How do I define color groups based on numerical threshold values for ggplot2 scatterplot

Question

I have a data set that contains 2 variables x = event number & y = assay amplitude. I am trying to create a scatterplot in ggplot2 where all of the points that are > 3000 are colored in one color and all of the points < 3000 are in a different color.

I can get the plot and change the color for all data points, but can't figure out how to define a color scheme based on the value threshold.

Here is a sample of the data I'm using:

dat <- data.frame(x=c(399, 16022, 14756, 2609, 1131, 12135, 
                                 7097, 12438, 12604, 14912, 11042, 
                                 14024, 7033, 4971, 15533, 4507, 4627, 
                                 12600, 7458, 14557, 3999, 3154, 6073),
                  y=c(3063.40137, 3687.42041, 3911.856, 
                                    4070.91748, 4089.99561, 4095.50317,
                                    4159.899, 4173.117, 4177.78955, 
                                    4186.46875, 4201.874, 4272.022, 
                                    638.615, 649.8995, 668.8346,
                                    688.754639, 711.92, 712.689636, 
                                    721.1352, 737.841, 741.0727, 
                                    755.2549, 756.730652))

Answer 1

You really just need to do a new indicator variable for this. As @hrbrmstr says, cut is a good general way to do this (works for as many cutpoints as you want).

dat$col <- cut(dat$y,
               breaks = c(-Inf, 3000, Inf),
               labels = c("<=3000", ">3000"))

ggplot(dat, aes(x = x, y = y, color = col)) +
  geom_point()

Answer 2

This can be done on the fly with an ifelse statement and without having to create an extra column in the dataset:

ggplot(dat, aes(x = x, y = y)) +
geom_point(aes(color = ifelse(y>3000, 'red', 'blue'))) +
scale_colour_manual(labels = c("<3000", ">3000"), values=c('blue', 'red')) + 
labs(color = "Values")