Default aes // x and y labels

[sinarueeger]

If observed=P, then xlab and ylab should be:

xlab(expression(Expected ~ ~-log[10](italic(P)))) +
ylab(expression(Observed ~ ~-log[10](italic(P))))
  

• https://community.rstudio.com/t/can-i-change-labels-with-custom-geom/955/5

[sinarueeger]

The default aes are for the geom_gwas_qq_{hex|}are y.

Problem
After plotting, the x-label is expected, the y-label whatever was passed onto y.
But we want -log10(expected) and -log10(P).

See also
See #47 (comment)

See also #2: the default aes to pass onto geom_qq is sample. And once plotted - the plot gets an x-axis theoretical (see below).

library(ggGWAS)
library(ggplot2)
#> Warning: package 'ggplot2' was built under R version 3.5.2
set.seed(3)
df <- data.frame(P = runif(1000))

ggplot(df, aes(sample = P)) +
  stat_qq(distribution = stats::qunif) +
  scale_x_continuous(trans = "mlog10") +
  scale_y_continuous(trans = "mlog10")

^{Created on 2019-05-14 by the reprex package (v0.2.1)}

This is part of the behaviour I would like to have for both geom_gwas_qq and geom_gwas_qq_hex.

[sinarueeger]

So stat_qq has a desirable behaviour. This is what has been implemented before in geom_gwas_qq. However - this won't work on the hex version of qq.

StatQq <- ggproto("StatQq", Stat,
  default_aes = aes(y = stat(sample), x = stat(theoretical)), #########<<<< ADD THIS

  required_aes = c("sample"), #########<<<< ADD THIS

  compute_group = function(data, scales, quantiles = NULL,
                           distribution = stats::qnorm, dparams = list(),
                           na.rm = FALSE) {

    sample <- sort(data$sample)
    n <- length(sample)

    # Compute theoretical quantiles
    if (is.null(quantiles)) {
      quantiles <- stats::ppoints(n)
    } else {
      stopifnot(length(quantiles) == n)
    }

    theoretical <- do.call(distribution, c(list(p = quote(quantiles)), dparams))

    new_data_frame(list(sample = sample, theoretical = theoretical)) #########<<<< ADD THIS
  }
)

To get around the hex geom, I will try to use hexbin with a point geom.

[sinarueeger]

This could help: draw_panels and draw_labels could work

  # ,
  # draw_panels = function(data, panel_scales, coord) {
  #   ## Transform the data first
  #   coords <- coord$transform(data, panel_scales)
  #
  #   ## Let's print out the structure of the 'coords' object
  #   str(coords)
  #
  #   ## Construct a grid grob
  #   pointsGrob(
  #     x = coords$x,
  #     y = coords$y,
  #     pch = coords$shape
  #   )
  # },
  # draw_labels <- function(data, panel_scales, coord) {
  #  has something to do with gtable: https://ggplot2.tidyverse.org/reference/ggplot2-ggproto.html
  ## labels from qqman::qq()
  #  xlab(expression(Expected ~ ~-log[10](italic(p)))) +
  #  ylab(expression(Observed ~ ~-log[10](italic(p))))
  #     }