Folder structure is inconsistent

frisen-lab / TREX

Simultaneous lineage TRacking and EXpression profiling of single cells using RNA-seq

MIT License

5 stars 6 forks source link

Folder structure is inconsistent #4

Closed Leonievb closed 2 years ago

Leonievb commented 2 years ago

The folder structure of this git repository is inconsistent and confusing. For example,

[x] #12
[x] the folder expected can be moved to /tests/expected
[x] the file chrTomato-N.fa in /tests/data/outs/ should be moved to /tests/data/
[ ] folder /tests/data/outs/filtered_gene_bc_matrices should be renamed to /tests/data/outs/filtered_feature_bc_matrices
[ ] The files barcodes.tsv, genes.tsv and matrix.mtx in /tests/data/outs/filtered_gene_bc_matrices)/hg38_Tomato-N/ are expected to be compressed (e.g. barcodes.tsv.gz by the algorithm and algorithm gives error if not

marcelm commented 2 years ago

Regarding compressed barcodes.tsv.gz: We need to add a test for this. What exactly gives an error?

marcelm commented 2 years ago

We support two CellRanger directory layouts, see https://github.com/frisen-lab/TREX/blob/main/src/trex/cellranger.py.

The CellRanger 2 layout:

uses outs/filtered_gene_bc_matrices/
uses barcodes.tsv

The CellRanger 3 layout:

uses outs/filtered_feature_bc_matrix/
uses barcodes.tsv.gz

We detect the format automatically depending on whether a filtered_gene_bc_matrices or filtered_feature_bc_matrix subfolder exists in the outs/ folder.

You wrote filtered_feature_bc_matrices ("feature" and plural "matrices"), which is a combination of the two. Is that a typo or is the directory really named that way?

Leonievb commented 2 years ago

To your first question: I haven't seen that error myself but I got told by someone using the code that this happened only when trying to run the pipeline on the test data with the -f flag. If required, I can request the person to reproduce the error and send the error message.

Leonievb commented 2 years ago

To the second question: This was a typo from my side, I am sorry for confusion! The layouts of CelllRanger 2 and 3 are exactly as you write it. And the person I talked to said she had to rename folders to the CellRanger 3 layout as soon as she used extra arguments. I will ask her to send more details. This is of interest for issue #10 and will be posted there

Leonievb commented 2 years ago

The person reporting issues on folder name and compression of files could not reproduce the error. Therefore I would keep the folder names and files as they are and also close this issue for now