matrix files gets changed during 'Reading feature-barcode matrix'

[athulmenon]

Hi Team,

Thanks for this amazing tool.

Describe the bug
I was trying to run scFlow on my dataset (Successfully ran test dataset!). But once it generates the 'sample_sce' folder the 'matrix.mtx.gz' file gets changed, and have only headers and a single line. Then scFlow throws error during 'Reading feature-barcode matrix' step. The error is as follows:

── Reading SingleCellExperiment ──
── Reading feature-barcode matrix ──
Reading: 'BL_sce/barcodes.tsv.gz'
Reading: 'BL_sce/features.tsv.gz'
Error in read.table(file = file, header = header, sep = sep, quote = quote, :
no lines available in input
Calls: lapply ... FUN -> read_sparse_matrix -> read.delim -> read.table
Execution halted

I cross checked the matrix files which I have provided, and they looks fine.

The command for runnning nfcore is as follows:
/home/athul/apps/nextflow run nf-core/scflow --manifest manifest.tsv --input Samplesheet.tsv -profile docker --outdir test_dir -r dev --qc_factor_vars group --cta_facet_vars manifest --max_cpus 20 --max_memory 40.GB --ensembl_mappings ensembl_mappings.tsv

Thanks in Advance.
Best,
Athul

[gl9797]

Hi, athulmenon!

ScFlow only accepts CellRanger's output files, if the upstream matrix file format of other analysis methods is likely to be different from CellRander to run this pipe, you can check the output document of CellRanger. IN ADDITION, I specified --CTD and --RIDDIM.YML, ALTHOUGH THIS IS NOT REQUIRED IN THE PARAMETER LIST, CTD MAY BE NECESSARY.
Hope this could be helpful to you~