Remove outlier pseudobulk samples using PCA-based Mahalanobis distance

Remove outlier pseudobulk samples using PCA-based Mahalanobis distance

Usage

remove_outliers(
  eset,
  n_components = 10,
  threshold_quantile = 0.99,
  split_by = NULL,
  min_samples = 10,
  verbose = TRUE
)

Arguments

eset

An ExpressionSet object.

n_components

Integer. Number of PCA components. Default 10.

threshold_quantile

Numeric in (0,1). Chi-squared quantile for outlier cutoff. Default 0.99.

split_by

Character or NULL. Column in pData to split by before outlier detection. If NULL, all samples are analyzed together. Default NULL.

min_samples

Integer. Minimum samples in a group to run outlier detection. Groups below this are kept as-is. Default 10.

verbose

Logical. Default TRUE.

Value

ExpressionSet with outlier samples removed.