Plots top features importance — plot_top

Plots the top features importance per dataset and latent dimension.

Usage

plot_top_features(
  method_output,
  latent_dimensions = NULL,
  group_latent_dims = TRUE,
  datasets = NULL,
  n_features = 20,
  mo_data = NULL,
  label_cols = NULL,
  truncate = NULL,
  nrow = 1
)

Arguments

method_output: Integration method output generated via the get_output() function.
latent_dimensions: Character vector giving the latent dimensions to display. Default value is NULL, i.e. all latent dimensions will be shown.
group_latent_dims: Logical, for integrations methods that construct datasets- specific versions of each latent dimension, should these be grouped? e.g. when DIABLO constructs a snps- and rnaseq version of component 1, should the two be grouped as "Component 1"? Default value is TRUE.
datasets: Character vector giving the datasets to display. Default value is NULL, i.e. all datasets will be shown.
n_features: Integer, number of top features to display per dataset and latent dimension.
mo_data: A MultiDataSet object. Only used if label_cols is not NULL.
label_cols: Character or named list of character, giving for each dataset the name of the column in the corresponding features metadata to use as label. If one value, will be used for all datasets. If list, the names must correspond to the names of the datasets in mo_data. If a dataset is missing from the list or no value is provided, feature IDs will be used as labels. Alternatively, use feature_id to get the feature IDs as labels.
truncate: Integer, width to which the labels should be truncated (to avoid issues with very long labels in plots). If NULL (default value), no truncation will be performed.
nrow: Integer, number of rows over which the dataset panels should be plotted for each latent dimensions.

Value

A patchwork of plots.