benchmark.integration: Benchmarks the cluster assignmentws

View source: R/benchmark.intergation.R

benchmark.integrationR Documentation

Benchmarks the cluster assignmentws

Description

Adopting scones method of Average Silhouette Width determination between batches. A depleted value is superior.

Usage

benchmark.integration(
  object,
  batch,
  assays,
  reduction,
  result.names,
  n.components = 2,
  threads = 1
)

Arguments

object

IBRAP S4 class object

batch

Character. Whcih column in the metadata dataframe contains the batches.

reduction

Character. Which reduction(s) should be used to observe integration.

result.names

Character. A vector of names for the benchmarking results.

n.components

Numerical. How many components of the reduced embeddings should be used. Default = 2

assay

Character. String containing indicating which assay to use

Value

Benchmarking scores for the supplied integrations

Examples

object <- benchmark.integration(object = object, 
                                          batch = 'original.project', 
                                          assays = c('SCT','SCRAN','SCANPY'), 
                                          reduction = c('pca_umap', 'pca_harmony_umap', 
                                                        'scanorama_umap', 'pca_bbknn_bbknn:umap',
                                                        'CCA_pca_umap'), 
                                          result.names = c('uncorrected', 'harmony', 
                                                           'scanorama', 'bbknn', 'cca'), 
                                          n.components = 2)


connorhknight/IBRAP documentation built on March 9, 2023, 7:01 p.m.