detect_batch_effect_express: Quicker function for detection of batch effects

In shdam/cyCombine: Robust integration of single-cell cytometry datasets within and across technologies

Quicker function for detection of batch effects

Description

This function can be used to check if a dataset contains batch effects. The function employs three different approaches to detect the effects:

1. The Earth Mover's Distance per marker when comparing each batch-batch pair.

2. Density plots per marker, per batch for visual inspection.

3. A MultiDimensional Scaling plot based on the median marker expression per sample. It can apply downsampling for a quicker analysis of larger datasets.

Usage

detect_batch_effect_express(
  df,
  out_dir,
  batch_col = "batch",
  downsample = NULL,
  seed = 472
)

Arguments

df	Tibble containing the expression data and batch information. See prepare_data.
out_dir	Directory for plot output
batch_col	Name of column containing batch information
downsample	Number of cells to include in detection. If not specified, all cells will be used.
seed	Random seed for reproducibility

See Also

Other detect_batch_effect: detect_batch_effect()

Examples

## Not run: 
detect_batch_effect_express(df = exprs, out_dir = '/my/cycombine/dir/')
detect_batch_effect_express(df = exprs, downsample = 100000, out_dir = '/my/cycombine/dir/')

## End(Not run)

shdam/cyCombine documentation built on June 12, 2022, 11:29 p.m.