Test Examples

[amadeovezz]

No description provided.

[erflynn]

this works now! I think we can go ahead and merge soon. but we may want to add more tests first?

[erflynn]

@dtm2451 and @syellamilli adding you both as reviewers to this PR.

Major changes to outputs are:

• no dttm info on seurat objects, vcf, or gzip
• rename barcodes_of_interest.filt.list to be ${library}_barcodes_of_interest.filt.list

Will this break anything for you all?

Otherwise this adds a working (!) example test that checks md5s for all generated non-PDF files. We need to expand to more tests, but very excited that it finally works.

I think we should wait to merge until we add a few more tests though.

[erflynn]

My plan for the major workflow tests are demuxlet & freemuxlet (merged + un-merged which the test data checks) for each pipeline script (pre_qc, pre_fmx_qc, post_qc, post_fmx_qc). I want these same tests for CITE + VDJ, but I don't have the VDJ test data. I'll try to set up for CITE soon.

I've been running into a few challenges/blockers with this. I'm listing them below with their status.

• (unsatisfactory workaround) I could not get functions to work (I left some of my code stubs for this aa9b3d7) I spent awhile on this and eventually just gave up, so we have a lot of duplicated code across the tests. I think probably not worth following up on this given we will eventually switch to snakemake, but something to note.
• (unsatisfactory workaround) I cannot figure out how to put tests for multiple scripts into one test file. I tried several things here and gave up, but this means that we will have to run multiple files every time we test.
• (unsatisfactory workaround) The input data for post-qc tests have been copied to additional directories within integration_test_user. This causes data duplication of the toy data but nextflow does not like symlinks.

[erflynn]

@dtm2451 @syellamilli:
@amadeovezz and I identified an error today in the test snapshots (they don't have md5s), pausing on this PR until this is addressed

other minor update -- I do have two sets of toy data for TCR/BCR/CITE, one that can be used for running cellranger and one for post cellranger. Let me know if you want to play with these.