Plot COLOC results

ggColocManhattanSingle.R

ggColocManhattanSingle <- function(
    summ.1, chr = "CHR", bp = "BP", p = "P", snp = "SNP",
    summ.1.name = expression(paste("", -log[10](P))),
    logp = TRUE, size = 2, lead.snp = NULL, r2 = NULL, ...
) {
  
  if (!is.null(summ.1[[snp]])) {
    d1 <- data.frame(CHR = summ.1[[chr]], BP = summ.1[[bp]], P = summ.1[[p]], SNP = summ.1[[snp]])
  } else {
    d1 <- data.frame(CHR = summ.1[[chr]], BP = summ.1[[bp]], P = summ.1[[p]])
  }
  
  # Calculate -log(P) for summ.1 and summ.2
  if (logp) {
    d1$logp <- -log10(d1$P)
  } else {
    d1$logp <- d1$P
  }
  
  # Map SNP r2 to discrete colors
  # This is inspired by locuscomparer package
  if (is.null(r2)) {
    snp.col <- rep("grey", nrow(d1))
  } else {
    snp.col <- as.character(cut(r2, breaks = c(0, 0.2, 0.4, 0.6, 0.8, 1),
                                labels = c('blue','lightblue','darkgreen','orange','red'),
                                include.lowest = T))
  }
  
  g1 <- ggplot() +
    cowplot::theme_cowplot(font_size = 4, line_size = 0.2, font_family = "Helvetica") +
    ggrastr::rasterise(geom_point(data = d1, mapping = aes(x = BP / 1e6, y = logp),
                                  col = snp.col, size = 0.5), dpi = 960, scale = 0.5) +
    ggrastr::rasterise(geom_point(data = d1 %>% filter(BP %in% lead.snp),
                                  mapping = aes(x = BP / 1e6, y = logp),
                                  col = "purple", shape = 18, size = 1), dpi = 960) +
    scale_y_continuous(limits = c(0, max(d1$logp) * 1.05), expand = c(0, 0)) +
    scale_x_continuous(expand = c(0, 0)) +
    xlab("") +
    ylab(summ.1.name) +
    theme(plot.margin = unit(c(0, 0, 0, 0), "in"), axis.title.y = ggtext::element_markdown())
  
  return(g1)
}

This function plots a single Manhattan plot. Two plots can be combined to visualize colocalization results.