This function plots the observed EAF values for each feature in the dataset. The features are ordered by their observed EAF values. The confidence intervals are plotted as error bars or ribbons.
Usage
plot_EAF_values(
qsip_data_object,
confidence = 0.9,
success_ratio = 0.9,
top = Inf,
error = "none",
alpha = 0.3,
zero_line = TRUE,
shared_y = FALSE,
title = NULL
)
Arguments
- qsip_data_object
(qsip_data) A qsip_data object or list of qsip_data objects
- confidence
(numeric) The confidence level for the confidence interval
- success_ratio
(numeric) The ratio of successful resamples to total resamples
- top
(numeric) The number of top features to plot. Use
Inf
for all- error
(character) The type of error bars to plot. Options are 'none', 'bar', 'ribbon'
- alpha
(numeric) The transparency of the error bar/ribbon
- zero_line
(logical) Add a line at EAF = 0
(logical) Use a shared y-axis for the facets
- title
(character) An optional title of the plot
Details
Either a single qsip object or a list of named qsip objects can be passed to this function. If giving a list of qsip objects the plot will be faceted by the list names.
If the resampling step was run with the default allow_failures = FALSE
, then the
points will just be colored a generic blue. But, if resampling was instead run
with allow_failures = TRUE
, then the points are colored based on the success
ratio of the resamples. If giving a list of qsip objects then the pass/fail color scheme
will be applied if any of the qsip objects have allow_failures = TRUE
.