How to assign heatmap color range

This issue is similar to this one I answered. It's about heatmaps in highcharter package. The difference is that highcharter is more convenient to assign colors. I use the same concept to solve this heatmap.2 case.

Sample Data

x <- rnorm(1000, 2, 1.5)
x[x < -2] <- -2
x[x > 6] <- 6
mat <- matrix(x, 20, 50)

Description

First, use symkey = F, symbreaks = F in heatmap.2 can make the color-key asymmetric. And then, you need to find where the specified value (i.e. 0) is located in your data. By the following formula, you can get a value within 0 ~ 1 to represent the position of a certain value.

In this case, (0-(-2))/(6-(-2)) is 0.25. If there are 100 breakpoints, then you need to concatenate 100 * 0.25 = 25 red colors and 100 - 25 = 75 blue colors. You can find that the white color will be consistent with zero.

library(gplots)
library(RColorBrewer)

hmcol <- function(n){
  colfun1 <- colorRampPalette(rev(brewer.pal(9, "Reds")))
  colfun2 <- colorRampPalette(brewer.pal(9, "Blues"))
  p <- (0 - min(mat)) / (max(mat) - min(mat))
  gap <- n * p
  c(colfun1(floor(gap)), colfun2(n - gap))
}

heatmap.2(mat, trace = "none", symkey = F, symbreaks = F, col = hmcol(50))

You can set any number in hmcol() to indicate the number of breakpoints.(You can try col = hmcol(10) and col = hmcol(100) to see the difference.)