A sigmoidal dose-response curve that plots biological response against drug concentration on a logarithmic x-axis, fitted using a four-parameter logistic (4PL) model. This visualization is essential for determining drug potency metrics such as EC50 (half-maximal effective concentration) or IC50 (half-maximal inhibitory concentration), Hill slope steepness, and upper/lower response asymptotes. It enables rapid visual comparison of compound efficacy and is a standard tool in pharmacological analysis.
#' anyplot.ai
#' curve-dose-response: Pharmacological Dose-Response Curve
#' Library: ggplot2 3.5.1 | R 4.4.1
#' Quality: 94/100 | Created: 2026-06-24
library(ggplot2)
library(dplyr)
library(tibble)
library(ragg)
set.seed(42)
# Theme tokens
THEME <- Sys.getenv("ANYPLOT_THEME", "light")
PAGE_BG <- if (THEME == "light") "#FAF8F1" else "#1A1A17"
ELEVATED_BG <- if (THEME == "light") "#FFFDF6" else "#242420"
INK <- if (THEME == "light") "#1A1A17" else "#F0EFE8"
INK_SOFT <- if (THEME == "light") "#4A4A44" else "#B8B7B0"
INK_MUTED <- if (THEME == "light") "#6B6A63" else "#A8A79F"
IMPRINT_PALETTE <- c(
"#009E73", "#C475FD", "#4467A3", "#BD8233",
"#AE3030", "#2ABCCD", "#954477", "#99B314"
)
# Data — concentrations in nM, 10 points per compound spanning 6 decades
conc_obs_nM <- 10^seq(-1, 5, length.out = 10)
resp_a_true <- 4 + (94 - 4) / (1 + (20 / conc_obs_nM)^1.8)
resp_b_true <- 6 + (88 - 6) / (1 + (800 / conc_obs_nM)^1.1)
df <- tibble(
conc = rep(conc_obs_nM, 2),
response = c(
pmax(0, pmin(100, resp_a_true + rnorm(10, 0, 4))),
pmax(0, pmin(100, resp_b_true + rnorm(10, 0, 4)))
),
sem = c(runif(10, 1.5, 3.5), runif(10, 1.5, 3.5)),
compound = rep(c("Compound A", "Compound B"), each = 10)
)
# Fit 4PL models via nonlinear least squares
fit_a <- nls(
response ~ Bottom + (Top - Bottom) / (1 + (EC50 / conc)^Hill),
data = filter(df, compound == "Compound A"),
start = list(Bottom = 4, Top = 94, EC50 = 20, Hill = 1.5)
)
fit_b <- nls(
response ~ Bottom + (Top - Bottom) / (1 + (EC50 / conc)^Hill),
data = filter(df, compound == "Compound B"),
start = list(Bottom = 6, Top = 88, EC50 = 800, Hill = 1.0)
)
# Fine concentration grid for smooth curves
conc_fine <- 10^seq(-1.5, 5.5, length.out = 400)
coef_a <- coef(fit_a)
coef_b <- coef(fit_b)
vcov_a <- vcov(fit_a)
pred_a <- predict(fit_a, newdata = data.frame(conc = conc_fine))
pred_b <- predict(fit_b, newdata = data.frame(conc = conc_fine))
# 95% confidence band for Compound A via delta method (gradient of 4PL w.r.t. params)
se_a <- sapply(conc_fine, function(x) {
B <- coef_a["Bottom"]; Tp <- coef_a["Top"]
E <- coef_a["EC50"]; H <- coef_a["Hill"]
r <- (E / x)^H; denom <- 1 + r
g <- c(r / denom, 1 / denom,
-(Tp - B) * H * r / (E * denom^2),
-(Tp - B) * r * log(E / x) / denom^2)
sqrt(as.numeric(t(g) %*% vcov_a %*% g))
})
df_curves <- tibble(
conc = rep(conc_fine, 2),
response = c(pred_a, pred_b),
compound = rep(c("Compound A", "Compound B"), each = 400)
)
df_ci_a <- tibble(
conc = conc_fine,
ymin = pred_a - 1.96 * se_a,
ymax = pred_a + 1.96 * se_a
)
# EC50 and half-response values for crosshair annotations
ec50_a <- coef_a["EC50"]
ec50_b <- coef_b["EC50"]
hmid_a <- (coef_a["Bottom"] + coef_a["Top"]) / 2
hmid_b <- (coef_b["Bottom"] + coef_b["Top"]) / 2
x_left <- 10^(-1.5)
x_right <- 10^5.5
comp_colors <- c(
"Compound A" = IMPRINT_PALETTE[1],
"Compound B" = IMPRINT_PALETTE[2]
)
# Plot
p <- ggplot() +
# 95% CI ribbon — Compound A only
geom_ribbon(
data = df_ci_a,
aes(x = conc, ymin = ymin, ymax = ymax),
fill = IMPRINT_PALETTE[1],
alpha = 0.18
) +
# Top and bottom asymptote dashed lines (Compound A)
geom_hline(
yintercept = coef_a["Top"],
linetype = "dashed", color = INK_SOFT, linewidth = 0.4
) +
geom_hline(
yintercept = coef_a["Bottom"],
linetype = "dashed", color = INK_SOFT, linewidth = 0.4
) +
# Asymptote labels at right margin
annotate(
"text",
x = x_right * 0.88, y = coef_a["Top"] + 3.5,
label = "Top", color = INK_MUTED, size = 2.6, hjust = 1, fontface = "italic"
) +
annotate(
"text",
x = x_right * 0.88, y = coef_a["Bottom"] + 3.5,
label = "Bottom", color = INK_MUTED, size = 2.6, hjust = 1, fontface = "italic"
) +
# EC50 crosshair — Compound A
annotate(
"segment",
x = ec50_a, xend = ec50_a, y = -3, yend = hmid_a,
linetype = "dashed", color = IMPRINT_PALETTE[1], linewidth = 0.7
) +
annotate(
"segment",
x = x_left, xend = ec50_a, y = hmid_a, yend = hmid_a,
linetype = "dashed", color = IMPRINT_PALETTE[1], linewidth = 0.7
) +
# EC50 crosshair — Compound B
annotate(
"segment",
x = ec50_b, xend = ec50_b, y = -3, yend = hmid_b,
linetype = "dashed", color = IMPRINT_PALETTE[2], linewidth = 0.7
) +
annotate(
"segment",
x = x_left, xend = ec50_b, y = hmid_b, yend = hmid_b,
linetype = "dashed", color = IMPRINT_PALETTE[2], linewidth = 0.7
) +
# Fitted 4PL curves
geom_line(
data = df_curves,
aes(x = conc, y = response, color = compound),
linewidth = 1.4
) +
# Observed data: error bars then points (bars behind points)
geom_errorbar(
data = df,
aes(x = conc, ymin = response - sem, ymax = response + sem, color = compound),
width = 0.08,
linewidth = 0.7
) +
geom_point(
data = df,
aes(x = conc, y = response, color = compound),
size = 2.5,
shape = 21,
fill = PAGE_BG,
stroke = 1.2
) +
# EC50 label annotations with background box for visual weight
annotate(
"label",
x = ec50_a * 2.0, y = hmid_a - 10,
label = sprintf("EC50 = %.0f nM", ec50_a),
color = IMPRINT_PALETTE[1],
fill = ELEVATED_BG,
label.size = 0.15,
label.padding = unit(0.12, "lines"),
size = 3.0, hjust = 0, fontface = "italic"
) +
annotate(
"label",
x = ec50_b * 1.8, y = hmid_b + 8,
label = sprintf("EC50 = %.0f nM", ec50_b),
color = IMPRINT_PALETTE[2],
fill = ELEVATED_BG,
label.size = 0.15,
label.padding = unit(0.12, "lines"),
size = 3.0, hjust = 0, fontface = "italic"
) +
# Potency ratio annotation
annotate(
"text",
x = 10^4, y = 30,
label = sprintf("Potency ratio (B/A): %.0f×", ec50_b / ec50_a),
color = INK_MUTED, size = 2.8, hjust = 0.5, fontface = "italic"
) +
# Axis scales
scale_x_log10(
breaks = c(0.1, 1, 10, 100, 1000, 10000, 100000),
labels = c("0.1", "1", "10", "100", "1k", "10k", "100k"),
limits = c(x_left, x_right),
name = "Concentration (nM)"
) +
scale_y_continuous(
breaks = seq(0, 100, 20),
limits = c(-5, 108),
name = "Response (%)"
) +
scale_color_manual(values = comp_colors, name = NULL) +
labs(title = "curve-dose-response · r · ggplot2 · anyplot.ai") +
theme_minimal(base_size = 8) +
theme(
plot.background = element_rect(fill = PAGE_BG, color = PAGE_BG),
panel.background = element_rect(fill = PAGE_BG, color = NA),
panel.grid.major = element_line(color = INK_MUTED, linewidth = 0.15),
panel.grid.minor = element_blank(),
panel.border = element_blank(),
axis.line.x = element_line(color = INK_SOFT, linewidth = 0.4),
axis.line.y = element_line(color = INK_SOFT, linewidth = 0.4),
axis.ticks = element_line(color = INK_SOFT, linewidth = 0.3),
axis.title = element_text(color = INK, size = 10),
axis.text = element_text(color = INK_SOFT, size = 8),
plot.title = element_text(
color = INK, size = 12, hjust = 0.5,
margin = margin(b = 8)
),
legend.background = element_rect(
fill = ELEVATED_BG, color = INK_MUTED, linewidth = 0.25
),
legend.text = element_text(color = INK_SOFT, size = 8),
legend.key.size = unit(1.2, "lines"),
legend.position = "inside",
legend.position.inside = c(0.13, 0.78),
plot.margin = margin(t = 10, r = 15, b = 10, l = 10)
)
# Save
ggsave(
filename = sprintf("plot-%s.png", THEME),
plot = p,
device = ragg::agg_png,
width = 8,
height = 4.5,
units = "in",
dpi = 400
)
Part of Pharmacological Dose-Response Curve on anyplot.ai.