Add perform_celltyping parameter #506
Conversation
|
A couple more things to do with this PR, based on things that happened last week: Lines 29 to 30 in 6fd98de and: Lines 73 to 74 in 6fd98de We just want to skip the supplemental report if we don't do cell typing. Though maybe we want to keep it if there are submitter cell types? I'm sort of rethinking what the exact logic should be, but skipping the extended report seems reasonable enough. |
|
Sorry I totally missed those things. I'll add them in and then re-request review.
I think definitely skipping the extended report, but I actually think keeping the cell typing with just submitter annotations is a good idea. To me, that's being added in separate from the SingleR and CellAssign cell types, so I think we want to be able to show those even if we didn't do the other methods. |
|
Okay I added in some logic to make sure that the main section of the cell type report is there if submitter annotations are there, regardless of if other cell type methods have been performed or not. Then, the supplemental report is only there if the Here's an example of a report with submitter cell types, but no other cell types. |
templates/qc_report/celltypes_qc.rmd
Outdated
| ```{r, eval = has_singler | has_cellassign} | ||
| knitr::asis_output( | ||
| glue::glue(" | ||
| For additional information about cell typing, including diagnostic plots and/or heatmap comparisons among annotations (if available), please refer to the [supplementary cell type QC report](`r params$celltype_report`). | ||
| ") | ||
| ) | ||
| ``` |
There was a problem hiding this comment.
Is there a reason we are wrapping here and above rather than using the chunk options? I think knitr::asis_output() is useful if we have a mix of outputs, but kind of distracting if the entire chunk is asis
| ```{r, eval = has_singler | has_cellassign} | |
| knitr::asis_output( | |
| glue::glue(" | |
| For additional information about cell typing, including diagnostic plots and/or heatmap comparisons among annotations (if available), please refer to the [supplementary cell type QC report](`r params$celltype_report`). | |
| ") | |
| ) | |
| ``` | |
| ````{r, eval = has_singler | has_cellassign, result = 'asis'} | |
| glue::glue(" | |
| For additional information about cell typing, including diagnostic plots and/or heatmap comparisons among annotations (if available), please refer to the [supplementary cell type QC report](`r params$celltype_report`). | |
| ") |
There was a problem hiding this comment.
Honestly, the only reason I did it this way was to mirror how the other sections were. But we can change to using result='asis' throughout instead of knitr::asis_output?
Co-authored-by: Joshua Shapiro <[email protected]>
|
Just noting that I decided to switch all the chunks that were just text to use |
Closes #494
Stacked on #503
Here I'm adding in the flag to specify whether or not we want to run the cell typing part of the workflow with
--perform_celltyping. As we discussed, this is set tofalseby default. This means to annotate cell types when running the workflow, you need to use the flag.There wasn't really anything too involved here since the output of
annotate_celltypesalready matches the output ofcluster_sce.I tested this with the stub workflow both with and without
--perform_celltypingand the expected processes were skipped when the flag was absent. I'm doing a test run with an actual library just to be sure everything is good.Edit: The tests finished so flag works as expected!