plot.gnbp and plot.dbn can be used to plot networks in which evidence has been absorbed and propagated. The beliefs or Jeffrey's signed information for conditional gaussian and phenotypic states and Fold Changes for discrete bayesian networks is mapped onto the network. plot.gpfit and plot.dbfit are plot methods for objects of class "gpfit" and "dbnfit" respectively.
"plot"(x, y="JSI",ncol = 1, col.palette,col.length = 100, fontsize=14, fontcolor="black",...)
"plot"(x, y="state",ncol = 1, col.palette,col.length = 100, fontsize=14, fontcolor="black",...)
"plot"(x, fontsize=14, fontcolor="black",...)
"plot"(x, fontsize=14, fontcolor="black",...)plot.gnbp), "dbn" (plot.dbn), "gpfit" (plot.gpfit) or "dbnfit" (plot.dbnfit)
"JSI" (default) or "belief" for Conditional Gaussian network and "FC" or "state" for Discrete Bayesian networks. plot.dbn plots only discrete bayesian networks. Note that y will be ignored for plot.dbnfit and gpfit
"JSI","belief" and "FC" a list of 6 elements specifying colors for colormap.All 6 elements should be character strings specifying the colour for
pos_high= high end of gradient of positive values (default = "red" (JSI, belief), "darkmagenta",(FC))
pos_low=low end of gradient of positive values (default = "wheat1" (JSI, belief), "palegoldenrod",(FC))
neg_high=high end of gradient of positive values (default = "cyan" (JSI, belief), "palegoldenrod",(FC))
neg_low=low end of gradient of positive values (default = "blue" (JSI, belief), "gold2",(FC))
dsep_col= d-separated nodes (default = "white")
qtl_col= discrete nodes (QTLs) (default = "grey")
node_abs_col= nodes for which evidence has been absorbed (default = "palegreen2")For "state", a list of 4 elements specifying colors for colormap should be specified. All 4 elements should be character strings specifying the colour for
col_nodes- a vector of colors for phenotype states should be specified. The length of the vector should be equal to the maximum number of phenotype states possible.
dsep_col= d-separated nodes (default = "white")
qtl_col= discrete nodes (QTLs) (default = "grey")
node_abs_col= nodes for which evidence has been absorbed (default = "palegreen2")
"belief", "JSI" and "FC". For "state", this argument will be ignored.
plot. These will be ignoredx is invisibly returned
plot.gpfit and plot.dbnfit are generic plot methods for objects of class"gpfit" and "dbnfit" respectively that are output from the fit-methods. These are networks in which evidence has not been absorbed and propagated. plot.gnbp and plot.dbn are generic plot methods for objects of class "gnbp" and "dbn" that are outputs from the absorb methods. These functions plot the genotype-phenotype networks in which evidence has been absorbed and propagated and maps the quantitative system wide effects on the network. Both conditional gaussian and discrete bayesian networks are supported. Users can specify the colormap options such as end colors for the positive and negative gradients and the resolution of the colormap. The default node shapes are "ellipse" for the phenotype nodes and "box" for genotype nodes.The d-separated nodes are white while the colored nodes are d-connected, with the color indicating the strength and direction of change. The node for which evidence is absorbed is colored green (default color).
absorb.gnbp, absorb.dbn, fit.gnbp, fit.dbn
## Fit, absorb and plot a genotype-phenotype network
data(mouse)
## Not run:
# mouse.cgbn<-fit.gnbp(mouse[,1:5],mouse[,6:19],alpha=0.1)
# plot(mouse.cgbn)
# mouse.cgbn.abs<-absorb.gnbp(mouse.cgbn,node="Tlr12",evidence=matrix(-0.99))
# plot(mouse.cgbn.abs)
# ## End(Not run)
Run the code above in your browser using DataLab