npData: The npData Object

In ZachHunter/bvt: Visualization Tools for Bioconductor

The npData Object

Description

Understanding the npData objects returned by BVT functions

Details

An object of class npData is returned from all BCT plotting functions. The primary use of these is to allow for the graph to be saved so it can be plotted later using plot or a BVT plotting function. Modifications can also be made such as changing a bar plot to a box plot or changing any of the bvt_graphic_options or plotColors settings. These saved objects can also be passed to functions such as link{pick_points}, to further analyze the data or make new factor levels. These objects also have some useful information stored in them stored in the a named list with the sections: summary, stats, plotType, and options. A breakdown of each section is below.

npData$summary

This section is a containing the summary data and descriptive statistics for the plot. The data is broken down by group, subgroup, and highlight factor levels if active and reflects the essential statistics of the plot. A bar plot might have mean and 2x standard error of the mean listed by factor level while a box plot would have the quartile values. These summary statistics will change with aggFun and errFun settings. Not yet implemented for scatter plots generated with geneScatter.

npData$stats

This section is a named list containing an overall p-value where applicable or NA (npData$stats$p.value), the name of the statistical test performed (e.g. pairwise.wilcox.test; npData$stats$test), and the object returned by that test (npData$stats$results). In the case of geneScatter, this section will provide a linear model from lm, and the output of cor.test for each trend line present.

npData$plotType

This is a text string that corresponds to the plot type. Valid options include box, dot, bar, density, and scatter.

npData$options

This section has all the options and data needed to regenerate the plot. The three parts worth noting are npData$options$x which as all the numeric data as a vector or data.frame, npData$options$by which contains a factor or a data.frame of factors used to subset or highlight the data, and npData$options$xypos which contains the actual x and y coordinates for all points on the graph including those subject to jitter, distribution, or beeswarm settings from the pointMethod option. Note that x and by are the original data with NA still present while xypos is post filtering and may be smaller and has the corresponding row numbers of the original x and by data to allow for mapping between the two systems. This can be useful when making custom modifications to a graph as shown in the example below.

See Also

genePlot, isoPlot, geneScatter

Examples

data(iris)
#plotting a saved graph with different options

a<-genePlot(t(iris[,1:4]), group=iris$Species, plotType="bar", main="npData Example")
#Note that only the first two columns of data are used  below
#as this is the maximum allowed for the \emph{density plots}.
plot(a, plotType="density", theme=npColorTheme)

#Example showing the use of npData$options$xypos with a toy data set

salesData<-data.frame(Q1=c(1,4,2,2), Q2=c(2,3,4,2), Q3=c(4,4,2,3))
stores<-factor(c("Store 1","Store 2","Store 3","Store 4"))

#Transposition is necessary since expression data is typical the transpose of tidy data
b<-genePlot(t(salesData), highlight=stores, plotType="dot",
   theme=npGGTheme, pointLaneWidth=2,
   plotColors=list(lines=setAlpha("black",.5)),
   legendSize=1, main="Toy Sales Data",
   pointMethod="jitter", ylab="Sales Volume (Millions)")

#Drawing connecting lines
for(i in 1:4){
 lines(
    x=b$options$xypos[b$options$xypos$ID==i,"x"],
    y=b$options$xypos[b$options$xypos$ID==i,"y"],
    col=npGGTheme$plotColors$points[i]
 )
}

ZachHunter/bvt documentation built on Sept. 18, 2024, 3:12 p.m.