plot.GGInetwork: Plot a Gene-Gene Interaction Study with an heatmap-like...
In GeneGeneInteR: Tools for Testing Gene-Gene Interaction at the Gene Level

Description Usage Arguments Details Value See Also Examples

GGI.plot is a graphical function that allow a heatmap-like or a network-like visualization of a Gene-Gene Interaction study based on a set of genes.

## S3 method for class 'GGInetwork'
plot(x,method=c("heatmap","network"),threshold=NULL, 
col=c("#D6604D","#104E8B"),colbar.width=0.15, title=NULL,hclust.order=FALSE, 
use.log=FALSE,NA.col="#D3D3D3",draw.pvals=NULL, 
draw.names=NULL,interact=FALSE, 
method.adjust=c("none","holm","hochberg","hommel","Bonferroni","BH","BY","fdr"), 
genes=seq_len(ncol(x$p.value)), plot.nointer=TRUE, ...)

`x`	`GGInetwork`. Given as output of `GGI`.
`method`	Output graph ("heatmap" for heatmap-like, "network" for network-like). Default is `heatmap`.
`threshold`	A numeric between 0 and 1. All p-value strictly greater than that `threshold` are distinctly colored (See `NA.col`). Default is `NULL` corresponding to no threshold for applied for an heatmap graph and a 0.05 threshold applied for the network representation.
`method.adjust`	correction method for multiple testing as proposed in the `p.adjust` function. By default, no correction is applied to the p-values.
`genes`	Numeric vector allowing a selection of the genes that will be included in the relations. Default is set to all genes.
`col`	(Only for `method=heatmap`.) String vector. Marker colors to be used for the gradient. The first element of the vector is the value for 0 and the last is for 1. If only one value is parsed, that color is used for 0 and white is automatically used for 1. Any value compatible with `colorRampPalette` function.
`colbar.width`	(Only for `method=heatmap`.) A positive number describing the gradient bar width ratio. That number is used to keep the gradient bar's width steady as the size of `GGI` increases.
`title`	(Only for `method=heatmap`.) A string used as the plot title. If left as NULL, a generic name is generated.
`hclust.order`	(Only for `method=heatmap`.) A boolean. Should a hierachical clustering procedure be performed on `GGI` to order the matrix ?
`use.log`	(Only for `method=heatmap`.) A boolean. Should the p-values be reported in -log10? FALSE by default.
`NA.col`	(Only for `method=heatmap`.) A string. The color used when a p-value is strictly greater than `threshold`.
`draw.pvals`	(Only for `method=heatmap`.) Either NULL or a boolean. Should p-values be plotted ? Disabled when the size of `GGI` exceeds 1515*.
`draw.names`	(Only for `method=heatmap`.) Either NULL or a boolean. Should genes' names be plotted on matrix margins? Disabled when the size of `GGI` exceeds 2525*.
`interact`	(Only for `method=heatmap`.) A boolean. Should the plot be clickable ? (See Details for more information). Disabled when open R session is not interactive.
`plot.nointer`	(Only for `method=network`.) A boolean. Set TRUE if the genes with no interaction should be plotted.
`...`	further arguments passed to or from other methods.

If method=heatmap, this function draw the upper half of a Gene-Gene Interaction results matrix without its diagonal. A gradient is created from 0 to 1 (by default from crimson to white) and the matrix cells are colored according to the corresponding p-value.

By default, when draw.pvals==NULL and draw.names==NULL, p-values and names are drawn to make matrix reading easier, but in case parameter GGI is large, p-values (and eventually gene names as GGI grows bigger) are not drawn. In that case, the default behavior of the function is to start an interactive process where user can click on a cell of interest to open a tooltip displaying which genes are involved in selected interaction and the p-value of the interaction test. Tooltips can be closed if user clicks anywhere else than on a cell. This process stops when the user presses the escape button (or terminates the locator procedure in general) or when the user clicks on any place other than a cell when no tooltip window is open.

To improve plot clarity, user may set a threshold above which cells are colored with a distinct color. By default, threshold is set to 1 and no cell is colored differently (as values must be strictly above the threshold).

If method=network, this function plots a graph representing the significative interactions between genes of a Gene-Gene Interaction study.

The form of the value returned by plot depends on the class of its argument. See Details.

GGI

## Not run: 
## Dataset is included in the package
ped <- system.file("extdata/example.ped", package="GeneGeneInteR")
info <- system.file("extdata/example.info", package="GeneGeneInteR")
posi <- system.file("extdata/example.txt", package="GeneGeneInteR")

## Importation of the genotypes
data.imported <- importFile(file=ped, snps=info, pos=posi, pos.sep="\t")
## Filtering of the data: SNPs with MAF < 0.05 or p.value for HWE < 1e-3 or 
## call rate < 0.9 are removed. 
data.scour <- snpMatrixScour(snpX=data.imported$snpX,genes.info=data.imported$genes.info,
                              min.maf=0.05,min.eq=1e-3,call.rate=0.9)
## Imputation of the missing genotypes
data.imputed <- imputeSnpMatrix(data.scour$snpX, genes.info = data.scour$genes.info)

## Importation of the phenotype
resp <- system.file("extdata/response.txt", package="GeneGeneInteR")
Y  <- read.csv(resp, header=FALSE)

## plot of the interaction between the 17 genes with the CLD method -- can take a few minutes
GGI.res <- GGI(Y=Y, snpX=dta$snpX, genes.info=dta$genes.info,method="CLD")
plot(GGI.res,threshold=0.05)

## Selection of 12 genes among 17
data.select <- selectSnps(data.imputed$snpX, data.imputed$genes.info, c("bub3","CDSN","Gc","GLRX",
                          "PADI1","PADI2","PADI4","PADI6","PRKD3","PSORS1C1","SERPINA1","SORBS1"))
GGI.res <- GGI(Y=Y, snpX=data.select$snpX, genes.info=data.select$genes.info,method="PCA")

## End(Not run)
## Equivalent importation of the GGI.res object
load(system.file("extdata/GGIRes.Rdata", package="GeneGeneInteR"))


## Plot of the results with default values
plot(GGI.res)

## Plot of the results with a threshold and an ordering of the genes. 
## Default method is an heatmap-like representation
plot(GGI.res,threshold=0.1,hclust.order=TRUE)

## Example of network with default threshold 0.05
plot(GGI.res,method="network")

## Example of network with threshold 0.01 where genes with no interaction are not plotted 
# (plot.nointer=FALSE)
plot(GGI.res,method="network",threshold=0.1,plot.nointer=FALSE)