In caleblareau/BuenColors: Plotting utils and colors for Buenrostrolab

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "figure/",
  fig.height = 1
)

BuenColors

A package that facilitates nice colors created for the Buenrostro Lab. The base of the package including some divergent color schemes were taken from the wesanderson package. This README provides a quick overview of how to use the color schemes with ggplot2 and which palettes are available.

Installation via GitHub

devtools::install_github("caleblareau/BuenColors")

Color Palettes

library(BuenColors)

With the library loaded, just type this to get either the continuous or discrete (by default) palette printed in your plot console.

jdb_palette("FantasticFox")
jdb_palette("aqua_brick", type = "continuous")

Here are all the names that are available...

sort(names(jdb_palettes))

Color Maps

To keep consistent color designations, one can use the color_map function to link features to their specific hex color annoations. For example,

jdb_color_map(c("HSC"))

returns the hex code associated with HSC in the Buenrostro Lab paradigm. This function may be applied over multiple features--

jdb_color_map(c("HSC", "CMP", "HSC"))

and will error out when a feature is not recognized--

jdb_color_map(c("WHAT"))
#>  Error: all(name %in% names(jdb_color_maps)) is not TRUE 

Here are all the names that are available...

sort(names(jdb_color_maps))

Here are what the mappings look like...

df <- data.frame(names = names(jdb_color_maps), 
                 jdb_color_maps = unname(jdb_color_maps), stringsAsFactors = FALSE)

ggplot(df, aes(x = names, y = -1, fill = names)) +
  geom_bar(stat = "identity") + pretty_plot(fontsize = 20) + 
  scale_fill_manual(values = jdb_color_maps) + coord_flip() +
  theme(legend.position = "none")+
  scale_y_continuous(expand = c(0, 0)) +
  labs(x = NULL, y = NULL, title = "Buenrostro Lab Cell Schemes") +
  theme(axis.title.x=element_blank(), axis.text.x=element_blank(),
        axis.ticks.x=element_blank(), panel.border = element_blank())

ggplot example

To coordinate a ggplot feature (e.g. data point in a scatter plot) with a particular color, this post was a life-saver. Specifically, we'll use a named vector to coordinate the discrete values. Here's an example--

xy <- 1:7
cell <- c("GMP-A", "Ery", "CD4", "Ery", "LMPP", "ERY", "MEP")

df <- data.frame(
  xy = xy,
  cell = cell, stringsAsFactors = FALSE
)

ggplot(df, aes(x = xy, y = xy, color = cell)) +
  geom_point(size = 10) + pretty_plot() +
  scale_color_manual(values = jdb_color_maps)

P.S.-- this call will return a blank color for features that are not found in the color map ("ERY" in this example).

Important Note on color mappings...

The above ggplot command works because jdb_color_maps (with an 's') exists as a named vector in the BuenColors NAMESPACE. The provided function (jdb_color_map) does not have an 's' by the way. This same syntax of supplying a named vector should work for all discrete color scale functionalities in ggplot.

Discrete colors

The trick here is to use scale_color_manual like it is shown here--

library(ggplot2)
ggplot(iris, aes(Sepal.Length, Sepal.Width, color = Species)) + 
  geom_point() + pretty_plot() + 
  scale_color_manual(values = jdb_palette("brewer_spectra"))

Continuous fill

The trick here is to use scale_color_gradientn like it is shown here--

df <- data.frame(x = rnorm(1000), y = 0)
ggplot(df, aes(x=x, y=y, colour=x)) + geom_point() + 
  scale_color_gradientn(colors = jdb_palette("flame_light")) +
  pretty_plot()

Density Plot

Best way that I've found to make the density color function represented in the points. Thanks to Kamil Slowikowski for figuring this out.

dat <- data.frame(
  x = c(
    rnorm(1e4, mean = 0, sd = 0.1),
    rnorm(1e3, mean = 0, sd = 0.1)
  ),
  y = c(
    rnorm(1e4, mean = 0, sd = 0.1),
    rnorm(1e3, mean = 0.1, sd = 0.2)
  )
)
dat$density <- get_density(dat$x, dat$y)
ggplot2::ggplot(dat) + geom_point(aes(x, y, color = density)) + 
  scale_color_gradientn(colors = jdb_palette("solar_extra")) +
  pretty_plot()

Shuffle Plot Order

Quick wrapper using shuf to change the order of plotting points (to a random presentation) to avoid hiding effects.

tdf<-paste(system.file('rds',package='BuenColors'),'basicTSNE.rds',sep='/')
df <- readRDS(tdf)
ggplot(shuf(df)) + geom_point(aes(X1, X2, color = counts)) + 
  scale_color_gradientn(colors = jdb_palette("solar_extra")) +
  pretty_plot()

Continuous Colors

Here's what each palette looks like on a continuous scale.

all <- sort(names(jdb_palettes))
names(all) <- sort(names(jdb_palettes))
lapply(all, function(p) jdb_palette(p, type = "continuous"))

Discrete colors

Here are the discrete color units that go into each scale. If n is small and discrete, note that the colors that are selected are read from left to right.

lapply(all, function(p) jdb_palette(p, type = "continuous"))

caleblareau/BuenColors documentation built on March 7, 2020, 3:34 p.m.