Skip to contents

Run outlier detection on per-cell QC metrics

run_cell_qc.Rd

Helper function to run initial quality control on cells.

Usage

run_cell_qc(
  metrics,
  cells_to_keep,
  directions = NULL,
  threshold = 3,
  groups = NULL,
  hard_thresholds = NULL,
  mad = TRUE
)

Arguments

metrics

Named list of numeric vectors.

cells_to_keep

Integer. 0-indexed cell positions.

directions

Named character vector, one of "twosided", "below", "above". Defaults to "twosided".

threshold

Numeric. MADs for MAD outlier detection.

groups

Optional grouping vector.

hard_thresholds

Optional named list of numeric vectors with lower and/or upper bounds, e.g. list(MT = c(upper = 15)). Applied independent of groups.

mad

Logical. If FALSE, skip MAD entirely; hard_thresholds must then be supplied.

Value

A CellQc object.

Examples

# MAD outlier detection over two metrics at once
set.seed(42L)
metrics <- list(
  lib_size = c(rnorm(99, 1000, 100), 50),
  pct_mt = runif(100, 0, 20)
)
run_cell_qc(
  metrics,
  cells_to_keep = 0:99,
  directions = c(lib_size = "below", pct_mt = "above")
)
#> CellQc: 100 cells, 5 outliers (5.0%)
#> Metrics:
#>   - lib_size: 5 outliers (mad = 5)
#>     MAD lower = 816.05
#>   - pct_mt: 0 outliers (mad = 0)
#>     MAD upper = 23.60