The p value position is incorrectly defined as zero using rstatix::wilcox_test(), rstatix::add_xy_position(), and ggpubr::stat_pvalue_manual()

Question

After performing an rstatix::wilcox_test() on my data I automatically added the y.position value for plotting the p value using rstatix::add_xy_position(). I then plotted the data and added the p value using ggpubr::stat_pvalue_manual(). The problem is that the y.position value is defined incorrectly as zero, causing the plotted p value and bracket geom to appear at the bottom of the y-axis at zero when it should appear near the top of the y-axis. The documentation for rstatix indicates that the y.position value, as calculated by the max default setting, should be have been equal to max(data$Response) = 0.00008556344, not zero. Am I using these functions incorrectly? If not, how can I get the p value and bracket geom to appear at the y-axis data value corresponding to the maximum?

library(tidyverse)
library(rstatix)
library(ggpubr)

# minimal data example
data <- tibble::tribble(
  ~Group,       ~Response,
  "Treatment",  3.210486e-06,
  "Control",    4.006825e-06,
  "Treatment",  4.350836e-06,
  "Control",    4.216934e-06,
  "Treatment",  4.415194e-06,
  "Control",    1.0260606e-05,
  "Treatment",  1.111064e-06,
  "Control",    1.0779088e-05,
  "Treatment",  3.57185e-07,
  "Control",    1.139097e-06,
  "Treatment",  0,
  "Control",    2.31074e-07,
  "Treatment",  5.78956e-07,
  "Control",    4.371157e-06,
  "Treatment",  6.5825e-08,
  "Control",    9.587202e-06,
  "Treatment",  2.65383e-07,
  "Control",    3.57337e-06,
  "Treatment",  7.14146e-07,
  "Control",    3.868605e-06,
  "Treatment",  1.2213951e-05,
  "Control",    6.936899e-06,
  "Treatment",  4.71707e-07,
  "Control",    5.5957173e-05,
  "Treatment",  1.265942e-06,
  "Control",    8.5563441e-05,
  "Treatment",  0,
  "Control",    0,
  "Treatment",  0,
  "Control",    2.1306289e-05,
  "Treatment",  5.2055e-07,
  "Control",    1.8420094e-05
)

# performing Wilcoxon test and adding 'y.position' for the p value
wilcoxon_result <- data %>%
  wilcox_test(Response ~ Group) %>%
  add_xy_position("Group")

# the 'y.position' is incorrectly defined as zero.
wilcoxon_result
#> # A tibble: 1 × 11
#>   .y.      group1  group2     n1    n2 statistic       p y.position groups  xmin
#>   <chr>    <chr>   <chr>   <int> <int>     <dbl>   <dbl>      <dbl> <name> <dbl>
#> 1 Response Control Treatm…    16    16      206. 0.00326          0 <chr>      1
#> # … with 1 more variable: xmax <dbl>

# the p value and bracket geom are plotted at zero.
ggplot(data, aes(Group, Response)) +
  stat_summary(geom = "errorbar", fun.data = mean_cl_boot, width = 0.25) +
  stat_summary(geom = "crossbar", fun = mean, fun.min = mean, fun.max = mean, fatten = FALSE, width = 0.5) +
  stat_pvalue_manual(wilcoxon_result, "P = {p}", size = 18 / ggplot2::.pt, tip.length = 0, bracket.size = 2 / ggplot2::.stroke) +
  theme_pubr(base_size = 18, x.text.angle = 45)

Answer 1

It seems your y position is being rounded to zero because you are dealing with very small numbers. If you multiply the Response variable by 1 million you get the same p-value, and you can then divide the y position to get it to sit wherever you want :

# performing Wilcoxon test and adding 'y.position' for the p value
wilcoxon_result <- data %>%
  mutate(Response = Response * 1e6) %>%
  wilcox_test(Response ~ Group) %>%
  add_xy_position("Group") %>%
  mutate(y.position = y.position /3e6)

This gives:

wilcoxon_result
#> # A tibble: 1 x 11
#>  .y.      group1  group2       n1    n2 statistic       p y.position groups        xmin  xmax
#>  <chr>    <chr>   <chr>     <int> <int>     <dbl>   <dbl>      <dbl> <named list> <dbl> <dbl>
#>  Response Control Treatment    16    16      206. 0.00326  0.0000315 <chr [2]>        1     2

And the plot looks like this:

ggplot(data, aes(Group, Response)) +
  stat_summary(geom = "errorbar", fun.data = mean_cl_boot, width = 0.25) +
  stat_summary(geom = "crossbar", fun = mean, fun.min = mean, 
               fun.max = mean, fatten = FALSE, width = 0.5) +
  stat_pvalue_manual(wilcoxon_result, "P = {p}", 
                     size = 18 / ggplot2::.pt, tip.length = 0, 
                     bracket.size = 2 / ggplot2::.stroke) +
  theme_pubr(base_size = 18, x.text.angle = 45)