Prerelease1 start with parameters

ktrns / scrnaseq

Workflow for single-cell RNA-seq analysis using Seurat

MIT License

37 stars 15 forks source link

Prerelease1 start with parameters #92

Closed tglomb closed 2 years ago

tglomb commented 3 years ago

The script 'scrnaseq.Rmd' has been adopted to receive parameter settings from outside including two small changes in functions_util.R. An R script for using the report from the command line is also included.

tglomb commented 3 years ago

Dear @tglomb,

Thanks a lot for your PR. It is great to see that things are moving and we are getting closer to our release!

My biggest request concerns run_scrnaseq.R. It would be very nice to have it in a generic way such that we can run it from command line, like we have it in place for the Pre-Workflow. Would you be able to implement this?

Dear @ktrns,
as previously anounced I have finally included the start script to be used from the command line. I also included two configuration files in the test directory (./test_datasets/10x_SmartSeq2_pbmc_GSE132044) to test the report with the settings from the original report.
What did not work for me was to start knitting the report from the command line using relative paths. I had to provide the full paths to the start script. Maybe you could test it on your side.

ktrns commented 3 years ago

Dear @tglomb,

Is your PR complete, and ready for us to test?

Thanks Katrin

tglomb commented 3 years ago

Dear @ktrns,

yes, it's ready from my part.
Happy testing! ... and success!

Kind regards
Torsten

andpet0101 commented 3 years ago

I have downloaded your PR and tried to merge the current master into it. There were two conflicts in scrnaseq.Rmd and one in R/functions_util.R. Can you have a look at them?

The first one in scrnaseq.Rmd is probably because because we added a param$assay_raw = "RNA" which contains the name of raw data assay. Could you simply add it as well?
The second one in scrnaseq.Rmd is likely caused by the additional params section in the YAML header. Here your version should be fine.
The one in R/functions_util.R: Here the master version is probably better since the & is used as a short-cut and version. If the first part is already false, the second part will not be evaluated. With &&, both parts are always evaluated (which could cause a name error/not in list error).

tglomb commented 3 years ago

Hi @ktrns and @andpet0101,

could you please have a look on the pull request again to see if I resolved all issues or if there is anything else to do.

Best regards
Torsten