right margins of the network, and color keys respectively.
key.lab.size, key.text.size	The size of color key label and text respectively.
main	The title in the static image. Default is NULL.
static	Logical. 'TRUE' and 'FALSE' return a static and interactive network graphs respectively.
dir	The directory to save nodes and edges of the module.
...	Other arguments passed to the generic function plot.default, e.g.: asp=1.

Value

A static or interactive network graph.

Author(s)

Jianhai Zhang [e-mail]
Dr. Thomas Girke [e-mail]

References

Martin Morgan, Valerie Obenchain, Jim Hester and Hervé Pagès (2018). SummarizedExperiment: SummarizedExperiment container. R package version 1.10.1 Csardi G, Nepusz T: The igraph software package for complex network research, InterJournal, Complex Systems 1695. 2006. http://igraph.org R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/ Winston Chang, Joe Cheng, JJ Allaire, Yihui Xie and Jonathan McPherson (2018). shiny: Web Application Framework for R. R package version 1.1.0. https://CRAN.R-project.org/package=shiny Winston Chang and Barbara Borges Ribeiro (2018). shinydashboard: Create Dashboards with 'Shiny'. R package version 0.7.1. https://CRAN.R-project.org/package=shinydashboard Almende B.V., Benoit Thieurmel and Titouan Robert (2018). visNetwork: Network Visualization using 'vis.js' Library. R package version 2.0.4. https://CRAN.R-project.org/package=visNetwork Keays, Maria. 2019. ExpressionAtlas: Download Datasets from EMBL-EBI Expression Atlas 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. doi:10.1186/s13059-014-0550-8 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 Dragulescu A, Arendt C (2020). _xlsx: Read, Write, Format Excel 2007 and Excel 97/2000/XP/2003 Files_. R package version 0.6.5, <https://CRAN.R-project.org/package=xlsx>.

Examples

## The example data included in this package come from an RNA-seq analysis on 
## development of 7 chicken organs under 9 time points (Cardoso-Moreira et al. 2019). 
## The complete raw count data are downloaded using the R package ExpressionAtlas
## (Keays 2019) with the accession number "E-MTAB-6769". 

# Access example count data. 
count.chk <- read.table(system.file('extdata/shinyApp/data/count_chicken.txt', 
package='spatialHeatmap'), header=TRUE, row.names=1, sep='\t')
count.chk[1:3, 1:5]

# A targets file describing spatial features and variables is made based on the 
# experiment design.
target.chk <- read.table(system.file('extdata/shinyApp/data/target_chicken.txt', 
package='spatialHeatmap'), header=TRUE, row.names=1, sep='\t')
# Every column in example data 2 corresponds with a row in the targets file. 
target.chk[1:5, ]
# Store example data in "SummarizedExperiment".
library(SummarizedExperiment)
se.chk <- SummarizedExperiment(assay=count.chk, colData=target.chk)

# Normalize data.
se.chk.nor <- norm_data(data=se.chk, norm.fun='CNF', log2.trans=TRUE)

# Aggregate replicates of "spatialFeature_variable", where spatial features are organs
# and variables are ages.
se.chk.aggr <- aggr_rep(data=se.chk.nor, sam.factor='organism_part', con.factor='age',
aggr='mean')
assay(se.chk.aggr)[1:3, 1:3]

# Genes with experssion values >= 5 in at least 1% of all samples (pOA), and coefficient
# of variance (CV) between 0.2 and 100 are retained.
se.chk.fil <- filter_data(data=se.chk.aggr, sam.factor='organism_part', con.factor='age', 
pOA=c(0.01, 5), CV=c(0.2, 100), file=NULL)

## Subset the data matrix for gene 'ENSGALG00000019846' and 'ENSGALG00000000112'.
se.sub.mat <- submatrix(data=se.chk.fil, ID=c('ENSGALG00000019846', 
'ENSGALG00000000112'), p=0.1) 

## Hierarchical clustering. 
library(dendextend)
# Static matrix heatmap.
mhm.res <- matrix_hm(ID=c('ENSGALG00000019846', 'ENSGALG00000000112'), data=se.sub.mat, 
angleCol=80, angleRow=35, cexRow=0.8, cexCol=0.8, margin=c(8, 10), static=TRUE, 
arg.lis1=list(offsetRow=0.01, offsetCol=0.01))
# Clusters containing "ENSGALG00000019846".
cut_dendro(mhm.res$rowDendrogram, h=15, 'ENSGALG00000019846')

# Interactive matrix heatmap.
 matrix_hm(ID=c('ENSGALG00000019846', 'ENSGALG00000000112'), data=se.sub.mat, 
angleCol=80, angleRow=35, cexRow=0.8, cexCol=0.8, margin=c(8, 10), static=FALSE, 
arg.lis1=list(offsetRow=0.01, offsetCol=0.01)) 


# In case the interactive heatmap is not automatically opened, run the following code snippet.
# It saves the heatmap as an HTML file that is assigned to the "file" argument.

mhm <- matrix_hm(ID=c('ENSGALG00000019846', 'ENSGALG00000000112'), data=se.sub.mat, 
angleCol=80, angleRow=35, cexRow=0.8, cexCol=0.8, margin=c(8, 10), static=FALSE, 
arg.lis1=list(offsetRow=0.01, offsetCol=0.01))
htmlwidgets::saveWidget(widget=mhm, file='mhm.html', selfcontained=FALSE)
browseURL('mhm.html')


## Adjacency matrix and module identification 
adj.mod <- adj_mod(data=se.sub.mat)

# The adjacency is a measure of co-expression similarity between genes, where larger
# value denotes higher similarity.
adj.mod[['adj']][1:3, 1:3]

# The modules are identified at four sensitivity levels (ds=0, 1, 2, or 3). From 0 to 3, 
# more modules are identified but module sizes are smaller. The 4 sets of module 
# assignments are returned in a data frame, where column names are sensitivity levels. 
# The numbers in each column are module labels, where "0" means genes not assigned to 
# any module.
adj.mod[['mod']][1:3, ]

# Static network graph. Nodes are genes and edges are adjacencies between genes. 
# The thicker edge denotes higher adjacency (co-expression similarity) while larger node
# indicates higher gene connectivity (sum of a gene's adjacencies with all its direct 
# neighbors). The target gene is labeled by "_target".
network(ID="ENSGALG00000019846", data=se.sub.mat, adj.mod=adj.mod, adj.min=0, 
vertex.label.cex=1.5, vertex.cex=4, static=TRUE)

# Interactive network. The target gene ID is appended "_target".  
 network(ID="ENSGALG00000019846", data=se.sub.mat, adj.mod=adj.mod, static=FALSE) 

jianhaizhang/spatialHeatmap documentation built on April 21, 2024, 7:43 a.m.