obsPlot: Function obsPlot

In jrthompson54/DGE.Tools2: A function library to facilitate Differential Gene Expression Analysis

Description

Provides a summary plot for each observation (gene), showing data for each experiment group. The plot can optionally include one or more of the following layers: boxplot, violin plot, individual points and/or mean of all points. The layers are built up in the order listed with user settable transparency, colors etc. By default, the Boxplot, Point and Mean layers are active. Also, by default, the plots are faceted. Facet plot can be turned off to return a list of individual ggplot graphics for each gene.

Usage

obsPlot(
  data,
  block,
  blockOrder,
  obsNames,
  sampNames,
  plotBy = "Gene",
  valType = "Log2CPM",
  boxLayer = TRUE,
  violinLayer = FALSE,
  pointLayer = TRUE,
  meanLayer = TRUE,
  xlab = NULL,
  ylab = NULL,
  title = NULL,
  boxColor = "grey30",
  boxFill = "deepskyblue3",
  boxAlpha = 0.5,
  boxNotch = FALSE,
  boxNotchWidth = 0.8,
  violinColor = "grey30",
  violinFill = "goldenrod1",
  ViolinAlpha = 0.5,
  pointColor = "grey30",
  pointFill = "dodgerblue4",
  pointShape = 21,
  pointAlpha = 1,
  pointSize = 2,
  pointJitter = 0,
  meanColor = "red2",
  meanFill = "goldenrod1",
  meanShape = 22,
  meanAlpha = 0.7,
  meanSize = 3,
  legenPosition = "right",
  baseFontSize = 12,
  themeStyle = "grey",
  facet = TRUE,
  facetCol = NULL,
  xAngle = 30,
  scales = "free_y",
  returnPlotDat = FALSE,
  debug = FALSE
)

Arguments

data	matrix or dataframe of whatever data you want to plot with samples in columns and observations (genes) in rows (with rownames and colnames). Bug: Currently requires at least 2 rows of data.
block	Character or numeric vector. Determines which samples belong to the same group. Must be same length as ncol(data). Assign the same value to each member of a group. Note that the block values are used to label the X Axis to identify the groups. Thus short pneumonic labels are useful here.
blockOrder	Character vector. Use this to explicitly set the display order of the blocks in the plots. blockOrder should contain unique(block) in the order you wish to items to be arranged along the X-axis. (Optional: Default = unique(block))
obsNames	Character vector of row (observation) names to use instead of actual rownames. (Default = NULL)
sampNames	Character vector of column (sample) names to use instead of actual colnames. (Default = NULL)
plotBy	Name for the type of observation being plotted (Default = "Gene")
valType	name for the Value being plotted. (Default = "Log2CPM")
boxLayer	Adds a boxplot layer (Default = TRUE)
violinLayer	Adds a violin layer (Default = FALSE)
pointLayer	Adds a point layer (Default = True)
meanLayer	Adds a mean layer (Default = True)
xlab	X axis label (defaults to "Block")
ylab	Y axis label (defaults to valType)
title	Plot title (optional; Defaults = NULL)
boxColor	Color for the boxplot layer (Default = "grey30")
boxFill	Fill Color for the boxplot layer (Default = "deepskyblue3")
boxAlpha	Transparency for the box layer (Default = 0.5)
boxNotch	turn on/off confidence interval notches on boxplots (Default = FALSE)
boxNotchWidth	Set the width of boxnotches (0-1) (Default = 0.8)
violinColor	Color for the violin layer (Default = "grey30")
violinFill	Fill Color for the violin (Default = "yellow")
pointColor	Color for the point layer (Default = "grey30")
pointFill	Fill color for the point layer (Default = "dodgerblue4")
pointShape	Shape for the point layer (Default = 21; fillable circle)
pointAlpha	Transparency for the box layer (Default = 1)
pointSize	Size of the points (Default = 4)
pointJitter	Amount to jitter the points (Default = 0) Try 0.2 if you have alot of points.
meanColor	Color for the mean layer (Default = "red2")
meanFill	Fill color for the mean layer (Default = "goldenrod1")
meanShape	Shape for the mean layer (Default = 21; fillable circle)
meanAlpha	Transparency for the mean layer (Default = 0.7)
meanSize	Size of the mean points (Default = 3)
baseFontSize	The smallest size font in the figure in points. (Default = 12)
themeStyle	"bw" or "grey" which correspond to theme_bw or theme_grey respectively. Default = bw"
facet	Specifies whether to facet (TRUE) or print individual plots (FALSE) (Default = TRUE)
xAngle	Angle to set the sample labels on the Xaxis. (Default = 30; Range = 0-90)
scales	Specify same scales or independent scales for each subplot (Default = "free_y"; Allowed values: "fixed", "free_x", "free_y", "free")
returnPlotDat	Returns the dataframe used for the plot as a list member (default=FALSE)
debug	Used to open a breakpoint just before the melt step (default=FALSE)
violinAlpha	Transparency for the box layer (Default = 0.5)
facetRow	Explicitly set the number of Rows for the facet plot. Default behavior will automatically set the columns. (Default = NULL)

Details

Input is a dataframe or matrix of observations (rows; usually genes) by Samples (columns) and requires rownames to identify the observations (genes). A Block vector is required to define which Samples (columns) belong to the same group and generate a summary plot by group with a separate plot for each observation (gene).

Rownames are required and all the data columns should be numeric. Each observation (gene) generates a separate plot, so you should pass a smallish list of genes unless you want alot of output. By default the plot is faceted. You should turn facet off if you have more than ~25 genes to plot. Colnames and Rownames will be used to label the plots by default but custom col/row names can also be supplied as separate vectors.

Value

ggplot If Facet=TRUE (default) returns a facetted ggplot object. If facet=FALSE or returnPlotDat=TRUE, returns a list of ggplot objects indexed by observation (gene) names. If returnPlotDat=TRUE, the last element of the list is the dataframe used to generate the plot.

Author(s)

John Thompson, [e-mail]

Examples

   Simple faceted plot with custom title

   #get Log2CPM from an DGEobj object
   dgeObj = readRDS("MyDGEobj.RDS")
   dgelist <- getItem(dgeObj, "DGEList")
   Log2CPM <- cpm(dgelist, log=TRUE)
   genes = Log2CPM[1:12,] #first dozen genes
   #define six treatment groups in triplicate
   MyBlock = c(1,1,1,2,2,2,3,3,3,4,4,4,5,5,5,6,6,6)
   MyPlot = obsPlot(genes, MyBlock, title = "Plot Title")

jrthompson54/DGE.Tools2 documentation built on May 12, 2021, 8:47 p.m.