Skip to contents

Extract Top Contributing Features for a DIVAS Component

Source: R/getTopFeatures.R
getTopFeatures.Rd

Identifies the most important features contributing to a specified component within a given modality by ranking features based on their loading values.

Usage

getTopFeatures(
  divasRes,
  compName,
  modName,
  n_top_pos = 100,
  n_top_neg = 100,
  return_values = FALSE
)

Arguments

divasRes

A DIVAS result object containing a Loadings element.

compName

Character string specifying the component name.

modName

Character string specifying the modality name.

n_top_pos

Integer specifying the number of top positive features to return. Default is 100.

n_top_neg

Integer specifying the number of top negative features to return. Default is 100.

return_values

Logical indicating whether to return loading values along with feature names. Default is FALSE.

Value

If return_values = FALSE, returns a list with two elements:

  • top_positive: Character vector of feature names with highest positive loadings

  • top_negative: Character vector of feature names with highest negative loadings (most negative)

If return_values = TRUE, returns a list with two elements:

  • top_positive: Named numeric vector of top positive loadings

  • top_negative: Named numeric vector of top negative loadings

Details

Features are ranked by their loading values for the specified component. Positive loadings indicate features that vary in the same direction as the component, while negative loadings indicate features that vary in the opposite direction. The function allows asymmetric selection of top positive and negative features, which can be useful when one direction is of greater biological interest.

Examples

if (FALSE) { # \dontrun{
# Get top 100 positive and 100 negative features
top_features <- getTopFeatures(
  divasRes = divasRes,
  compName = "CD4_T_combined+CD8_T_combined+CD14_Monocyte+NK+proteomics+metabolomics-1",
  modName = "CD4_T_combined"
)

# Get top 50 positive and 200 negative features with loading values
top_features_values <- getTopFeatures(
  divasRes = divasRes,
  compName = "CD4_T_combined+CD8_T_combined+CD14_Monocyte+NK+proteomics+metabolomics-1",
  modName = "CD4_T_combined",
  n_top_pos = 50,
  n_top_neg = 200,
  return_values = TRUE
)

# Access results
pos_features <- top_features$top_positive
neg_features <- top_features$top_negative
} # }