This is a test version of Biostars. For the public version, visit https://www.biostars.org.
subsetting out cells from seurat object based on expression of 1 gene

I have a seurat object, with raw counts stored in the RNA assay at object@assays[["RNA"]]@counts

Lets say that the count matrix is simple and looks like this, where letters are genes and numbers are cells:

      [1] [2]  [3] 
[A]    2    4    3 
[B]    1    5    7 
[C]    2    3    4

I want to subset out all the cells that have expression of gene A >= 3. So that if I did that the resulting matrix would look like this:

      [2]  [3] 
[A]    4    3 
[B]    5    7 
[C]    3    4

Im having trouble getting the right syntax for this, specifically how to call only that one gene within a Subset function and still retain all the other genes in the list

seurat r
• 25,332 views • link
updated by Ondina • Ondina • written by cook.675 • cook.675

I tried subsetting out the count matrix like this:

data.frame = as.data.frame(Object[["RNA"]]@counts)

then running: Subset <- data.frame[, "A" >= 4]

and it returned all values

I also tried Subset <- data.frame[, data.frame$"A" >= 4]

• 0 views • link
written by cook.675 • cook.675

I found a solution if anyone has interest. Take the data out of the object first as a data frame (df) then:

new_df <- df[, df["A",] >= 4]
• 0 views • link
written by cook.675 • cook.675

1 answer

For the future, the correct way to do this is:

subset(x = my.seurat.object, subset = my.gene >= 4)

As stated in the Seurat cheat sheet.

• 0 views • link
written by jared.andrews07 jared.andrews07

Thanks thats much simpler and more elegant. I have a follow up if you wouldn't mind:

How would I subset a list of genes based on the same criteria? I tried:

subset(object, features = gene.list > 0)

And got an error here:

Error: Under current subsetting parameters, the default assay will be removed. Please adjust subsetting parameters or change default assay.
• 0 views • link
written by cook.675 • cook.675

sorry for the late reply, yes a vector, I created it by entering the following:

gene.list <- c("GeneX", "GeneY", "GeneZ")

It works if I type out each gene as follows:

Object.subset <- subset(Object, GeneX > 0 & GeneY > 0 & GeneZ > 0)

Anytime I try to pass a gene name or a list to features it gives me that same error, and I cant seem to pass gene.list to subset either

• 0 views • link
written by cook.675 • cook.675

Old post, but I came across it today since I'm trying to subset an object in a bunch of different ways. My "solution" isn't elegant or robust at all, but it'll do in a pinch:

### create list w/ genes of interest
genelist <- c("Gene1", "Gene2", "Gene3", "Gene4", "Gene5", "Gene6", "Gene7", "Gene8")

### init empty list
genelist_subset_query <- ""

### for loop to construct query
for(i in 1:length(genelist)){
   if(i != length(genelist)){
      genelist_subset_query <- paste(genelist_subset_query, genelist[i], "> 0 &", sep = " ") #### can change & to | (logical "or") if needed
   }
   if(i == length(genelist)){
      genelist_subset_query <- paste(genelist_subset_query, genelist[i], "> 0", sep = " ")
   }
}

### copy and paste this output into the subset command
genelist_subset_query

### subset w/ pasted output
subset_SeuratObj <- subset(SeuratObj, Gene1 > 0 & Gene2 > 0 & Gene3 > 0 & Gene4 > 0 & Gene5 > 0 & Gene6 > 0 & Gene7 > 0 & Gene8 > 0)

If you try to pass genelist_subset_query to the subset function it will fail since it is interpreted as one long character, hence the copy & paste step.

• 0 views • link
written by angelo.pelonero • angelo.pelonero

I don't know if it's normal bu when I filter my seurat object based on this way of doing it, I have the following results:

My data goes like this based on the quantiles:

`0%    25%    50%    75%   100%  : 0.0   10.0   61.5  300.0 4029.0`  So the maximum is 4029 and the minimum is 0.

I can't use counts_svep1 <- subset(x = scrna_m, my_gene > 60) because the following error appears Error in CellsByIdentities(object = object, cells = cells) : Cannot find cells provided.

But if I do this counts_svep1 <- subset(x = scrna_m, my_gene > 2)

It cuts the data like this:

`0%  25%  50%  75% 100% : 529  860 1087 1733 4029`

How does this work? What is the logic behind it?

• 0 views • link
written by Ondina • Ondina

Open another question with your full example. Piggybacking on questions like this makes it hard for others to find your question (and any answers to it).

• 0 views • link
written by jared.andrews07 jared.andrews07

Okay, I will do that :)

• 0 views • link
written by Ondina • Ondina

Log in to answer this question.

More posts like this