Plot feature contributions to selected samples — plot_contributions_to

Plot waterfall-style SHAP contributions for explicitly selected perturbations and samples. Training samples are used by default. Set source = "user" to plot samples predicted with add_powerup_predictions().

Usage

plot_contributions_to_sample(
  models,
  prepared,
  perturbations,
  samples,
  source = "training",
  n_features = 5,
  n_columns = 1,
  short_title = FALSE,
  fixed_axis = FALSE,
  axis_limits = c(-0.05, 1),
  nudges_lr = c(0, 0),
  values_are_percentages = TRUE,
  show_error = FALSE,
  highlight_significant = FALSE,
  labels_data = NULL,
  sec_label = NULL
)

Arguments

models: A named list of models returned by fit_powerup_models().
prepared: The prepared POWERUP object used for model training and prediction.
perturbations: Character vector of perturbation names to plot.
samples: Character vector of sample names to plot.
source: Either "training" (default) or "user".
n_features: Number of top feature contributions to show individually. Remaining contributions are combined.
n_columns: Number of columns in the resulting plot grid.
short_title: If TRUE, use a compact plot title.
fixed_axis: If TRUE, use axis_limits for every plot.
axis_limits: Numeric vector of length two used when fixed_axis = TRUE.
nudges_lr: Numeric vector of length two controlling contribution-label nudges.
values_are_percentages: If TRUE, display predictions and contributions as percentages.
show_error: If TRUE, shade the central 50 and 80 percent prediction intervals and show the 95 percent interval with dashed boundaries.
highlight_significant: If TRUE, highlight plots whose prediction interval supports the modeled event.
labels_data: Optional data frame containing old_label and new_label columns for perturbation-label replacement.
sec_label: If non-NULL, use labels_data$second_label as the subtitle for mapped perturbations.

Value

A combined ggplot object containing one waterfall plot per selected perturbation and sample.

Examples

if (FALSE) { # \dontrun{
plot_contributions_to_sample(
  models,
  prepared,
  perturbations = "CTNNB1",
  samples = "ACH-000957"
)

models <- add_powerup_predictions(models, prepared)
plot_contributions_to_sample(
  models,
  prepared,
  perturbations = "CTNNB1",
  samples = "ACH-000957",
  source = "user"
)
} # }