Identifies specific and non-specific features based on the Gini inequality
coefficient computed on ranked expression values.
Usage
gini.rank.fs(mat, gini_thresh = 0.5, noise_thresh = NULL, num_threads = -1)
Arguments
- mat
A matrix with features as rows and cells as columns.
- gini_thresh
Numeric; Gini threshold for selecting specific features.
- noise_thresh
Integer; minimum number of cells required for a feature to be retained.
- num_threads
Integer; number of threads to use. -1 uses all available cores.
Value
A list containing specific_features, non_specific_features,
and no_occurrence.
Details
This method removes globally low-ranked features while retaining features
with consistently high ranks across cells for downstream analysis.
Examples
if (FALSE) { # \dontrun{
fs <- gini.rank.fs(mat)
} # }