Identify artifact genes

For each gene in each dataset, compute how many cells in that dataset have more UMI than a specified threshold (umi.count.threshold) of that gene. A plot comparing the largest number across data sets with the third largest number is generated. For the majority of genes, these values are expected to be similar, and therefore lie on a diagonal. Genes that exhibit differences (difference.threshold) between the largest and third-largest number of cells are flagged and returned. This approach is adopted from work by Lause, Berens, Kobak (2021) BioRxiv (See Figure S6)

findArtifactGenes(
  object,
  assay = NULL,
  features = NULL,
  meta.feature = "Barcode",
  umi.count.threshold = 5,
  nth.max = 2,
  difference.threshold = 30,
  group.specific.is.artefact = F,
  verbose = F
)

Arguments

object

seurat object or named list of seurat objects.

assay

assay containing count matrix. If unspecified, "RNA" assay is used. If "RNA" assay is missing, the default assay is used.

features

genes used for analysis. If unspecified, variable genes are used.

meta.feature

feature in meta data that provides dataset grouping information. Default is "Barcode".

umi.count.threshold

cells that exceed this UMI count threshold are counted.

difference.threshold

fold difference between largest and 3rd largest number of cells that is used to flag artifact genes for omission.

group.specific.is.artefact

include genes that are only expressed in one dataset as artifacts. Default is FALSE.

verbose

Print progress (T or F). Default is F.

Author

Nicholas Mikolajewicz

Examples

ag.res <-  findArtifactGenes(object = so, assay = NULL, features = NULL, meta.feature = "Barcode", umi.count.threshold = 5, difference.threshold = 100, verbose = T)
#> Error in findArtifactGenes(object = so, assay = NULL, features = NULL,     meta.feature = "Barcode", umi.count.threshold = 5, difference.threshold = 100,     verbose = T): object 'so' not found