Annotating single cell types using the SingleR module in BETSY

betsy_run.py --num_cores 40 \
--input SignalFile --input_file counts.txt \
--mattr keep_genes_expressed_in_perc_samples=5 \
--dattr SignalFile.preprocess=counts \
--output SingleRResults --output_file out-singler_annot.besty.txt \
--mattr singler_reference=human_primary_cell_atlas \
--run

singler_reference: "human_primary_cell_atlas", "blueprint_encode", "database_immune_cell_expression", "novershtern_hematopoietic", "monaco_immune", "immgen", “mouse_rnase"

Annotating immune single cell types using the ImmClassifier module in BETSY

betsy_run.py --num_cores 40
--input SignalFile --input_file counts.txt
--dattr SignalFile.preprocess=counts
--output SingleRResults --output_file out-ImmClass_annot.betsy.txt \
--run

Annotating single cell types using the SingleR package

❗❗❗ This is the legacy version, please use betsy instead ❗❗❗

writeups:

The input "counts_matrix.txt" file looks like so:

	Cell1	Cell2	Cell3
a	1	0	3
b	1	1	20
c	1	0	0
d	0	5	4

For more details please refer to the 📖manual :))

---

Scripts:

set.seed(41)

# install packages
# library(devtools)
# devtools::install_github('dviraran/SingleR')

# load pkgs
library(dplyr)
library(SingleR)

# run SingleR 
singler <-   
  CreateSinglerSeuratObject(  
  'counts_matrix.txt', annot = NULL, "ProjectName",  
  min.genes = 10, technology = "10X", species = "Human", citation = "",  
  ref.list = list(), normalize.gene.length = F, variable.genes = "de",  
  fine.tune = T, reduce.file.size = T, do.signatures = T, min.cells = 2,  
  npca = 10, regress.out = "nUMI", do.main.types = T,  
  reduce.seurat.object = T, numCores = 29  
  )  
  
# save the singler object  
save(singler, file = 'singler_object.RData')

# extract singler annotations as a dataframe
singler_annot <- singler$singler[[1]]$SingleR.single.main$labels %>% as.data.frame %>% `colnames<-`(c('singler.annot')) %>% mutate(Cell = rownames(.))