| Type: | Package |
| Title: | Multiscale Clustering of Geometrical Network |
| Version: | 1.3.7 |
| Date: | 2018-08-30 |
| Author: | Won-Min Song, Bin Zhang |
| Maintainer: | Won-Min Song <wonmin1984@gmail.com> |
| Description: | Co-Expression Network Analysis by adopting network embedding technique. Song W.-M., Zhang B. (2015) Multiscale Embedded Gene Co-expression Network Analysis. PLoS Comput Biol 11(11): e1004574. <doi:10.1371/journal.pcbi.1004574>. |
| URL: | https://github.com/songw01/MEGENA |
| License: | GPL (≥ 3) |
| Imports: | Rcpp (≥ 0.11.3),Matrix (≥ 1.1-5),ggplot2 (≥ 1.0.0),reshape (≥ 0.8.5),fpc (≥ 2.1-11),cluster (≥ 2.0.7-1),ggrepel (≥ 0.5), ggraph (≥ 1.0.1) |
| Depends: | R (≥ 3.4.0),doParallel (≥ 1.0.11),foreach (≥ 1.4.4),igraph (≥ 1.2.1) |
| LinkingTo: | Rcpp,BH |
| Packaged: | 2018-09-10 15:10:45 UTC; songw01 |
| Suggests: | knitr, rmarkdown |
| VignetteBuilder: | knitr |
| NeedsCompilation: | yes |
| Repository: | CRAN |
| Date/Publication: | 2018-09-10 19:00:03 UTC |
co-expression network analysis
Description
construction of gene-gene interaction network and dissection into multi-scale functional modules, and network key drivers.
Details
| Package: | MEGENA |
| Type: | Package |
| Version: | 1.3 |
| Date: | 2015-12-18 |
| License: | GPL (>= 2) |
Multiscale Embedded Gene Co-expression Network Analysis (MEGENA)
Author(s)
Won-Min Song
Maintainer: Won-Min Song <won-min.song@mssm.edu>
References
Song W-M, Zhang B (2015) Multiscale Embedded Gene Co-expression Network Analysis. PLoS Comput Biol 11(11): e1004574. doi: 10.1371/journal.pcbi.1004574
MEGENA module summary
Description
Summarizes modules into a table.
Usage
MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 2500,
annot.table = NULL,symbol.col = NULL,id.col = NULL,
output.sig = TRUE)
Arguments
MEGENA.output |
A list object. The output from "do.MEGENA()". |
mod.pvalue |
module compactness significance p-value, to identify modules with significant compactness. |
hub.pvalue |
node degree significance p-value to identify nodes with significantly high degree. |
min.size |
minimum module size allowed to finalize in the summary output. |
max.size |
maximum module size allowed to finalize in the summary output. |
annot.table |
Default value is NULL, indicating no mapping is provided between node names to gene symbols. If provided, the mapping between node names (id.col) and gene symbol (symbol.col) are used. |
id.col |
column index of annot.table for node names. |
symbol.col |
column index of annot.table for gene symbols. |
output.sig |
Default value is TRUE, indicating significant modules are outputted. |
Details
output$module.table contains many important information including module hierarchy, as indicated by
Value
A list object with the components:
modules |
Final set of modules obtained upon apply mod.pvalue for significance, min.size and max.size for module size thresholding. |
mapped.modules |
gene symbol mapped modules when "annot.table" is provided. |
module.table |
data.frame object for module summary table. Columns include: id, module.size, module.parent, module.hub, module.scale and module.pvalue. |
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
## End(Not run)
Toy example data
Description
A portion of TCGA breast cancer data to test run MEGENA.
Usage
data(Sample_Expression)
Format
Contains a matrix object, "datExpr". Use data(Sample_Expression) to load.
Details
a gene expression matrix.
References
1. Song, W.M. and B. Zhang, Multiscale Embedded Gene Co-expression Network Analysis. PLoS Comput Biol, 2015. 11(11): p. e1004574.
PFN calculation
Description
main function to calculate PFN a ranked list of edge pairs
Usage
calculate.PFN(edgelist,max.skipEdges = NULL,maxENum = NULL,doPar = FALSE,
num.cores = NULL,keep.track = TRUE)
Arguments
edgelist |
three column edgelist: first two columns are topological edges, and the third column is the weight. Must be a data.frame object. |
max.skipEdges |
Maximum number of edges to be searched by planarity test without any inclusion to PFN. If set NULL, it will be automatically set to number of cores x 1000. It acts as a threhold to quicken PFN construction termination during PCP. |
maxENum |
maximum number of edges to include in final PFN. Default value is NULL, which invokes maximal number of edges allowed in planar network. |
doPar |
TRUE/FALSE logical variable to choose parallelization. |
num.cores |
number of cores to use in parallelization. |
keep.track |
If TRUE, pfg_el.RData will be created in working folder. This file can be used later for restart in case PFN construction did not finish successfully. Default is TRUE. |
Details
If doPar = TRUE, then num.cores are registered for PCP.
Value
output is three column edgelist data.frame, third column being the weight.
Author(s)
Won-Min Song
Examples
# test simplest case of planar network (a 3-clique).
a <- c(1,1,2);b <- c(2,3,3);w <- runif(3,0,1);
el <- cbind(a,b,w);el <- as.data.frame(el[order(el[,3],decreasing = TRUE),])
calculate.PFN(edgelist = el,max.skipEdges = Inf,doPar = FALSE,num.cores = NULL)
correlation calculation
Description
correlation analysis with FDR calculation
Usage
calculate.correlation(datExpr,doPerm = 100,doPar = FALSE,num.cores = 8,method = "pearson",
FDR.cutoff = 0.05,n.increment = 100,is.signed = FALSE,
output.permFDR = TRUE,output.corTable = TRUE,saveto = NULL)
Arguments
datExpr |
gene expression data matrix |
doPerm |
Number of permutations to perform. If |
doPar |
TRUE/FALSE logical variable to choose parallelization. Parallelization is utilized when BH FDR p-values are calculated for all pairs. |
num.cores |
number of cores to use in parallelization. |
method |
correlation method to be passed to |
FDR.cutoff |
FDR threshold to output final results of significant correlations. |
n.increment |
When permutation is utilized, 0 <= |rho| <= 1 is broken down into n.increment to map each |rho| cutoff to respective FDR. |
is.signed |
TRUE/FALSE to indicate using signed/unsigned correlation. |
output.permFDR, output.corTable |
TRUE/FALSE to choose to output permutation indices and FDR table. |
saveto |
folder to output results. |
Details
If doPar = TRUE, then num.cores are registered for PCP.
Value
output is three column edgelist data.frame, third column being the weight.
Author(s)
Won-Min Song
Examples
# test simplest case of planar network (a 3-clique).
data(Sample_Expression)
calculate.correlation(datExpr[1:100,],doPerm = 5)
Parallelized PFN computation
Description
PFN construction by parallelized edge screening.
Usage
compute.PFN.par(sortedEdge,Ng,maxENum,Njob,Ncore,max.skipEdges = NULL,
keep.track = TRUE,initial.links = NULL)
Arguments
sortedEdge |
3-column matrix for the input edgelist (e.g. - correlation pair list). Must be sorted by third column, which is usually weight vector. |
Ng |
integer. number of genes included in sortedEdge. |
maxENum |
Maximum number of edges to include in final PFN. The theoretical maximal number enforced by Euler's formula is 3(Ng-2). |
max.skipEdges |
Maximum number of edges to be counted before any valid edge to be included in PFN. This works as a termination condition to avoid exhaustive planarity testing over all edges provided in sortedEdge. |
Njob |
Number of edges to be passed to each core for parallelized edge screening. |
Ncore |
Number of cores to utilize. |
keep.track |
TRUE/FALSE logical. Indicate if the record of PFN construction is saved in temporary file "pfg_el.RData". Default is TRUE. |
initial.links |
If provided, PFN construction will restart by regarding these initial.links as already-built PFN. |
Details
This is parallelized implementation of PFN construction, where it is possible to re-capture PFN construction by providing already computed edgelist into initial.links. Although provivded, this function itself may require careful caution and users are encouraged to use more user-friendly "calculate.PFN()" instead.
Value
A 3-column matrices, where first two columns are integer indices for vertices, and third is the weight vector.
Author(s)
Won-Min Song
Toy example data
Description
A portion of TCGA breast cancer data to test run MEGENA.
Usage
Sample_Expression.RData
Format
Contains a matrix object, "datExpr". Use data(Sample_Expression) to load.
Details
a gene expression matrix.
References
1. Song, W.M. and B. Zhang, Multiscale Embedded Gene Co-expression Network Analysis. PLoS Comput Biol, 2015. 11(11): p. e1004574.
MEGENA clustering + MHA
Description
multiscale clustering analysis (MCA) and multiscale hub analysis (MHA) pipeline
Usage
do.MEGENA(g,
do.hubAnalysis = TRUE,
mod.pval = 0.05,hub.pval = 0.05,remove.unsig = TRUE,
min.size = 10,max.size = 2500,
doPar = FALSE,num.cores = 4,n.perm = 100,singleton.size = 3,
save.output = FALSE)
Arguments
g |
igraph object of PFN. |
do.hubAnalysis |
TRUE/FALSE indicating to perform multiscale hub analysis (MHA) in downstream. Default is TRUE. |
mod.pval |
cluster significance p-value threshold w.r.t random planar networks |
hub.pval |
hub significance p-value threshold w.r.t random planar networks |
remove.unsig |
TRUE/FALSE indicating to remove insignificant clusters in MHA. |
min.size |
minimum cluster size |
max.size |
maximum cluster size |
doPar |
TRUE/FALSE indicating parallelization usage |
num.cores |
number of cores to use in parallelization. |
n.perm |
number of permutations to calculate hub significance p-values/cluster significance p-values. |
singleton.size |
Minimum module size to regard as non-singleton module. Default is 3. |
save.output |
TRUE/FALSE to save outputs from each step of analysis |
Details
Performs MCA and MHA by taking PFN as input. Returns a list object containing clustering outputs, hub analysis outputs, and node summary table.
Value
A series of output files are written in wkdir. Major outputs are,
module.output |
outputs from MCA |
hub.output |
outputs from MHA |
node.summary |
node table summarizing clustering results. |
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
## End(Not run)
Draw sunburst plot showing MEGENA module hierarchy.
Description
Sunburst plot and colored heatmaps
Usage
draw_sunburst_wt_fill(module.df,
parent.col = "module.parent",id.col = "id",
min.angle = 5,
feat.col,
fill.type = "continuous",log.transform = TRUE,
fill.scale = NULL,
theme.adjust = NULL
)
Arguments
module.df |
A data.frame table summarizing module information. Must contain module parent and child relation for hierarchy visualization. |
parent.col |
Character object, name for the parent module column in module.df. |
id.col |
Character object for the module id column in module df. |
min.angle |
Minimum angle that rectangles in the sunburst are labeled with respective module id. |
feat.col |
Chracter object, for the feature column in module.df to color the heatmaps. |
fill.type |
continuous/discrete, is the variable numeric (continuous) or factor (discrete)? |
log.transform |
TRUE/FALSE. do log10 transform for p-values? |
fill.scale |
A ggplot object to specify heatmap coloring scheme. Permissible functions are: scale_fill_gradient,scale_fill_gradient2,scale_fill_gradientn,scale_fill_manual. |
theme.adjust |
A ggplot object to specify theme for plotting. |
Details
makes use of ggraph scheme to manipulate and draw sunburst plot in ggplot2 framework. fill.scale and theme.adjust provide flexibility to designate heatmap coloring schemes and figure aesthetics.
Value
ggplot object for the figure
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
# no coloring
sbobj = draw_sunburst_wt_fill(module.df = output.summary$module.table,
feat.col = NULL,id.col = "module.id",parent.col = "module.parent")
sbobj
# get some coloring (with log transform option)
mdf= output.summary$module.table
mdf$heat.pvalue = runif(nrow(mdf),0,0.1)
sbobj = draw_sunburst_wt_fill(module.df = mdf,feat.col = "heat.pvalue",log.transform = TRUE,
fill.type = "continuous",
fill.scale = scale_fill_gradient2(low = "white",mid = "white",high = "red",
midpoint = -log10(0.05),na.value = "white"),
id.col = "module.id",parent.col = "module.parent")
sbobj
# get discrete coloring done
mdf$category = factor(sample(x = c("A","B"),size = nrow(mdf),replace = TRUE))
sbobj = draw_sunburst_wt_fill(module.df = mdf,feat.col = "category",
fill.type = "discrete",
fill.scale = scale_fill_manual(values = c("A" = "red","B" = "blue")),
id.col = "module.id",parent.col = "module.parent")
sbobj
## End(Not run)
calculate module degree statistics based on random triangulation model via T1 and T2 moves.
Description
calculation of module p-values.
Usage
get.DegreeHubStatistic(subnetwork,n.perm = 100,doPar = FALSE,n.core = 4)
Arguments
subnetwork |
a planar network as an igraph object. |
n.perm |
number of random networks generated, constraint with number of links and nodes same to "subnetwork". |
doPar |
TRUE/FALSE to parallelize. |
n.core |
number of cores/threads to use. |
Details
Hub significance calculation functionality. Make sure that, if doPar = TRUE, register cores using registerDoParallel() from doParallel package.
Value
a data.frame table showing node-wise statistics.
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
out <- get.DegreeHubStatistic(subnetwork = g,n.perm = 100,doPar = FALSE,n.core = 4)
## End(Not run)
summarize hub information.
Description
hubs in different scales are summarized.
Usage
get.hub.summary(MEGENA.output)
Arguments
MEGENA.output |
A list object. The output from "do.MEGENA()". |
Details
returns a data.frame object
Value
A data.frame object with columns:
node |
hub gene node names |
S1, ... |
binary vector indicating hubs in each scale |
frequency |
number of scales that respective gene emerges as hub. |
scale.summary |
list of scales that respective gene as hub. |
Author(s)
Won-Min Song
Scale-thresholding of multiscale modules.
Description
obtain a discrete, disjoint clustering results from multiscale MEGENA modules for a given alpha value.
Usage
get.union.cut(module.output,alpha.cut,output.plot = T,
plotfname = "validModules_alpha",module.pval = 0.05,remove.unsig = T)
Arguments
module.output |
A direct output from "do.MEGENA". (i.e. MEGENA.output$module.output). |
alpha.cut |
Resolution parameter cut-off (i.e. alpha) value. alpha.cut = 1 corresponds to classical definition of "small-world" compactness. |
output.plot |
TRUE/FALSE to indicate outputting a .png file showing hierarchical structure with final outputted modules highlighted in red. |
plotfname |
.png file outputname. |
module.pval |
module significance p-value. |
remove.unsig |
TRUE/FALSE indicating to remove insignificant clusters. |
Details
Returns a list object where each entry is a module.
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
get.union.cut(module.output = MEGENA.output$module.output,alpha.cut = 1,
output.plot = FALSE,plotfname = NULL,module.pval = 0.05,remove.unsig = TRUE)
## End(Not run)
conversion of module list object to a data.frame table format
Description
Summarizes module hub/hierarchy/membership into a data.frame table format.
Usage
module_convert_to_table(MEGENA.output,mod.pval = 0.05,
hub.pval = 0.05,min.size = 10,max.size)
Arguments
MEGENA.output |
A list object. The output from "do.MEGENA()". |
mod.pval |
module compactness significance p-value, to identify modules with significant compactness. |
hub.pval |
node degree significance p-value to identify nodes with significantly high degree. |
min.size |
minimum module size allowed to finalize in the summary output. |
max.size |
maximum module size allowed to finalize in the summary output. |
Details
the resulting data.frame contains the following essential columns: id, module.parent and module. If the co-expression network bears significant hubs, it will additionally have node.degree (connectivity), node.strength (sum of edge weights) and is.hub column to supplement hub information.
Value
A data.frame with the columns:
id |
gene name |
module.parent |
parent module id |
module |
module name. |
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
module.df = module_convert_to_table(MEGENA.output,mod.pval = 0.05,
hub.pval = 0.05,min.size = 10,max.size)
head(module.df)
## End(Not run)
output gene signatures into .gmt file format
Description
An interface function to output .gmt format gene signature file.
Usage
output.geneSet.file(geneSet,outputfname)
Arguments
geneSet |
a list object |
outputfname |
output file name |
Details
Outputs each signature into a single line of lists in outputfname.
Author(s)
Won-Min Song
Boyer-Myvold Planarity test of a network
Description
wrapper function of _MEGENA_planaritytest. imports from Boost graph library, and test planarity of a network
Usage
planaritytest(N, rows, cols)
Arguments
N |
must be an integer. number of nodes in the network. |
rows |
first column of edgelist. a vector of integers. |
cols |
second column of edgelist. a vector of integers. |
Details
cbind(rows,cols) is equivalent to the two column edge list of the network. We assume that the network is undirected.
Value
TRUE/FALSE is returned to indicate planarity. (TRUE -> network is planar).
Author(s)
Won-Min Song
Examples
# test simplest case of planar network (a 3-clique).
planaritytest(as.integer(3),c(1,1,2),c(2,3,3))
Module plotting function.
Description
Extract subnetworks for modules and plot.
Usage
plot_module(output.summary,PFN,subset.module = NULL,col.names,
gene.set = NULL,color.code = "logFC",show.legend = TRUE,
label.hubs.only = TRUE,hubLabel.col = "red",hubLabel.sizeProp = 0.5,show.topn.hubs = 10,
node.sizeProp = 13,label.sizeProp = 13,label.scaleFactor = 10,label.alpha = 0.5,
layout = "kamada.kawai",output.plot = TRUE,out.dir = "modulePlot")
Arguments
output.summary |
output from summary function, "MEGENA.ModuleSummary". |
PFN |
igraph object retaining PFN topology. |
subset.module |
A character vector for list of module names to plot. Default = NULL plots all modules in output.summary. |
col.names |
a character vector for list of colors to be used for coloring children modules. |
gene.set |
A list object containing signatures for customized coloring of nodes in resulting network plot. |
color.code |
A character vector with matched length to "gene.set", to specify colors for each signature. |
label.hubs.only |
TRUE/FALSE to show labels for significant hub genes only, or all genes. Defauly is TRUE. |
hubLabel.col |
Label color for hubs. Default is "red" |
show.legend |
TRUE/FALSE for showing node legend on the bottom of the figure. |
hubLabel.sizeProp |
A multiplicative factor to adjust hub label sizes with respect to node size values. Default is 0.5 |
show.topn.hubs |
Maximal number of hubs to label on module subnetwork. Default is 10. |
node.sizeProp |
A multiplicative factor to adjust node sizes with respect to 90th percentile degree node size. Default is 13 |
label.sizeProp |
A multiplicative factor to adjust node label sizes with respect to 90th percentile degree node size. Default is 13 |
label.scaleFactor |
Overall scale factor to control the final size of node labels appearing in figure. Default is 10. |
label.alpha |
Transparency value ranging from 0 (transparent) to 1 (solid). Default is 0.5. |
layout |
Network layout algorithm to apply. Options are: "kamada.kawai", "fruchterman.reingold". |
output.plot |
logical value. output.plot = TRUE generates figure files under folder, "modulePlot". |
out.dir |
if output.plot = TRUE, then out.dir is created and resulting figures are exported to .png files to the folder. |
Details
Subnetwork plot functionality with application of "ggrepel" package for node labeling. The most effective way to control overall node label size is through label.scaleFactor.
Value
A list object holding ggplot objects for plotted modules.
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
library(MEGENA)
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
pnet.obj <- plot_module(output = output.summary,PFN = g,subset.module = "comp1_2",
layout = "kamada.kawai",label.hubs.only = FALSE,
gene.set = list("hub.set" = c("CD3E","CD2")),color.code = c("red"),
output.plot = FALSE,out.dir = "modulePlot",col.names = c("grey","grey","grey"),
hubLabel.col = "black",hubLabel.sizeProp = 1,show.topn.hubs = Inf,show.legend = TRUE)
pnet.obj
## End(Not run)
Plot module hierarchy
Description
visualized module hierarchical structure.
Usage
plot_module_hierarchy(module.table,plot.coord = NULL,
edge.color = "grey",node.color = "black",node.label.color = "black",
label.scaleFactor = 0.5,node.scaleFactor = 0.2,arrow.size = 0.015,
data.col = NULL,low.color = "blue",mid.color = "white",
high.color = "red",mid.value = 0.05)
Arguments
module.table |
output from MEGENA.ModuleSummary. Specifically $module.table component of the output. |
plot.coord |
Two column coordinate matrix. rownames must be labelled according to module.table$id. |
edge.color |
Edge color to be shown. |
node.color |
If data.col = NULL, node.color is used to color nodes in figure. |
node.label.color |
Node label color. |
label.scaleFactor |
scale number to adjust node label sizes. |
node.scaleFactor |
scale number to adjust node sizes. |
arrow.size |
scale number to arrow size. |
data.col |
A character to specify data vector to color nodes in module.table. |
low.color |
If data.col != NULL, color to be used in lower value spectrum. |
mid.color |
If data.col != NULL, color to be used in middle value spectrum. |
high.color |
If data.col != NULL, color to be used in high value spectrum. |
mid.value |
If data.col != NULL, value to define middle value spectrum. |
Details
Module hierarchy plotting functionality using ggplot2.
Value
A list containing output$hierarchy.obj = ggplot2 object, output$node.data = node attributes, output$edge.data = edge attributes.
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
data(Sample_Expression)
ijw <- calculate.correlation(datExpr,doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
module.table = output.summary$module.table
colnames(module.table)[1] <- "id"
output.obj <- plot_module_hierarchy(module.table = module.table,
label.scaleFactor = 0.15,arrow.size = 0.005,node.label.color = "blue")
print(output.obj[[1]])
## End(Not run)
subnetwork plotting functionality.
Description
A modification of plot_module() function for more general subnetwork plotting purpose.
Usage
plot_subgraph(module,hub = NULL,PFN,node.default.color = "black",
gene.set = NULL,color.code = "grey",show.legend = TRUE,
label.hubs.only = TRUE,hubLabel.col = "red",hubLabel.sizeProp = 0.5,show.topn.hubs = 10,
node.sizeProp = 13,label.sizeProp = 13,label.scaleFactor = 10,layout = "kamada.kawai")
Arguments
module |
A character vector containing gene names to be subsetted. |
hub |
If provided, genes in hub will be highlighted as triangles in resulting figure. |
PFN |
igraph object retaining PFN topology. |
node.default.color |
Default node colors for those that do not intersect with signatures in gene.set. |
gene.set |
A list object containing signatures for customized coloring of nodes in resulting network plot. |
color.code |
A character vector with matched length to "gene.set", to specify colors for each signature. |
show.legend |
TRUE/FALSE for showing node legend on the bottom of the figure. |
label.hubs.only |
TRUE/FALSE to show labels for significant hub genes only, or all genes. Defauly is TRUE. |
hubLabel.col |
Label color for hubs. Default is "red" |
hubLabel.sizeProp |
A multiplicative factor to adjust hub label sizes with respect to node size values. Default is 0.5 |
show.topn.hubs |
Maximal number of hubs to label on module subnetwork. Default is 10. |
node.sizeProp |
A multiplicative factor to adjust node sizes with respect to 90th percentile degree node size. Default is 13 |
label.sizeProp |
A multiplicative factor to adjust node label sizes with respect to 90th percentile degree node size. Default is 13 |
label.scaleFactor |
Overall scale factor to control the final size of node labels appearing in figure. Default is 10. |
layout |
Network layout algorithm to apply. Options are: "kamada.kawai", "fruchterman.reingold". |
Details
Subnetwork plot functionality with application of "ggrepel" package for node labeling. The most effective way to control overall node label size is through label.scaleFactor.
Value
A list object holding ggplot object and node annotation table.
Author(s)
Won-Min Song
Examples
## Not run:
rm(list = ls())
library(MEGENA)
data(Sample_Expression)
ijw <- calculate.correlation(datExpr[1:100,],doPerm = 2)
el <- calculate.PFN(ijw[,1:3])
g <- graph.data.frame(el,directed = FALSE)
MEGENA.output <- do.MEGENA(g = g,remove.unsig = FALSE,doPar = FALSE,n.perm = 10)
output.summary <- MEGENA.ModuleSummary(MEGENA.output,
mod.pvalue = 0.05,hub.pvalue = 0.05,
min.size = 10,max.size = 5000,
annot.table = NULL,id.col = NULL,symbol.col = NULL,
output.sig = TRUE)
pnet.obj <- plot_subgraph(module = output.summary$modules[[1]],
hub = c("CD3E","CD2"),PFN = g,node.default.color = "black",
gene.set = NULL,color.code = c("grey"),show.legend = TRUE,
label.hubs.only = TRUE,hubLabel.col = "red",hubLabel.sizeProp = 0.5,
show.topn.hubs = 10,node.sizeProp = 13,label.sizeProp = 13,
label.scaleFactor = 10,layout = "kamada.kawai")
# the plot
pnet.obj[[1]]
# the annotation
pnet.obj[[2]]
## End(Not run)
.gmt file reader function
Description
An interface function to read-in .gmt format gene signature file.
Usage
read.geneSet(geneSet.file)
Arguments
geneSet.file |
text file containing gene signatures in .gmt format |
Details
Each line of lists in geneset.file is a single set of signature.
Value
loads signatures into a list object.
Author(s)
Won-Min Song