Dose-response curves

The chart we are set to recreate, originally obtained from the article “The discriminatory power of the T cell receptor” (doi: 10.7554/eLife.67092) published in eLife, stands as a prime example of a dose-response curves chart.

ORIGINAL CHART: Example dose-responses for naïve T cells.

A dose-response curves chart is a graphical representation used in various scientific fields, including pharmacology, toxicology, and environmental science, to illustrate the association between the dose of a substance and the response it elicits in a biological system. The x-axis typically represents the dose or concentration of the substance administered, while the y-axis represents the magnitude of the response.

Used geometries

Main geometries:

geom_line()
geom_point()

Secondary geometries:

geom_text()
geom_hline()

We need to load the following packages:

library(drc)
library(tidyverse)
library(here)
library(broom)
library(ggtext)

We load the file drc.xlsx in R.

library(readxl)
drc <- read_excel("data/drc.xlsx")

head(drc)

# A tibble: 6 × 49
    dose NYE9V1 NYE6V1 NYE3Y1 NYE6T1 NYE4D1 NYE4A1 NYE5Y1 NYE5F1 NYE9V2 NYE6V2
   <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>  <dbl>
1 125      95     95.6   92     95.5   89.5   55.3   89.6   91.3   83.9   78.5
2  25      96.9   96.3   95.3   94.8   91.5   65.5   85.1   81.5   86.4   79.8
3   5      96.9   97     95.2   95.8   92.1   50     89.8   74.4   86.7   83.2
4   1      97.7   95.9   94.8   95.3   91     46.9   86     53.9   82.2   84.7
5   0.2    96.7   96.4   95     94.2   89     41.3   74.8   41.4   84.4   80.9
6   0.04   97.4   94.9   92.5   91.9   85.7   35.7   44.7   36.0   70.1   67.3
# ℹ 38 more variables: NYE3Y2 <dbl>, NYE6T2 <dbl>, NYE4D2 <dbl>, NYE4A2 <dbl>,
#   NYE5Y2 <dbl>, NYE5F2 <dbl>, NYE9V3 <dbl>, NYE6V3 <dbl>, NYE3Y3 <dbl>,
#   NYE6T3 <dbl>, NYE4D3 <dbl>, NYE4A3 <dbl>, NYE5Y3 <dbl>, NYE5F3 <dbl>,
#   NYE9V4 <dbl>, NYE6V4 <dbl>, NYE3Y4 <dbl>, NYE6T4 <dbl>, NYE4D4 <dbl>,
#   NYE4A4 <dbl>, NYE5Y4 <dbl>, NYE5F4 <dbl>, NYE9V5 <dbl>, NYE6V5 <dbl>,
#   NYE3Y5 <dbl>, NYE6T5 <dbl>, NYE4D5 <dbl>, NYE4A5 <dbl>, NYE5Y5 <dbl>,
#   NYE5F5 <dbl>, NYE9V6 <dbl>, NYE6V6 <dbl>, NYE3Y6 <dbl>, NYE6T6 <dbl>, …

We have eight peptides labeled as “NYE9V”, “NYE6V”, “NYE3Y”, “NYE6T”, “NYE4D”, “NYE4A”, “NYE5Y”, and “NYE5F”, each with six replicates denoted from 1 to 6 (e.g., NYE9V1, NYE9V2, NYE9V3, etc.). The data values for these peptides appear to be within the range of 0 to 100. To represent the data, we will calculate the mean value for each dose per peptide.

drc_tidy <- drc |> 
  pivot_longer(cols = -dose,
               names_to = "sample",
               values_to = "response") |> 
  mutate(
    peptides = str_sub(sample, 1, 5),
    replicate = str_sub(sample, 6, 6)
  )


drc_per_peptide <- drc_tidy |> 
  group_by(dose, peptides) |> 
  summarize(
    #sem = sd(response)/sqrt(n()),
            response = mean(response)
            )


model <- drm(data = drc_tidy,
             formula = response ~ dose,
             curveid = peptides,
             fct = LL.4(names = c("Hill slope", "Min", "Max", "EC50")))  


predicted_data <- data.frame(
  dose = rep(drc$dose, times = 8),
  peptides = rep(c("NYE9V", "NYE6V", "NYE3Y", "NYE6T", 
                    "NYE4D", "NYE4A", "NYE5Y", "NYE5F"), 
                  each = nrow(drc))
  
)

predicted_data$prediction <- predict(model, newdata = predicted_data)



drc_per_peptide <- drc_per_peptide |> 
  mutate(peptides = factor(peptides, levels =  c("NYE9V", "NYE6V", "NYE3Y", 
                                                   "NYE6T", "NYE4D", "NYE4A", 
                                                   "NYE5Y", "NYE5F"),
                            labels =  c("NYE 9V", "NYE 6V", "NYE 3Y", 
                                        "NYE 6T", "NYE 4D", "NYE 4A", 
                                        "NYE 5Y", "NYE 5F"))
         )


predicted_data <- predicted_data |> 
  mutate(peptides = factor(peptides, levels =  c("NYE9V", "NYE6V", "NYE3Y", 
                                                   "NYE6T", "NYE4D", "NYE4A", 
                                                   "NYE5Y", "NYE5F"),
                            labels =  c("NYE 9V", "NYE 6V", "NYE 3Y", 
                                        "NYE 6T", "NYE 4D", "NYE 4A", 
                                        "NYE 5Y", "NYE 5F"))
         )

We also set particular colors for the selected countries:

# Colors

my_colors <- c("#E81B67", "#F5921C", "#6E4C9F", "#97CB62",
               "#6BD2C7", "#3D78C7", "#3D78c7", "#3952A4", "#2B267C")

drc_per_peptide |> 
  ggplot(aes(x = log10(dose), y = response, color = peptides)) + 
  geom_point(size = 3) +
  geom_line(data = predicted_data, linewidth = 1.3, show.legend = FALSE,
            aes(x = log10(dose), y = prediction)) +
  geom_hline(yintercept = 15, col="grey20", linetype="dotted", linewidth = 0.8
  ) +
  scale_y_continuous(
    limits = c(0, 115), 
    breaks = seq(0, 100, 50),
    minor_breaks = seq(0, 115, 10)
  ) +
  scale_x_continuous(
    limits = c(-6, 3), 
    breaks = -6:3,
    expand = c(0,0),
    minor_breaks = log10(rep(1:9, 7)*(10^rep(-10:3, each = 9))),
    labels = function(lab) {
      do.call(
        expression,
        lapply(paste(lab), function(x) bquote("10"^.(x)))
      )
    }
  ) +
  guides(x = guide_axis(minor.ticks = TRUE),
         y = guide_axis(minor.ticks = TRUE),
         color = guide_legend(override.aes = list(size = 3.5))
         ) +
  scale_color_manual(
    labels = paste("<span style='color:",
                                   my_colors,
                                   "'>",
                   levels(drc_per_peptide$peptides),
                                   "</span>"),
    values = my_colors) +
  annotate("text", x = log(10^-2.45), y = 20, size = 6,
           label = expression("P"["15"]), color = "grey20") +
  annotate("text", x = log(10^-2.55), y = 110, hjust = 0, size = 6,
           label = "1G-4 electroporated \nnaïve T cells", color = "grey20") +
  labs(x = "Peptide (μΜ)", y = "Normalised \nCD69-positive cells (%)",
       tag = "B") +
  theme(panel.background = element_blank(),
        axis.title = element_text(size = 18),
        axis.text = element_text(size = 16),
        axis.line = element_line(color="black", linewidth = 0.7),
        axis.ticks.length = unit(10, "pt"),
        axis.minor.ticks.length = rel(0.5),
        axis.ticks = element_line(size = 0.9),
        legend.title = element_blank(),
        legend.text = element_markdown(size = 16),
        legend.key.size = unit(1.1, "cm"),
        legend.key.width = unit(0.9,"cm"),
        legend.margin = margin(l = -0.8, b = -1.45, unit = "cm"),
        plot.tag = element_text(size = 40),
        plot.tag.position = c(0.01, 0.96)
        )