Manhattan plot

R code of the ggplot2 example

library(ggplot2)
library(dplyr)
library(qqman)

## Simulated association tests along 22 chromosomes, one point per SNP,
## laid end to end; the line marks genome-wide significance (p = 5e-8).
snps <- gwasResults |>
  arrange(CHR, BP) |>
  mutate(position = row_number(), hit = P < 5e-8,
         shade = ifelse(hit, "hit", ifelse(CHR %% 2 == 0, "even", "odd")))
centres <- snps |>
  group_by(CHR) |>
  summarise(centre = mean(position))

p <- ggplot(snps, aes(x = position, y = -log10(P), colour = shade)) +
  geom_hline(yintercept = -log10(5e-8), colour = "#535353", linewidth = 0.3, linetype = "dashed") +
  geom_point(data = filter(snps, !hit), size = 0.5) +
  ## Significant SNPs: black outline, half-transparent fill.
  geom_point(data = filter(snps, hit), shape = 21, size = 1.6, colour = "black",
             fill = scales::alpha("#9b2367", 0.5), stroke = 0.4) +
  scale_colour_manual(values = c(odd = "#a6a6a6", even = "#535353", hit = "#9b2367"), guide = "none") +
  scale_x_continuous(breaks = centres$centre[c(1:10, seq(12, 22, 2))],
                     labels = centres$CHR[c(1:10, seq(12, 22, 2))], expand = expansion(add = 50)) +
  scale_y_continuous(expand = expansion(mult = c(0, 0.05))) +
  labs(title = "Six SNPs on chromosome 3 pass genome-wide significance",
       subtitle = "Simulated association study of 16 470 SNPs; dashed line at p = 5 × 10⁻⁸",
       x = "Chromosome", y = expression(-log[10](italic(p))),
       caption = "Data: simulated GWAS results, via the qqman R package") +
  theme_lex_viz(grid = "y", arrows = "y")