Subsetting Seurat object based on list of cell barcodes

Question

[bgbs]

Hi everyone,

I have a Seurat object that I want to subset based on a csv file I have of "good" cell barcodes (56,020 cells) that have already been through QC. The list of cell barcodes have the following sample prefixes 1_MI1_S3 2_MI1_S4 3_C_MI2_S1 4_D_MI2_S2 4_MI1_S5 5_MI1_S6 7_MI1_S7

I read in the csv file of good cells as follows

all_good_cells_rhemac10 <- read.csv("master_list_good_cells_rhemac10.csv", header = FALSE)

head(all_good_cells_rhemac10)
                           V1
1 1_MI1_S3_AAACCCAAGAGACAAG-1
2 1_MI1_S3_AAACCCAAGAGTATAC-1
3 1_MI1_S3_AAACCCAAGTCCCGAC-1
4 1_MI1_S3_AAACCCAAGTGAGGCT-1
5 1_MI1_S3_AAACCCACAGAAGCTG-1
6 1_MI1_S3_AAACCCACATCTATCT-1

The seurat object I made with a raw counts matrix (69,801 cells) has cell barcode row names with the same sample prefixes. The original identity of each cell corresponds to the sample prefix (side question: Is there a way to change the orig.ident column to distinguish between the 4_D_MI2_S2 and the 4_MI1_S5 sample)?

Sample name (orig.ident column in seurat object meta.data) 1_MI1_S3 (1) 2_MI1_S4 (2) 3_C_MI2_S1 (3) 4_D_MI2_S2 (4) 4_MI1_S5 (4) 5_MI1_S6 (5) 7_MI1_S7 (7)

head([e-mail])
                              orig.ident nCount_RNA nFeature_RNA
3_C_MI2_S1_AAACCCAAGCTAGATA-1          3       2133         1348
3_C_MI2_S1_AAACCCAAGCTCACTA-1          3       1443          999
3_C_MI2_S1_AAACCCACAACAAGAT-1          3       3058         1630
3_C_MI2_S1_AAACCCACACTGCGAC-1          3       1214          877
3_C_MI2_S1_AAACCCAGTCGGTAAG-1          3       1513          998
3_C_MI2_S1_AAACCCAGTTTGTGGT-1          3       2238         1397

I have tried several methods to filter the seurat object using the subset() function but I've been running into errors

I tried adding the V1 column of the seurat object meta.data before using the subset() function but that didn't work

seurat.obj_combined[["V1"]] <- all_good_cells_rhemac10[["V1"]]
seurat.obj_combined<-subset(seurat.obj_combined, subset = all_good_cells_rhemac10$V1)

Do the cell barcodes in the seurat object and the csv file need to be in the same order? Is there a parameter I need to use in the subset function in order to subset the seurat object? There doesn't seem to be a cell barcodes parameter to use, but there is when I use WhichCells.

When I use that function, it does return a list of subsetted barcodes, but not a subsetted Seurat object which is what I want.

WhichCells(seurat.obj_combined, cells = all_good_cells_rhemac10)

1    1_MI1_S3_AAACCCAAGAGACAAG-1
2    1_MI1_S3_AAACCCAAGAGTATAC-1
3    1_MI1_S3_AAACCCAAGTCCCGAC-1
4    1_MI1_S3_AAACCCAAGTGAGGCT-1
5    1_MI1_S3_AAACCCACAGAAGCTG-1
6    1_MI1_S3_AAACCCACATCTATCT-1
7    1_MI1_S3_AAACCCACATGGAGAC-1
8    1_MI1_S3_AAACCCAGTCAAGCGA-1
9    1_MI1_S3_AAACCCAGTCCTATAG-1
10   1_MI1_S3_AAACCCAGTCTGGTTA-1
[ reached 'max' / getOption("max.print") -- omitted 55020 rows ]

Answer 1

[Ming Tommy Tang]

index<- rownames(seurat.obj_combined) %in% all_good_cells_rhemac10$V1
seurat.obj_combined[index, ]

should do the trick

[bgbs]

Hi, thanks so much!

Answer 2

[fracarb8]

subset(seurat.obj_combined, cells = all_good_cells_rhemac10$V1)

You need to pass cells not subset. From the documentation:

subset    Logical expression indicating features/variables to keep
cells, j    A vector of cell names or indices to keep

[bgbs]

Thank you!