Intro to ggplot2

Ted Laderas

March 11, 2015

Intro to Hadleyland

Hadley Wickham started building solutions to specific R problems as part of his dissertation work, expanding to form an empire within R.

plyr/dplyr (How do I apply functions to different data structures?)
reshape2 (How can I change/summarize data from one form to another)
ggplot2 (How can I make pretty graphics?)
testthat (Unit testing framework for R)

The Grammar of Graphics

ggplot2 is a graphical plotting package that utilizes a “grammar of graphics”:

ggplot(iris) + geom_point(aes(x = Sepal.Length, y = Sepal.Width))

# + theme_minimal()

Specify data with ggplot()
Specify data representation with geoms (layer)
- (bar, line, scatter, heatmap, etc)
Specify data mapping with aes()
- (x, y, Color, Size, fill)
Subset data with facets
Change the look around with themes
Add other properties (titles, etc)

Simple Example

library(ggplot2)
data(iris)
ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length)) + theme_minimal()

aes() - the Aesthetic Function

aes() lets you map a variable to a visual property of the geom. Aesthetics can include:

x or y (almost any geom)
size
color (point or line)
shape (point)
linetype (line)
fill (rect, bar, histogram)

Documentation for aes() is remarkably terrible, as it’s different for each geom. Look at the documentation for each geom for a list of mappable aesthetics.

Let’s get fancier and add coloring

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length, color=Species, size=Petal.Length))

Using a continuous variable

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length, color=Petal.Length)) +
  scale_color_continuous(low = "white", high = "green")

Other aesthetic attributes: Size

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length, size=Petal.Length, color=Species))

Geoms - lots of them

Common geoms to use are below. Each of these requires a different aesthetic mapping.

geom_point() - scatterplot
geom_line() - line
geom_bar() - barplot - requires
geom_boxplot() - boxplot
geom_tile() - “heatmap”
geom_histogram() - histogram

I will try to talk about two or three of them.

Histograms

ggplot(iris) + geom_histogram(aes(x= Sepal.Width, fill=Species, group=Species))

## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.

Boxplots

ggplot(iris) + geom_violin(aes(x=Species, y=Sepal.Width))

Layering geoms

Where ggplot2 becomes very powerful is that it allows you to layer geoms. Here we’re looking at a smoothed version of the data over all coordinates using loess smoothing.

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length)) + 
  geom_smooth(aes(x= Sepal.Width, y=Sepal.Length))

## geom_smooth: method="auto" and size of largest group is <1000, so using loess. Use 'method = x' to change the smoothing method.

Facets - Subsetting the data

Perhaps we want to visualize the data onto three separate plots (one for each species). We can do this with facets:

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length)) + 
  facet_grid(facets=Species ~ .)

For rows, we can change the facet_grid() order:

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length)) + 
  facet_grid(facets=. ~ Species)

Theme properties and scaling

The look of everything on the graph can be customized, including

color scale (continuous or discrete colors)
order of elements
fonts of everything
labels and legends

See ?theme for more information. Also: http://sharpstatistics.co.uk/r/ggplot2-guide/

Theme properties

The default theme is called theme_grey(). You can see all of the properties of a theme by looking at it.

theme_grey()

## List of 40
##  $ line                :List of 4
##   ..$ colour  : chr "black"
##   ..$ size    : num 0.5
##   ..$ linetype: num 1
##   ..$ lineend : chr "butt"
##   ..- attr(*, "class")= chr [1:2] "element_line" "element"
##  $ rect                :List of 4
##   ..$ fill    : chr "white"
##   ..$ colour  : chr "black"
##   ..$ size    : num 0.5
##   ..$ linetype: num 1
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ text                :List of 8
##   ..$ family    : chr ""
##   ..$ face      : chr "plain"
##   ..$ colour    : chr "black"
##   ..$ size      : num 12
##   ..$ hjust     : num 0.5
##   ..$ vjust     : num 0.5
##   ..$ angle     : num 0
##   ..$ lineheight: num 0.9
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.text           :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : chr "grey50"
##   ..$ size      :Class 'rel'  num 0.8
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ strip.text          :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      :Class 'rel'  num 0.8
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.line           : list()
##   ..- attr(*, "class")= chr [1:2] "element_blank" "element"
##  $ axis.text.x         :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : NULL
##   ..$ vjust     : num 1
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.text.y         :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : num 1
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.ticks          :List of 4
##   ..$ colour  : chr "grey50"
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..$ lineend : NULL
##   ..- attr(*, "class")= chr [1:2] "element_line" "element"
##  $ axis.title.x        :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.title.y        :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : num 90
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ axis.ticks.length   :Class 'unit'  atomic [1:1] 0.15
##   .. ..- attr(*, "unit")= chr "cm"
##   .. ..- attr(*, "valid.unit")= int 1
##  $ axis.ticks.margin   :Class 'unit'  atomic [1:1] 0.1
##   .. ..- attr(*, "unit")= chr "cm"
##   .. ..- attr(*, "valid.unit")= int 1
##  $ legend.background   :List of 4
##   ..$ fill    : NULL
##   ..$ colour  : logi NA
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ legend.margin       :Class 'unit'  atomic [1:1] 0.2
##   .. ..- attr(*, "unit")= chr "cm"
##   .. ..- attr(*, "valid.unit")= int 1
##  $ legend.key          :List of 4
##   ..$ fill    : chr "grey95"
##   ..$ colour  : chr "white"
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ legend.key.size     :Class 'unit'  atomic [1:1] 1.2
##   .. ..- attr(*, "unit")= chr "lines"
##   .. ..- attr(*, "valid.unit")= int 3
##  $ legend.key.height   : NULL
##  $ legend.key.width    : NULL
##  $ legend.text         :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      :Class 'rel'  num 0.8
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ legend.text.align   : NULL
##  $ legend.title        :List of 8
##   ..$ family    : NULL
##   ..$ face      : chr "bold"
##   ..$ colour    : NULL
##   ..$ size      :Class 'rel'  num 0.8
##   ..$ hjust     : num 0
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ legend.title.align  : NULL
##  $ legend.position     : chr "right"
##  $ legend.direction    : NULL
##  $ legend.justification: chr "center"
##  $ legend.box          : NULL
##  $ panel.background    :List of 4
##   ..$ fill    : chr "grey90"
##   ..$ colour  : logi NA
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ panel.border        : list()
##   ..- attr(*, "class")= chr [1:2] "element_blank" "element"
##  $ panel.grid.major    :List of 4
##   ..$ colour  : chr "white"
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..$ lineend : NULL
##   ..- attr(*, "class")= chr [1:2] "element_line" "element"
##  $ panel.grid.minor    :List of 4
##   ..$ colour  : chr "grey95"
##   ..$ size    : num 0.25
##   ..$ linetype: NULL
##   ..$ lineend : NULL
##   ..- attr(*, "class")= chr [1:2] "element_line" "element"
##  $ panel.margin        :Class 'unit'  atomic [1:1] 0.25
##   .. ..- attr(*, "unit")= chr "lines"
##   .. ..- attr(*, "valid.unit")= int 3
##  $ panel.margin.x      : NULL
##  $ panel.margin.y      : NULL
##  $ strip.background    :List of 4
##   ..$ fill    : chr "grey80"
##   ..$ colour  : logi NA
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ strip.text.x        :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ strip.text.y        :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      : NULL
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : num -90
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ plot.background     :List of 4
##   ..$ fill    : NULL
##   ..$ colour  : chr "white"
##   ..$ size    : NULL
##   ..$ linetype: NULL
##   ..- attr(*, "class")= chr [1:2] "element_rect" "element"
##  $ plot.title          :List of 8
##   ..$ family    : NULL
##   ..$ face      : NULL
##   ..$ colour    : NULL
##   ..$ size      :Class 'rel'  num 1.2
##   ..$ hjust     : NULL
##   ..$ vjust     : NULL
##   ..$ angle     : NULL
##   ..$ lineheight: NULL
##   ..- attr(*, "class")= chr [1:2] "element_text" "element"
##  $ plot.margin         :Class 'unit'  atomic [1:4] 1 1 0.5 0.5
##   .. ..- attr(*, "unit")= chr "lines"
##   .. ..- attr(*, "valid.unit")= int 3
##  - attr(*, "class")= chr [1:2] "theme" "gg"
##  - attr(*, "complete")= logi TRUE

Making a new theme

To customize the theme, you must use elements to fill out the various properties of the theme list. Each of these properties maps to a specific element, either element_text, element_rect, or element_line.

For example, we want a white background and different fonts for the axis titles.

theme_new <- theme(panel.background = element_rect(fill = "white"), 
                   axis.title.x = element_text(family = "Times New Roman"),
                   axis.title.y = element_text(family = "Arial"))

Trying our new theme out

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length)) + theme_new

Changing a discrete color scale

ggplot(iris) + geom_point(aes(x= Sepal.Width, y=Sepal.Length, color=Species)) +
  scale_color_manual(values= c("blue", "purple", "gold"))

If your data is originally in a matrix like format

For bioinformaticians, we often work with a matrix like format where genes are rows and samples are columns. This “wide” data format is not really usable by ggplot.

We need to “melt” the data frame to make it easier for ggplot to work with. Our melted data frame will have one data point for row, and additional Gene and Sample Columns. First we load our data in.

library(reshape2)

## Warning: package 'reshape2' was built under R version 3.1.2

expData <- data.frame( 
  Gene = c("G1","G2","G3","G4", "G5", "G6", "G7", "G8"),
  Tumor1 = c( 5.6, 6.2, 5.6, 6.6, 1.3, 1.4, 2.1, 2.3),
  Tumor2 = c(5.4, 6.1, 5.8, 6.2, 1.3, 1.8, 2.2, 2.1),
  Normal1 = c(2.0, 2.5, 1.1, 2.8, 5.5, 5.8, 6.1, 5.7),
  Normal2 = c(2.1, 2.6, 1.3, 2.4, 5.7, 5.4, 5.8, 5.9)
)
expData

##   Gene Tumor1 Tumor2 Normal1 Normal2
## 1   G1    5.6    5.4     2.0     2.1
## 2   G2    6.2    6.1     2.5     2.6
## 3   G3    5.6    5.8     1.1     1.3
## 4   G4    6.6    6.2     2.8     2.4
## 5   G5    1.3    1.3     5.5     5.7
## 6   G6    1.4    1.8     5.8     5.4
## 7   G7    2.1    2.2     6.1     5.8
## 8   G8    2.3    2.1     5.7     5.9

Melting the data into longform data

Then we melt the data using the melt() function:

expDataMelt <- melt(expData, id.vars="Gene", variable.name="Sample",
                    value.name = "value")
expDataMelt

##    Gene  Sample value
## 1    G1  Tumor1   5.6
## 2    G2  Tumor1   6.2
## 3    G3  Tumor1   5.6
## 4    G4  Tumor1   6.6
## 5    G5  Tumor1   1.3
## 6    G6  Tumor1   1.4
## 7    G7  Tumor1   2.1
## 8    G8  Tumor1   2.3
## 9    G1  Tumor2   5.4
## 10   G2  Tumor2   6.1
## 11   G3  Tumor2   5.8
## 12   G4  Tumor2   6.2
## 13   G5  Tumor2   1.3
## 14   G6  Tumor2   1.8
## 15   G7  Tumor2   2.2
## 16   G8  Tumor2   2.1
## 17   G1 Normal1   2.0
## 18   G2 Normal1   2.5
## 19   G3 Normal1   1.1
## 20   G4 Normal1   2.8
## 21   G5 Normal1   5.5
## 22   G6 Normal1   5.8
## 23   G7 Normal1   6.1
## 24   G8 Normal1   5.7
## 25   G1 Normal2   2.1
## 26   G2 Normal2   2.6
## 27   G3 Normal2   1.3
## 28   G4 Normal2   2.4
## 29   G5 Normal2   5.7
## 30   G6 Normal2   5.4
## 31   G7 Normal2   5.8
## 32   G8 Normal2   5.9

Making a heatmap with our melted data

ggplot(expDataMelt) + geom_tile(aes(x=Sample, y=Gene, fill=value)) + 
scale_fill_gradient(low="white", high = "darkgreen")

Using the built in Interface to ColorBrewer

Try different palette numbers here!

ggplot(expDataMelt) + geom_tile(aes(x=Sample, y=Gene, fill=value)) + 
 scale_fill_distiller(type="div", palette = 2)

Rotating labels and prettifying

You can change labels using the labs() function and you can also use one theme as a starting point and modify it as you wish.

ggplot(expDataMelt) + geom_tile(aes(x=Sample, y=Gene, fill=value)) + 
  scale_fill_gradient(low="white", high = "darkgreen") + 
  labs(x= "Sample Label", fill="Expression") +
  theme_minimal() + 
  theme(axis.text.x = element_text(angle=90))

Making a line plot with our melted data

Let’s plot the expression values as a line plot. To do this we need to tell ggplot2 that we want to plot each line by gene group. Additionaly we want a facet for each gene.

ggplot(expDataMelt) + geom_line(aes(x=Sample, y=value, color=Gene, group=Gene)) + 
  facet_grid(facets= Gene ~.)

A More Complex Example

ggplot(drugmelt[drugmelt$variable == "S6_pS235_S236" & !is.na(drugmelt$Stim) &
                  drugmelt$Cell == "BT20" & !drugmelt$Drug %in% c("DAS", "MEKi") & 
                  drugmelt$Time < 720,], aes(Time, value, group=DS, colour=DS)) + 
    geom_point() + geom_line() + facet_grid(Stim ~ Drug, margin=T) + 
  scale_x_continuous(breaks = c(30, 120, 480, 720)) + 
  theme(legend.position = "none", axis.text.x = element_text(angle=90))

S6 example

Interactive plots with ggvis

ggplot2 is great for static graphics. But for interactive graphics that allow you to change parameters, you will need to learn ggvis.

The good news: ggvis uses a similar grammar. The bad news: everything has a different name. Learning how to parametrize a plot can be tricky.

Thanks!

Hadley Wickham for ggplot2 and ggvis
Eric Leung for showing us knitR/Sweave

This presentation is available here: http://bit.ly/1GJu8GV

Future Topics?

Faculty Research Map
OHSU Graduates in the real world