Make a boxplot

Takes a data.frame and creates a boxplot

Usage

create.boxplot(
  formula,
  data,
  filename = NULL,
  main = NULL,
  main.just = 'center',
  main.x = 0.5,
  main.y = 0.5,
  main.cex = 3,
  add.stripplot = FALSE,
  jitter.factor = 1,
  jitter.amount = NULL,
  points.pch = 19,
  points.col = 'darkgrey',
  points.cex = 0.5,
  points.alpha = 1,
  abline.h = NULL,
  abline.v = NULL,
  abline.lty = NULL,
  abline.lwd = NULL,
  abline.col = 'black',
  add.rectangle = FALSE,
  xleft.rectangle = NULL,
  ybottom.rectangle = NULL,
  xright.rectangle = NULL,
  ytop.rectangle = NULL,
  col.rectangle = 'transparent',
  alpha.rectangle = 1,
  box.ratio = 1,
  col = 'transparent',
  alpha = 1,
  border.col = 'black',
  symbol.cex = 0.8,
  lwd = 1,
  outliers = TRUE,
  sample.order = 'none',
  order.by = 'median',
  xlab.label = tail(sub('~', '', formula[-2]), 1),
  ylab.label = tail(sub('~', '', formula[-3]), 1),
  xlab.cex = 2,
  ylab.cex = 2,
  xlab.col = 'black',
  ylab.col = 'black',
  xlab.top.label = NULL,
  xlab.top.cex = 2,
  xlab.top.col = 'black',
  xlab.top.just = 'center',
  xlab.top.x = 0.5,
  xlab.top.y = 0,
  xlimits = NULL,
  ylimits = NULL,
  xat = TRUE,
  yat = TRUE,
  xaxis.lab = TRUE,
  yaxis.lab = TRUE,
  xaxis.cex = 1.5,
  yaxis.cex = 1.5,
  xaxis.col = 'black',
  yaxis.col = 'black',
  xaxis.fontface = 'bold',
  yaxis.fontface = 'bold',
  xaxis.rot = 0,
  yaxis.rot = 0,
  xaxis.tck = c(1,0),
  yaxis.tck = 1,
  layout = NULL,
  as.table = FALSE,
  x.spacing = 0,
  y.spacing = 0,
  x.relation = 'same',
  y.relation = 'same',
  top.padding = 0.5,
  bottom.padding = 2,
  right.padding = 1,
  left.padding = 2,
  ylab.axis.padding = 0,
  add.text = FALSE,
  text.labels = NULL,
  text.x = NULL,
  text.y = NULL,
  text.anchor = 'centre',
  text.col = 'black',
  text.cex = 1,
  text.fontface = 'bold',
  key = NULL,
  legend = NULL,
  strip.col = 'white',
  strip.cex = 1,
  strip.fontface = 'bold',
  line.func = NULL,
  line.from = 0,
  line.to = 0,
  line.col = 'transparent',
  line.infront = TRUE,
  height = 6,
  width = 6,
  size.units = 'in',
  resolution = 1600,
  enable.warnings = FALSE,
  description = 'Created with BoutrosLab.plotting.general',
  style = 'BoutrosLab',
  preload.default = 'custom',
        use.legacy.settings = FALSE,
  disable.factor.sorting = FALSE
  );

Arguments

formula: The formula used to extract the x & y components from the data-frame. Transforming data within formula is not compatible with automatic scaling with `xat` or `yat`.
data: The data-frame to plot
filename: Filename for tiff output, or if NULL returns the trellis object itself
main: The main title for the plot (space is reclaimed if NULL)
main.just: The justification of the main title for the plot, default is centered
main.x: The x location of the main title, deault is 0.5
main.y: The y location of the main title, default is 0.5
main.cex: Size of text for main plot title, defaults to 3
add.stripplot: logical whether to plot all points, defaults to FALSE
jitter.factor: Numeric value to apply to jitter, default is 1
jitter.amount: Numeric; amount of noise to add, default is NULL
points.pch: pch value to use for stripplot
points.col: colour(s) to use for stripplot (either a single colour or a vector)
points.cex: cex value to use for stripplot
points.alpha: alpha value to use for stripplot
abline.h: Specify the horizontal superimpose line
abline.v: Specify the vertical superimpose line
abline.lty: Specify the superimpose line type
abline.lwd: Specify the superimpose line width
abline.col: Specify the superimpose line colour (defaults to black)
add.rectangle: Allow a rectangle to be drawn, default is FALSE
xleft.rectangle: Specifies the left x ooordinate of the rectangle to be drawn
ybottom.rectangle: Specifies the bottom y coordinate of the rectangle to be drawn
xright.rectangle: Specifies the right x coordinate of the rectangle to be drawn
ytop.rectangle: Specifies the top y coordinate of the rectangle to be drawn
col.rectangle: Specifies the colour of the rectangle to be drawn
alpha.rectangle: Specifies the colour bias of the rectangle to be drawn
box.ratio: ability to change the box width, defaults to 1
col: The colour to fill the interior of the boxplot, defaults to white
alpha: The alpha of the interior boxplot colour specified in `col`. Defaults to 1 (opaque)
border.col: Colour of the boxplot, defaults to black
symbol.cex: Size of the boxplot outlier-symbol
lwd: Line width, defaults to 1
outliers: logical whether to plot outliers, defaults to TRUE
sample.order: String specifying how samples should be ordered. Either none, increasing, or decreasing.
order.by: A string specifying what the sample order should be ordered by, either max, min, median or mean
xlab.label: The label for the x-axis
ylab.label: The label for the y-axis
xlab.cex: Size of x-axis label, defaults to 3
ylab.cex: Size of y-axis label, defaults to 3
xlab.col: Colour of the x-axis label, defaults to “black”
ylab.col: Colour of the y-axis label, defaults to “black”
xlab.top.label: The label for the top x-axis
xlab.top.cex: Size of top x-axis label
xlab.top.col: Colour of the top x-axis label
xlab.top.just: Justification of the top x-axis label, defaults to centered
xlab.top.x: The x location of the top x-axis label
xlab.top.y: The y location of the top y-axis label
xlimits: Two-element vector giving the x-axis limits
ylimits: Two-element vector giving the y-axis limits
xat: Accepts a vector listing where x-axis ticks should be drawn or if automatic scaling is desired, one of three strings: “auto”, “auto.linear” or “auto.log”. Automatic scaling fixes x-axis tick locations, labels, and data values dependent on given data. “auto” will determine whether linear or logarithmic scaling fits the given data best, “auto.linear” or “auto.log” will force data to be scaled linearly or logarithmically respectively. Defaults to lattice automatic (TRUE). For more details see `auto.axis()`.
yat: Accepts a vector listing where y-axis ticks should be drawn or if automatic scaling is desired, one of three strings: “auto”, “auto.linear” or “auto.log”. Automatic scaling fixes y-axis tick locations, labels, and data values dependent on given data. “auto” will determine whether linear or logarithmic scaling fits the given data best, “auto.linear” or “auto.log” will force data to be scaled linearly or logarithmically respectively. Defaults to lattice automatic (TRUE). For more details see `auto.axis()`.
xaxis.lab: Vector listing x-axis tick labels, defaults to automatic (TRUE). Using automatic scaling with xat will overwrite user input. Set to NULL to remove x-axis labels.
yaxis.lab: Vector listing y-axis tick labels, defaults to automatic (TRUE). Using automatic scaling with yat will overwrite user input. Set to NULL to remove y-axis labels.
xaxis.cex: Size of x-axis tick labels, defaults to 2
yaxis.cex: Size of y-axis tick labels, defaults to 2
xaxis.col: Colour of the x-axis tick labels, defaults to “black”
yaxis.col: Colour of the y-axis tick labels, defaults to “black”
xaxis.fontface: Fontface for the x-axis scales
yaxis.fontface: Fontface for the y-axis scales
xaxis.rot: Rotation of x-axis tick labels; defaults to 0
yaxis.rot: Rotation of y-axis tick labels; defaults to 0
xaxis.tck: Specifies the length of the tick marks for x-axis, defaults to 1 (bottom) and 0 (top)
yaxis.tck: Specifies the length of the tick marks for y-axis, defaults to 1
layout: A vector specifying the number of columns, rows (e.g., c(2,1). Default is NULL; see lattice::xyplot for more details

as.table: Specifies panel drawing order, default is FALSE which draws panels from bottom left corner, moving right then up. Set to TRUE to draw from top left corner, moving right then down
x.spacing: A number specifying the distance between panels along the x-axis, defaults to 0
y.spacing: A number specifying the distance between panels along the y-axis, defaults to 0
x.relation: Allows x-axis scales to vary if set to “free”, defaults to “same”
y.relation: Allows y-axis scales to vary if set to “free”, defaults to “same”
top.padding: A number specifying the distance to the top margin, defaults to 0.5
bottom.padding: A number specifying the distance to the bottom margin, defaults to 2
right.padding: A number specifying the distance to the right margin, defaults to 1
left.padding: A number specifying the distance to the left margin, defaults to 2
ylab.axis.padding: A number specifying the distance of y-axis label to the y-axis, defaults to 0

add.text: Allow additional text to be drawn, default is FALSE
text.labels: Labels for additional text. If the formula contains group, the length of this argument should match with the number of groups.
text.x: The x co-ordinates where additional text should be placed
text.y: The y co-ordinates where additional text should be placed
text.anchor: Part of text that should be anchored to x/y coordinates. Defaults to 'centre'. Use 'left' or 'right' to left or right-align text.
text.col: The colour of additional text
text.cex: The size of additional text
text.fontface: The fontface for additional text
key: Add a key to the plot. See xyplot.
legend: Add a legend to the plot. Helpful for adding multiple keys and adding keys to the margins of the plot. See xyplot.
strip.col: Strip background colour, defaults to “white”
strip.cex: Strip title character expansion
strip.fontface: Strip title fontface, defaults to bold
line.func: Function for the line that should be drawn on top of plot
line.from: The starting point of the line on the plot
line.to: The ending point of the line on the plot
line.col: Colour of the line on the plot
line.infront: Should the line appear in front of the plot or not
height: Figure height, defaults to 6 inches
width: Figure width, defaults to 6 inches
size.units: Figure units, defaults to inches
resolution: Figure resolution in dpi, defaults to 1600
enable.warnings: Print warnings if set to TRUE, defaults to FALSE
description: Description of image/plot; default NULL.
style: defaults to “BoutrosLab”, also accepts “Nature”, which changes parameters according to Nature formatting requirements
preload.default: ability to set multiple sets of diffrent defaults depending on publication needs
use.legacy.settings: boolean to set wheter or not to use legacy mode settings (font)
disable.factor.sorting: Disable barplot auto sorting factors alphabetically/numerically

Value

If filename is NULL then returns the trellis object, otherwise creates a plot and returns a 0/1 success code.

Author

Maud H.W. Starmans

Warning

If this function is called without capturing the return value, or specifying a filename, it may crash while trying to draw the histogram. In particular, if a script that uses such a call of create histogram is called by reading the script in from the command line, it will fail badly, with an error message about unavailable fonts:


    Error in grid.Call.graphics("L_text", as.graphicsAnnot(x$label), x$x,  )
        Invalid font type
    Calls: print ... drawDetails.text -> grid.Call.graphics -> .Call.graphics

Examples


set.seed(12345);

simple.data <- data.frame(
    x = rnorm(1000),
    y = rep('A',1000)
    );

create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Simple', fileext = '.tiff'),
    formula = y ~ x,
    data = simple.data,
    main = 'Simple',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 50
    );


# add stripplot behind boxplot
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_with_Stripplot', fileext = '.tiff'),
    formula = y ~ x,
    data = simple.data,
    main = 'With Stripplot',
    add.stripplot = TRUE,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 50
    );


# Multi-coloured stripplot
strip.data <- data.frame(
    score = c(rnorm(30, 15, 3), rnorm(50, 20, 4)),
    sex = sample(c('male', 'female'), 80, replace = TRUE),
    gene = sample(c('a', 'b'), 80, replace = TRUE)
    );

create.boxplot(
    filename = NULL,
    formula = score ~ sex | gene,
    data = strip.data,
    main = 'Multi-Coloured Stripplot',
    add.stripplot = TRUE,
    points.col = c('pink', 'dodgerblue'),
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );



# format data
reformatted.data <- data.frame(
    x = as.vector(t(microarray[1:10,1:58])),
    y = as.factor(rep(rownames(microarray[1:10,1:58]),each = 58)),
    z = sample(1:10, 580, replace = TRUE)
    );

# Minimal Input
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Minimal_Input', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Minimal input',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 50
    );


# Minimal Input
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Disable_Factor_Sorting_Input', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'No Factor Sorting',
    disable.factor.sorting = TRUE,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 50
    );


# Axes and labels
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Axes_Labels', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Axes & labels',
    # Adjusting axes size
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    # Adding y-axis label
    ylab.label = 'Gene',
    # setting axes limits
    xlimits = c(0,13),
    xat = seq(0,12,2),
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );


# \donttest{
# Sorting
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Sorted', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Sorting',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # Reordered by median
    sample.order = 'increasing',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );


# Colour change
sex.colour <- as.character(patient$sex);
sex.colour[sex.colour == 'male'] <- 'dodgerblue';
sex.colour[sex.colour == 'female'] <- 'pink';

create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Colour_Change', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Colour change',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # Colour change
    col = sex.colour,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );

    
# Remove y-axis labels
create.boxplot(
    formula = y ~ x,
    data = reformatted.data,
    main = 'Remove y-axis labels',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    yaxis.lab = NULL, # Remove labels with NULL
    # Colour change
    col = sex.colour,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );


# Log Scaled Axis
log.data <- data.frame(
    x = 10 ** rnorm(1000, 5, 2),
    y = rep('A',1000)
    );

create.boxplot(
    formula = x ~ y,
    data = log.data,
    # Log base 10 scale y axis
    yat = 'auto.log',
    main = 'Log Scale',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );


# Legend
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Legend', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Legend',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    col = sex.colour,
    # legend
    legend = list(
        inside = list(
            fun = draw.key,
            args = list(
                key = list(
                    points = list(
                        col = 'black',
                        pch = 22,
                        cex = 1.5,
                        fill = c('dodgerblue','pink')
                        ),
                    text = list(
                        lab = c('male','female')
                        ),
                    cex = 1
                    )
                ),
            x = 0.03,
            y = 0.97,
            corner = c(0,1),
            draw = FALSE
            )
        ),
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 100
    );


# Orientation
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Orientation', fileext = '.tiff'),
    # switch the order
    formula = x ~ y,
    data = reformatted.data,
    main = 'Orientation',
    xaxis.cex = 1,
    yaxis.cex = 1,
    # adjust the axes
    ylimits = c(0,13),
    yat = seq(0,12,2),
    # rotate the labels
    xaxis.rot = 90,
    xlab.label = 'Gene',
    xlab.cex = 1.5,
    col = sex.colour,
    # legend
    legend = list(
        inside = list(
            fun = draw.key,
            args = list(
                key = list(
                    points = list(
                        col = 'black',
                        pch = 22,
                        cex = 1.5,
                        fill = c('dodgerblue','pink')
                        ),
                    text = list(
                        lab = c('male','female')
                        ),
                    cex = 1
                    )
                ),
            x = 0.23,
            y = 0.97,
            corner = c(0,1),
            draw = FALSE
            )
        ),
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );


# Background rectangle
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_BG_Rect', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Bg rectangle',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # draw rectangle
    add.rectangle = TRUE,
    xleft.rectangle = 0,
    xright.rectangle = 13,
    ybottom.rectangle = seq(0.5, 8.5, 2),
    ytop.rectangle = seq(1.5, 9.5, 2),
    col.rectangle = 'grey',
    alpha.rectangle = 0.5,
    col = sex.colour,
    # legend
    legend = list(
        inside = list(
            fun = draw.key,
            args = list(
                key = list(
                    points = list(
                        col = 'black',
                        pch = 22,
                        cex = 1.5,
                        fill = c('dodgerblue','pink')
                        ),
                    text = list(
                        lab = c('male','female')
                        ),
                    cex = 1
                    )
                ),
            x = 0.03,
            y = 0.97,
            corner = c(0,1),
            draw = FALSE
            )
        ),
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );


# Line
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Line', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Line',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # draw line
    line.func = function(x){c(0.5, 10.5)},
    line.from = 11,
    line.to = 11,
    line.col = 'grey',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );


# Panel Organization
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Panels_numeric_conditioning', fileext = '.tiff'),
    formula = ~ x | z,
    data = reformatted.data,
    main = 'Panels',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # Setting up the layout
    layout = c(2,5),
    x.relation = 'free',
    x.spacing = 1,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );
#> Warning: Numeric values detected for conditional variable. If text labels are desired, please convert conditional variable to character.


create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Panels_factor_conditioning', fileext = '.tiff'),
    formula = ~ x | y,
    data = reformatted.data,
    main = 'Panels',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # Setting up the layout
    layout = c(2,5),
    x.relation = 'free',
    x.spacing = 1,
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );


# Nature format
create.boxplot(
    # filename = tempfile(pattern = 'Boxplot_Nature_style', fileext = '.tiff'),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Nature style',
    xaxis.cex = 1,
    yaxis.cex = 1,

    # set style to Nature
    style = 'Nature',

    # demonstrating how to italicize character variables
    ylab.lab = expression(paste('italicized ', italic('a'))),

    # demonstrating how to create en-dashes
    xlab.lab = expression(paste('en dashs: 1','\u2013', '10'^'\u2013', ''^3)),
    resolution = 1200
    );
#> Warning: Nature also requires italicized single-letter variables and en-dashes
#> 			for ranges and negatives. See example in documentation for how to do this.
#> Warning: Avoid red-green colour schemes, create TIFF files, do not outline the figure or legend.


# Sorting by mean and multiple filenames
create.boxplot(
    filename = c(
        tempfile(pattern = 'Boxplot_Sorted1', fileext = '.tiff'),
        tempfile(pattern = 'Boxplot_Sorted2', fileext = '.tiff')
        ),
    formula = y ~ x,
    data = reformatted.data,
    main = 'Sorting',
    xaxis.cex = 1,
    yaxis.cex = 1,
    xlab.cex = 1.5,
    ylab.cex = 1.5,
    ylab.label = 'Gene',
    xlimits = c(0,13),
    xat = seq(0,12,2),
    # Reordered by median
    sample.order = 'increasing',
    order.by = 'mean',
    description = 'Boxplot created by BoutrosLab.plotting.general',
    resolution = 200
    );


# Adding text to plot
# Generate normally distributed variables with two different means
set.seed(779);
groupA <- rnorm(n = 100, mean = 10, sd = 2);
groupB <- rnorm(n = 134, mean = 10.5, sd = 2);

# Create data frame for plotting
to.plot <- data.frame(
  y = rep(
    c('1', '2'),
    times = c(100, 134)
    ),
  x = c(groupA, groupB)
  );

# Get difference between means
diff.mean <- round(mean(groupB) - mean(groupA), 2);

# Plot and display difference
create.boxplot(
  formula = x ~ y,
  # filename = tempfile(pattern = 'boxplot_with_text', fileext = '.tiff'),
  data = to.plot,
  add.stripplot = TRUE,
  add.text = TRUE,
  text.labels = bquote(mu[B] - mu[A] == .(diff.mean)),
  text.x = 2.1,
  text.y = 15.3,
  text.col = 'black',
  text.cex = 1.5,
  text.fontface = 'bold',
  ylimits = c(
    min(to.plot$x) - abs(min(to.plot$x) * 0.1),
    max(to.plot$x) + abs(max(to.plot$x) * 0.1)
    ),
  resolution = 200
    );
#> Warning: NAs introduced by coercion

# }