I have been using Cellenics to evaluate public scRNAseq datasets. I’ve noticed a lot of public data are in the .tsv and .mtx format (yay), but instead of having individual files for each sample they have multiplexed the data. I saw your tutorial on how to demultiplex.rds files, but didn’t see anything on how to demultiplex files already .tsv and .mtx. I can load this onto Cellenics, but can’t analyze the individual samples.
Hi! The problem with demultiplexing is that it’s almost an art: heavily dependent on how the sample information is encoded for a specific dataset.
The tutorial for rds demultiplexing can be extended to mtx/tsv files; the requirements are that you
- load these files in R (you can use utils::read.delim for TSV files, or Matrix::readMM for mtx files)
- convert them to a (sparse) matrix data structure (.mtx files are sparse matrices, so no conversion needed).
- have a way of getting the sample information. (In the tutorial we presented how to do so when the sample information is encoded in the barcode name.)
We also offer these kinds of conversion and demultiplexing for a fee as a consulting service, don’t hesitate to write to us at email@example.com for more information.
Thanks so much, this is very helpful.