Get the top markers for fast annotation — Find_Annotation

This function is a wrapper around FindMarkers that allows for parallelization and filtering of mitochondrial, ribosomal and non-coding RNA features in human, as well as filtering of pseudogenes in mouse. It will also directly return the top X markers for each identity.

Usage

Find_Annotation_Markers(
  seurat_object,
  ident.1 = NULL,
  ident.2 = NULL,
  min.pct = 0.25,
  top.markers = 5,
  unique.markers = TRUE,
  filter.mito = TRUE,
  filter.ribo = TRUE,
  filter.ncRNA = TRUE,
  species = "human",
  parallelized = FALSE,
  BPPARAM = NULL,
  name.features = FALSE,
  output.df = FALSE,
  output.list = FALSE,
  verbose = TRUE,
  ...
)

Arguments

seurat_object: A Seurat object.
ident.1: Character. (from FindMarkers documentation) Identity class to define markers for; pass an object of class phylo or 'clustertree' to find markers for a node in a cluster tree; passing 'clustertree' requires BuildClusterTree to have been run. Leave NULL to find markers for all clusters.
ident.2: Character. (from FindMarkers documentation) A second identity class for comparison; if NULL, use all other cells for comparison; if an object of class phylo or 'clustertree' is passed to ident.1, must pass a node to find markers for.
min.pct: Numeric. (from FindMarkers documentation) Only test features that are detected in a minimum fraction of min.pct cells in either of the two populations. Meant to speed up the function by not testing features that are very infrequently expressed.
top.markers: Numeric. The number of top markers to return. If set to Inf, all markers will be returned.
unique.markers: Logical. If TRUE, unique markers will be returned for each identity in order to prevent features repeated multiple times.
filter.mito: Logical. If TRUE, mitochondrial features will be filtered out.
filter.ribo: Logical. If TRUE, ribosomal features will be filtered out.
filter.ncRNA: Logical. If TRUE, non-coding RNA features will be filtered out.
species: Character. The species from which to pull data from to filter out features. If 'human', non-coding RNA features will be filtered out from a dataset named ncRNA_human built from genenames database. If 'mouse', only pseudogenes will be filtered out based on a dataset named pseudogenes_mouse and built from dreamBase2 database. These datasets are loaded with RightOmicsTools and may be checked for more information.
parallelized: Logical. If TRUE, FindMarkers will be parallelized using BiocParallel. Please note that parallelization is complex and depends on your system operating system (Windows users might not see a gain or might even experience a slowdown).
BPPARAM: A BiocParallelParam object to be used for parallelization. If NULL, the function will set this parameter to SerialParam, which uses a single worker (core) and is therefore not parallelized, in order to prevent accidental use of large computation resources. Ignored if parallelized = FALSE.
name.features: Logical. If TRUE, and if output.df = FALSE, each feature will be named with the corresponding cluster identity.
output.df: Logical. If TRUE, a data frame of features names and associated statistics will be returned. If FALSE, a character vector of features names will be returned.
output.list: Logical. If TRUE, a list of data frames for each identity with features names and statistics or a list of character vectors containing features names if output.df = FALSE will be returned.
verbose: Logical. If FALSE, does not print progress messages and output, but warnings and errors will still be printed.
...: Additional arguments to be passed to FindMarkers, such as test.use, or passed to other methods and to specific DE methods.

Value

A data frame or a list of data frames with features names and associated statistics, or a character vector or a list of character vectors with features names.

Examples

# Prepare data
pbmc3k <- Right_Data("pbmc3k")

# Example 1: default parameters and origin of markers
pbmc3k.markers <- Find_Annotation_Markers(pbmc3k,
                                          name.features = TRUE)
#> Finding markers for cluster Naive CD4 T against all other clusters
#> Finding markers for cluster CD14+ Mono against all other clusters
#> Finding markers for cluster Memory CD4 T against all other clusters
#> Finding markers for cluster B against all other clusters
#> Finding markers for cluster CD8 T against all other clusters
#> Finding markers for cluster FCGR3A+ Mono against all other clusters
#> Finding markers for cluster NK against all other clusters
#> Finding markers for cluster DC against all other clusters
#> Finding markers for cluster Platelets against all other clusters
pbmc3k.markers
#>  Naive CD4 T  Naive CD4 T  Naive CD4 T  Naive CD4 T  Naive CD4 T   CD14+ Mono 
#>       "CCR7"       "LEF1"        "MAL"    "PIK3IP1"       "TCF7"      "FOLR3" 
#>   CD14+ Mono   CD14+ Mono   CD14+ Mono   CD14+ Mono Memory CD4 T Memory CD4 T 
#>    "S100A12"     "S100A8"     "S100A9"       "CD14"     "CD40LG"       "AQP3" 
#> Memory CD4 T Memory CD4 T Memory CD4 T            B            B            B 
#>      "SUSD3"        "CD2"      "TRAT1"     "VPREB3"      "CD79A"      "FCRLA" 
#>            B            B        CD8 T        CD8 T        CD8 T        CD8 T 
#>      "TCL1A"      "FCER2"       "GZMK"       "GZMH"       "CD8A"       "CCL5" 
#>        CD8 T FCGR3A+ Mono FCGR3A+ Mono FCGR3A+ Mono FCGR3A+ Mono FCGR3A+ Mono 
#>      "KLRG1"        "CKB"     "CDKN1C"     "MS4A4A"       "HES4"      "BATF3" 
#>           NK           NK           NK           NK           NK           DC 
#>     "AKR1C3"       "GZMB"     "SH2D1B"      "SPON2"     "FGFBP2"   "SERPINF1" 
#>           DC           DC           DC           DC    Platelets    Platelets 
#>     "FCER1A"      "CLIC2"    "CLEC10A"       "ENHO"     "LY6G6F"      "CLDN5" 
#>    Platelets    Platelets    Platelets 
#>        "GP9"     "ITGA2B"      "SEPT5" 

# Example 2: parallelized FindAllMarkers
BPPARAM <- BiocParallel::registered()[[1]]
if (BPPARAM$workers > 4) BPPARAM$workers <- 4
pbmc3k.markers <- Find_Annotation_Markers(pbmc3k,
                                          min.pct = 0.01,
                                          top.markers = Inf,
                                          unique.markers = FALSE,
                                          filter.mito = FALSE,
                                          filter.ribo = FALSE,
                                          filter.ncRNA = FALSE,
                                          parallelized = TRUE,
                                          BPPARAM = BPPARAM,
                                          output.df = TRUE)
#> Finding markers for cluster FCGR3A+ Mono against all other clusters
#> Finding markers for cluster NK against all other clusters
#> Finding markers for cluster DC against all other clusters
#> Finding markers for cluster Platelets against all other clusters
#> Finding markers for cluster Naive CD4 T against all other clusters
#> Finding markers for cluster CD14+ Mono against all other clusters
#> Finding markers for cluster Memory CD4 T against all other clusters
#> Finding markers for cluster B against all other clusters
#> Finding markers for cluster CD8 T against all other clusters
head(pbmc3k.markers)
#>                p_val avg_log2FC pct.1 pct.2    p_val_adj     cluster feature
#> GTSCR1  1.720280e-08   7.163733 0.016 0.000 2.359193e-04 Naive CD4 T  GTSCR1
#> REG4    4.484775e-10   5.902153 0.022 0.000 6.150421e-06 Naive CD4 T    REG4
#> C2orf40 1.024936e-10   5.644114 0.023 0.000 1.405597e-06 Naive CD4 T C2orf40
#> MMP28   1.729283e-07   4.380462 0.016 0.000 2.371539e-03 Naive CD4 T   MMP28
#> NOG     5.025979e-10   4.148065 0.032 0.003 6.892627e-06 Naive CD4 T     NOG
#> FAM153A 8.848417e-08   3.880304 0.020 0.001 1.213472e-03 Naive CD4 T FAM153A