#### Data processing related
arcsinh_cofactor: 5
subsample: FALSE
subsample_n: 4000
exclude_controls: FALSE
...but only the arcsinh transformation is part of the prepare_checkpoint1() functionality.
Potentially, could add subsampling and control exclusion.
For my use case, I am ignoring controls that were only included for batch correction purposes. I had to remove them manually from my trans_exp before running the function.
Not a huge deal, but might be nice?
Additionally, if the prepare_checkpoint1() function actually handled everything / most of what checkpoint1 does accomplish, we could maintain just prepare_checkpoint1() and plug that in to cytof_pipeline.R in place of the relevant code.
Seems like it should be fairly simple to do. I could take this work on myself if you'd like. Just let me know!
In the config file there are params...
...but only the arcsinh transformation is part of the
prepare_checkpoint1()
functionality.Potentially, could add subsampling and control exclusion.
For my use case, I am ignoring controls that were only included for batch correction purposes. I had to remove them manually from my trans_exp before running the function.
Not a huge deal, but might be nice?
Additionally, if the
prepare_checkpoint1()
function actually handled everything / most of what checkpoint1 does accomplish, we could maintain justprepare_checkpoint1()
and plug that in to cytof_pipeline.R in place of the relevant code.Seems like it should be fairly simple to do. I could take this work on myself if you'd like. Just let me know!