Compare variant effects between two GWAS — effect_compare

Scatter the per-variant effect sizes of two studies against each other for the variants they share, to assess replication and effect concordance (for example discovery versus replication, or two ancestries). A y = x reference line is drawn; points are colored by whether the two effects agree in sign among variants significant in either study.

Usage

effect_compare_plot(
  gwas1,
  gwas2,
  snp = NULL,
  beta = NULL,
  se = NULL,
  p = NULL,
  labels = c("Study 1", "Study 2"),
  p_threshold = 5e-08,
  show_ci = TRUE,
  ci = 0.95,
  label_top_n = 10,
  colors = c(concordant = "#2C7FB8", discordant = "#D7301F", ns = "#BDC3C7"),
  point_size = 1.8,
  title = NULL
)

Arguments

gwas1, gwas2: gwas_data objects or data.frames. Variants are matched on the SNP column.
snp, beta, se, p: Column name overrides applied to both inputs when they are plain data.frames.
labels: Length-2 character vector of axis labels.
p_threshold: Significance threshold used to classify variants.
show_ci: If TRUE, draw confidence-interval crosses for the significant variants (requires SE columns).
ci: Confidence level for the crosses.
label_top_n: Label the N most significant shared variants.
colors: Named colors for "concordant", "discordant" and "ns".
point_size: Point size.
title: Plot title.

Value

A ggplot object.

Examples

data(example_gwas)
g2 <- example_gwas
g2$BETA <- g2$BETA + stats::rnorm(nrow(g2), 0, 0.02)
effect_compare_plot(example_gwas, g2,
                    labels = c("Discovery", "Replication"),
                    p_threshold = 1e-3)