Skip to contents

SuSiE Plots.

Source: R/susie_plot.R
susie_plots.Rd

susie_plot produces a per-variable summary of the SuSiE credible sets. susie_plot_iteration produces a diagnostic plot for the susie model fitting. For susie_plot_iteration, several plots will be created if track_fit = TRUE when calling susie.

Usage

susie_plot(
  model,
  y,
  add_bar = FALSE,
  pos = NULL,
  b = NULL,
  max_cs = 400,
  add_legend = NULL,
  ...
)

susie_plot_iteration(model, L, file_prefix, pos = NULL)

Arguments

model

A SuSiE fit, typically an output from susie or one of its variants. For suse_plot, the susie fit must have model$z, model$PIP, and may include model$sets. model may also be a vector of z-scores or PIPs.

y

A string indicating what to plot: either "z_original" for z-scores, "z" for z-score derived p-values on (base-10) log-scale, "PIP" for posterior inclusion probabilities, "log10PIP" for posterior inclusion probabiliities on the (base-10) log-scale. For any other setting, the data are plotted as is.

add_bar

If add_bar = TRUE, add horizontal bar to signals in credible interval.

pos

Indices of variables to plot. If pos = NULL all variables are plotted.

b

For simulated data, set b = TRUE to highlight "true" effects (highlights in red).

max_cs

The largest credible set to display, either based on purity (set max_cs between 0 and 1), or based on size (set max_cs > 1).

add_legend

If add_legend = TRUE, add a legend to annotate the size and purity of each CS discovered. It can also be specified as location where legends should be added, e.g., add_legend = "bottomright" (default location is "topright").

...

Additional arguments passed to plot.

L

An integer specifying the number of credible sets to plot.

file_prefix

Prefix to path of output plot file. If not specified, the plot, or plots, will be saved to a temporary directory generated using tempdir.

Value

Invisibly returns NULL.

See also

susie_plot_changepoint

Examples

set.seed(1)
n <- 1000
p <- 1000
beta <- rep(0, p)
beta[sample(1:1000, 4)] <- 1
X <- matrix(rnorm(n * p), nrow = n, ncol = p)
X <- scale(X, center = TRUE, scale = TRUE)
y <- drop(X %*% beta + rnorm(n))
res <- susie(X, y, L = 10)
susie_plot(res, "PIP")

susie_plot(res, "PIP", add_bar = TRUE)

susie_plot(res, "PIP", add_legend = TRUE)

susie_plot(res, "PIP", pos = 1:500, add_legend = TRUE)

# Plot selected regions with adjusted x-axis position label
res$genomic_position <- 1000 + (1:length(res$pip))
susie_plot(res, "PIP",
  add_legend = TRUE,
  pos = list(attr = "genomic_position", start = 1000, end = 1500)
)

# True effects are shown in red.
susie_plot(res, "PIP", b = beta, add_legend = TRUE)


set.seed(1)
n <- 1000
p <- 1000
beta <- rep(0, p)
beta[sample(1:1000, 4)] <- 1
X <- matrix(rnorm(n * p), nrow = n, ncol = p)
X <- scale(X, center = TRUE, scale = TRUE)
y <- drop(X %*% beta + rnorm(n))
res <- susie(X, y, L = 10)
susie_plot_iteration(res, L = 10)
#> Iterplot saved to /tmp/RtmpNk3Z5y/susie_plot.pdf