xGGnetwork: Function to visualise an igraph object using ggnetwork

In Pi: Leveraging Genetic Evidence to Prioritise Drug Targets at the Gene and Pathway Level

Description

xGGnetwork is supposed to visualise an igraph object using ggnetwork.

Usage

xGGnetwork(
g,
node.label = NULL,
label.wrap.width = NULL,
label.wrap.lineheight = 0.8,
node.label.size = NULL,
node.label.fontface = "plain",
node.label.color = "darkblue",
node.label.alpha = 0.9,
node.label.padding = 1,
node.label.arrow = 0.01,
node.label.force = 1,
node.shape = 19,
node.shape.title = NULL,
node.xcoord = NULL,
node.ycoord = NULL,
node.color = NULL,
node.color.title = NULL,
colormap = "grey-orange-darkred",
ncolors = 64,
zlim = NULL,
na.color = "grey80",
node.color.alpha = 1,
node.size = NULL,
node.size.title = NULL,
node.size.range = c(1, 4),
slim = NULL,
title = "",
edge.size = 0.5,
edge.color = "black",
edge.color.alpha = 0.5,
edge.curve = 0.1,
edge.arrow = 2,
edge.arrow.gap = 0.02,
ncolumns = NULL
)

Arguments

g	an object of class "igraph". For an advanced use, it can be a list of igraph objects; in this case, multiple panels will be shown (particularly useful when visualising the same network but color-coded differently)
node.label	either a vector labelling nodes or a character specifying which node attribute used for the labelling. If NULL (by default), no node labelling
label.wrap.width	a positive integer specifying wrap width of node labelling
label.wrap.lineheight	line height spacing for text in ggplot. By default it is 0.8
node.label.size	a character specifying which node attribute used for node label size
node.label.fontface	a character specifying which node attribute used for node label fontface ('plain', 'bold', 'italic', 'bold.italic')
node.label.color	a character specifying which node attribute used for the node label color
node.label.alpha	the 0-1 value specifying transparency of node labelling
node.label.padding	the padding around the labeled node
node.label.arrow	the arrow pointing to the labeled node
node.label.force	the repelling force between overlapping labels
node.shape	an integer specifying node shape or a character specifying which node attribute used for the node shape (no matter whether it is numeric or character)
node.shape.title	a character specifying the title for node shaping
node.xcoord	a vector specifying x coordinates. If NULL, it will be created using igraph::layout_as_tree
node.ycoord	a vector specifying y coordinates. If NULL, it will be created using igraph::layout_as_tree
node.color	a character specifying which node attribute used for node coloring
node.color.title	a character specifying the title for node coloring
colormap	short name for the colormap. It can be one of "jet" (jet colormap), "bwr" (blue-white-red colormap), "gbr" (green-black-red colormap), "wyr" (white-yellow-red colormap), "br" (black-red colormap), "yr" (yellow-red colormap), "wb" (white-black colormap), "rainbow" (rainbow colormap, that is, red-yellow-green-cyan-blue-magenta), and "ggplot2" (emulating ggplot2 default color palette). Alternatively, any hyphen-separated HTML color names, e.g. "lightyellow-orange" (by default), "blue-black-yellow", "royalblue-white-sandybrown", "darkgreen-white-darkviolet". A list of standard color names can be found in http://html-color-codes.info/color-names
ncolors	the number of colors specified over the colormap
zlim	the minimum and maximum values for which colors should be plotted
na.color	the color for NAs. By default, it is 'grey80'
node.color.alpha	the 0-1 value specifying transparency of node colors
node.size	either a vector specifying node size or a character specifying which node attribute used for the node size
node.size.title	a character specifying the title for node sizing
node.size.range	the range of actual node size. Can be two values (range) or a value (fixed size)
slim	the minimum and maximum values for which sizes should be plotted
title	a character specifying the title for the plot
edge.size	a numeric value specifying the edge size. By default, it is 0.5. It can be a character specifying which edge attribute defining the edge size (though without the legend)
edge.color	a character specifying the edge color. By default, it is "black". It can be a character specifying which edge attribute defining the edge color (though
edge.color.alpha	the 0-1 value specifying transparency of edge color. By default, it is 0.5. It can be a character specifying which edge attribute defining the transparency of edge color (though without the legend)
edge.curve	a numeric value specifying the edge curve. 0 for the straight line
edge.arrow	a numeric value specifying the edge arrow. By default, it is 2
edge.arrow.gap	a gap between the arrow and the node
ncolumns	an integer specifying the number of columns for facet_wrap. By defaul, it is NULL (decided on according to the number of groups that will be visualised)

Value

a ggplot object, appended with 'data_nodes' and 'data_edges'

Note

none

See Also

xGGnetwork

Examples

RData.location <- "http://galahad.well.ox.ac.uk/bigdata"
## Not run: 
###########################
# load REACTOME
# restricted to Immune System ('R-HSA-168256') or Signal Transduction ('R-HSA-162582')
g <- xRDataLoader(RData.customised='ig.REACTOME',
RData.location=RData.location)
neighs.out <- igraph::neighborhood(g, order=vcount(g),
nodes="R-HSA-168256", mode="out")
nodeInduced <- V(g)[unique(unlist(neighs.out))]$name
ig <- igraph::induced.subgraph(g, vids=nodeInduced)

# visualise the graph with vertices being color-coded
V(ig)$degree <- igraph::degree(ig)
gp <- xGGnetwork(g=ig, node.label='term_id', label.wrap.width=30,
node.label.size=2, node.label.color='black', node.label.alpha=0.8,
node.label.padding=0, node.label.arrow=0, node.label.force=1,
node.shape=19, node.xcoord='xcoord', node.ycoord='ycoord',
node.color='degree', node.color.title='Degree',
colormap='grey-orange-darkred', ncolors=64, zlim=c(0,10),
node.size.range=3,
edge.color="black",edge.color.alpha=0.3,edge.curve=0.05,edge.arrow.gap=0.02,
title='')
# advanced use: visualise the list of graphs
ls_ig <- list(ig, ig)
gp <- xGGnetwork(g=ls_ig, node.label='term_id', label.wrap.width=30,
node.label.size=2, node.label.color='black', node.label.alpha=0.8,
node.label.padding=0, node.label.arrow=0, node.label.force=1,
node.shape=19, node.xcoord='xcoord', node.ycoord='ycoord',
node.color='degree', node.color.title='Degree',
colormap='grey-orange-darkred', ncolors=64, zlim=c(0,10),
node.size.range=3,
edge.color="black",edge.color.alpha=0.3,edge.curve=0.05,edge.arrow.gap=0.02,
title='')

###########################
# load PhasedTargets
# restricted to disease ('EFO:0000408') or immune system disease ('EFO:0000540')
g <- xRDataLoader(RData.customised='ig.PhasedTargets',
RData.location=RData.location)
neighs.out <- igraph::neighborhood(g, order=vcount(g),
nodes="EFO:0000408", mode="out")
nodeInduced <- V(g)[unique(unlist(neighs.out))]$name
ig <- igraph::induced.subgraph(g, vids=nodeInduced)

# append with the number of approved and phased targets
dag <- ig
V(dag)$num_approved <- sapply(V(ig)$max_phase,function(x)
sum(x$max_phase>=4))
V(dag)$num_phased <- sapply(V(ig)$max_phase,function(x)
sum(x$max_phase>=0))
# keep nodes with num_approved >=20
dag_ig <- igraph::induced.subgraph(dag,
vids=which(V(dag)$num_approved>=20))
# (optional) further restricted to the direct children of the root
root <- dnet::dDAGroot(dag_ig)
neighs.out <- igraph::neighborhood(dag_ig, order=1, nodes=root,
mode="out")
nodeInduced <- V(dag_ig)[unique(unlist(neighs.out))]$name
dag_ig <- igraph::induced.subgraph(dag_ig, vids=nodeInduced)
# nodes colored by num_approved
V(dag_ig)$node_color <- log2(V(dag_ig)$num_approved)
glayout <- igraph::layout_with_kk(dag_ig)
V(dag_ig)$xcoord <- glayout[,1]
V(dag_ig)$ycoord <- glayout[,2]
gp <- xGGnetwork(g=dag_ig, node.label='term_name', label.wrap.width=30,
node.label.size=2, node.label.color='black', node.label.alpha=0.9,
node.label.padding=0, node.label.arrow=0, node.label.force=0.5,
node.shape=19, node.xcoord='xcoord', node.ycoord='ycoord',
node.color='node_color', node.color.title='Approved\n(log2-scale)',
colormap='ggplot2.top', ncolors=64, node.size.range=3,
edge.color="orange",edge.color.alpha=0.5,edge.curve=0.05,edge.arrow.gap=0.02,
title='')

###########################
# visualise gene network
glayout <- igraph::layout_with_kk(g)
V(g)$xcoord <- glayout[,1]
V(g)$ycoord <- glayout[,2]
V(g)$degree <- igraph::degree(g)
gp <- xGGnetwork(g=g, node.label='name', node.label.size=2,
node.label.color='black', node.label.alpha=0.8, node.label.padding=0,
node.label.arrow=0, node.label.force=0.01, node.shape=19,
node.xcoord='xcoord', node.ycoord='ycoord', node.color='priority',
node.color.title='5-star\nrating', colormap='yellow-red', ncolors=64,
zlim=c(0,5), node.size='degree', node.size.title='Degree', slim=c(0,5),
edge.color="orange",edge.color.alpha=0.5,edge.curve=0,edge.arrow.gap=0.025,
title='')
gp_rating <- xGGnetwork(g=g, node.label='name', node.label.size=2,
node.label.color='black', node.label.alpha=0.8, node.label.padding=0.1,
node.label.arrow=0, node.label.force=0.01, node.shape=19,
node.xcoord='xcoord', node.ycoord='ycoord', node.color='priority',
node.color.title='5-star\nrating', colormap='white-yellow-red',
ncolors=64, zlim=c(0,5), node.size.range=5,
edge.color="orange",edge.color.alpha=0.3,edge.curve=0,edge.arrow.gap=0.02,
title='')

###########################
# use edge weight to color/size edges (without legends)
# edge color
#E(g)$color <- xColormap(colormap='RdYlBu', data=E(g)$weight)
#E(g)$size <- (E(g)$weight - min(E(g)$weight)) / (max(E(g)$weight) - min(E(g)$weight))
e.color <- subset(gp$data, !is.na(na.y))$e.color
gp + ggnetwork::geom_edges(color=e.color, show.legend=FALSE)
# edge size/thickness
e.size <- subset(gp$data, !is.na(na.y))$e.size
gp + ggnetwork::geom_edges(size=e.size, show.legend=FALSE)

## End(Not run)

Pi documentation built on Nov. 29, 2021, 3 p.m.