Add p values not working :((

[uak20]

Hi guys - i'm trying to plot this graph (csv with data attached), it's a plot comparing two groups' ages, with error bars for SEM, and i want to plot the P value of the difference between the two groups. i keep getting the error

! Problem while computing aesthetics.
ℹ Error occurred in the 4th layer.
Caused by error:
! object 'group' not found

Would be grateful for any advice - code below! thanks!

# Load the necessary libraries
library(rstatix)
library(dplyr)
library(ggplot2)
library(ggprism)
library(ggpubr)
library(extrafont)
library(patchwork)
library(readr)
library(tidyr)
library(magrittr)

# Load the data, group1 control, group2 treatment
A_Age <- read.csv("A Age.csv")

# Calculate the mean age, standard deviation and standard error for each group
mean_age <- A_Age %>%
  group_by(group) %>%
  summarise(Mean_Age = mean(age), SD_Age = sd(age), SE_Age = sd(age) / sqrt(n()), .groups = 'drop')

# Perform a standard t-test (assuming equal variances)
t_test <- t.test(age ~ group, data = A_Age, var.equal = TRUE)
p_value <- t_test$p.value
p_value <- round(p_value, 4)

# Determine the upper limit for the y-axis as the next multiple of 20 above the max Mean_Age
max_y_value <- max(mean_age$Mean_Age + mean_age$SD_Age)
max_y <- ceiling(max_y_value / 15) * 15

# # Create a data frame for p-value annotation
df_p_val <- tibble::tribble(
  ~group1, ~group2, ~p, ~y.position,
  "Control", "Treatment", round(p_value,4), max_y_value + max_y_value/20
)

# Set Baskerville font family
loadfonts()

# Create the bar plot with ggprism
plotAge <- ggplot(mean_age, aes(x = group, y = Mean_Age, fill = group)) +
  geom_bar(stat = "identity",
           position = position_dodge(),
           width = .75,
           color = "black",
           linewidth = .25) +
  geom_errorbar(aes(ymin = Mean_Age - SD_Age,
                    ymax = Mean_Age + SD_Age),
                color = "black",
                position = position_dodge(0.5), 
                width = 0.2,
                linewidth = .25) +
  theme_prism(axis_text_angle = 45, base_size = 14) +
  theme(text = element_text(family = "Source Sans 3"),  # Specify Baskerville font for all text
        plot.title = element_text(face = "bold"),  # Make the title bold
        axis.title = element_text(face = "plain"),  # Make axis titles non-bold
        axis.text = element_text(face = "plain"),  # Make axis labels non-bold
        axis.line = element_line(linewidth = .5),  # Adjust axis line thickness here
        axis.ticks = element_line(linewidth = .5),
        axis.ticks.length.y = unit(0.1, "cm"),
        prism.ticks.length = unit(0.05, "cm"),
        plot.margin = margin(0.25, 0.25, 0.25, 0.25, "cm")) +  # Adjust plot margins here
  labs(y = "Mean Age (years)", x = NULL, title = NULL) +
  theme(legend.position = "none") +
  scale_y_continuous(expand = c(0, 0), limits = c(0, max_y), breaks = seq(0, max_y, by = 10), minor_breaks = seq(0, max_y, by = 2), guide = "prism_minor") +
  # scale_fill_prism(palette = "colors")
  scale_fill_manual(values = c("Control" = "#4771FF", "Treatment" = "#FF9B57"))  # Use specific color codes for accuracy

plotAge <- plotAge + add_pvalue(df_p_val,
                                xmin = "group1",
                                xmax = "group2",
                                label = "p",
                                y.position = "y.position")

ggsave("Age.png", plotAge, units = "px", dpi = 900, height = 3500, width = 2500)

CSV with data

A Age.csv

[csdaw]

Hi there,

Thanks for the reproducible example. The error message is a bit opaque unfortunately, but to fix this you just need to remove aes(fill = group) as a global aesthetic, and put it into geom_bar() only.

From this:

plotAge <- ggplot(mean_age, aes(x = group, y = Mean_Age, fill = group)) +
  geom_bar(stat = "identity",
           position = position_dodge(),
           width = .75,
           color = "black",
           linewidth = .25) +
  ...

To this:

plotAge <- ggplot(mean_age, aes(x = group, y = Mean_Age)) +
  geom_bar(aes(fill = group),
           stat = "identity",
           position = position_dodge(),
           width = .75,
           color = "black",
           linewidth = .25) +
  ...

add_pvalue() does not understand the fill aesthetic (as there is no shapes to fill!). Where it is defined as a global aesthetic (as in your example), add_pvalue() doesn't handle it gracefully and produces this error. In future if I re-write add_pvalue() I will fix this.

Hope this helps,
Charlotte