Skip to contents

Score an embedding or a knn graph with the kBET test

Source: R/metrics_kbet.R
score-kbet.Rd

Compute scores based on the k-nearest neighbour batch effect test (kBET). Accepts a DimReduc instance or a graph object (Graph or Neighbor) provided that it contains distances or connectivities.

If 'what' is a DimReduc object, a knn graph is computed for k = 50. Then, (or if 'what' is a Graph or a Neighbor object), connectivities are derives from 50 nearest neighbours distances. Finally, the kBET test is computed for each cell-type label of all provided cell.var. kBET measures if local batch label distribution is similar to the global one. The result of kBET is the average test rejection rate (between 0 and 1). The closest to zero, the less bias is attributable to the batch effect.

Usage

ScoreKBET(
  object,
  batch.var,
  cell.var,
  what,
  graph.type = c("distances", "connectivities"),
  seed.use = 42L,
  verbose = TRUE,
  assay = NULL,
  layer = NULL
)

AddScoreKBET(
  object,
  integration,
  batch.var,
  cell.var,
  what,
  graph.type = c("distances", "connectivities"),
  seed.use = 42L,
  verbose = TRUE,
  assay = NULL,
  layer = NULL
)

Arguments

object

A Seurat object

batch.var

The name of the batch variable (must be in the object metadata)

cell.var

The name of the column with cell type label variable (must be in the object metadata). If NULL is passed, the kBET score is computed for all cells at once.

what

the name of the dimension reduction or Graph/Neighbor object to score. Must be in the Seurat object.

graph.type

one of 'distances' (default) or 'connectivities', to indicate the type of graph when what is not a dimension reduction.

seed.use

the value of the seed to obtain reproducible results. NULL disables to use of a seed

verbose

whether to print progress messages

assay

assay to use. Passed to Seurat to automatically construct the batch.var when not provided (DefaultAssay() is used by default). Useless otherwise

layer

layer to use. Passed to Seurat to automatically construct the batch.var when not provided. Useless otherwise

integration

name of the integration to score

Value

ScoreKBET: a list with one element per cell.var. Each is a named numeric vector of kBET score (floats between 0 and 1), one per cell-label.

AddScoreKBET: the updated Seurat object with the mean kBET score(s) set for the integration.

References

Büttner, M., Miao, Z., Wolf, F. A., Teichmann, S. A. & Theis, F. J. A test metric for assessing single-cell RNA-seq batch correction. Nat Methods 16, 43–49 (2018). DOI

Luecken, M. D., Büttner, M., Chaichoompu, K., Danese, A., Interlandi, M., Mueller, M. F., Strobl, D. C., Zappia, L., Dugas, M., Colomé-Tatché, M. & Theis, F. J. Benchmarking atlas-level data integration in single-cell genomics. Nat Methods 19, 41–50 (2021). DOI