runCellQC on multiple samples

[cmtom]

Hello,

I have the pipeline working for a single sample but run into issues when I load in multiple samples. Below is my code and the error I get. I was wondering if I'm doing something wrong or if you have any suggestions?

scex4 <- importCellRanger(cellRangerDirs ='./datasets',sampleNames = c("sample1","sample2","sample3","sample4"))
unique(scex4$sample)
[1] "sample1" "sample2" "sample3" "sample4"
head(scex4)
class: SingleCellExperiment
dim: 6 5727894
metadata(1): sctk
assays(1): counts
rownames(6): ENSMUSG00000051951 ENSMUSG00000089699 ... ENSMUSG00000025900 ENSMUSG00000025902
rowData names(3): feature_ID feature_name feature_type
colnames(5727894): sample1_AAACCCAAGAAACCCA-1 sample1_AAACCCAAGAAACCCG-1 ... sample4_TTTGTTGTCTTTGCTG-1 sample4_TTTGTTGTCTTTGGAG-1
colData names(3): cell_barcode column_name sample
reducedDimNames(0):
mainExpName: NULL
altExpNames(0):
scex4 <- runCellQC(scex4, sample = c("sample1","sample2","sample3","sample4"),
algorithms = c("QCMetrics", "doubletFinder", "decontX"),
mitoRef = "mouse", mitoIDType = "ensembl",
mitoGeneLocation = "rownames", seed = 12345)
Error in .manageCellVar(inSCE, var = sample) : Invalid variable length

If I run runCellQC() as above except with sample = NULL, it uses 30GB of RAM and takes a very long time, it still has yet to finish. If I use "scDblFinder" instead then I get an error: Cholmod error 'problem too large' at file ../Core/cholmod_dense.c

[joshua-d-campbell]

Hello @cmtom, thanks for using our tool! The "sample" parameter needs to be given a vector of sample labels that is the same length as the number of cells. It denotes which sample each cell belongs to. So your code should look like this:

scex4 <- runCellQC(scex4, sample = scex4$sample, ....)

with the "..." being the rest of the parameters you set. Just let me know if you have any more questions. I will move this to a Discussion as others may have the same question.

[cmtom]

Hello,

Thank you very much for the help! I am still having some issues, I still get "Cholmod error 'problem too large' at file ../Core/cholmod_dense.c" with scDblFinder, and with doubletFinder I get the error "cannot allocate vector of size 358.0 Gb". But if I load just one sample it works.

The full error for scDblFinder:
Error: BiocParallel errors
1 remote errors, element index: 1
3 unevaluated and other errors
first remote error:
Error in value[3L]: An error occured while processing sample 'sample1':
Error in h(simpleError(msg, call)): error in evaluating the argument 'x' in selecting a method for function 'colSums': Cholmod error 'problem too large' at file ../Core/cholmod_dense.c, line 102

Thanks again!