Conversation
| - name: Filter rank genes groups | ||
| description: Arguments for scanpy `filter_rank_genes_groups()` | ||
| arguments: | ||
| - name: --filter_results |
There was a problem hiding this comment.
If we wish to align with other components, I think we used do_subset for this.
There was a problem hiding this comment.
I'm happy to change the name but I think this is doing something different. It doesn't remove/mask vars before the test, instead it runs a second function filter_rank_genes_groups() which applies some filters to the testing results. No vars are removed at any stage.
There is a masking argument to rank_genes_groups() but it's not implemented in the component.
| logger.info("Using .X matrix") | ||
|
|
||
| logger.info("Using '%s' method", par["method"]) | ||
| sc.tl.rank_genes_groups( |
There was a problem hiding this comment.
Does this method from scanpy automatically log-normalize the input layer when it does not have the correct items set in uns? Otherwise we might need to do something different here.
There was a problem hiding this comment.
The function doesn't do anything to the input but it will give a warning if it thinks the data is not log-normalised (i.e. if it is all non-negative integers).
| - name: Outputs | ||
| description: Arguments that define the output | ||
| arguments: | ||
| - name: --output |
There was a problem hiding this comment.
Which slots are added to the output anndata? If there are none, lets remove this argument. If there are, please allow setting the slot name(s) using arguments.
There was a problem hiding this comment.
The results are stored in .uns. There is already a key_added argument that matches the function argument which controls the name. Happy to rename it if you like.
…genes * origin/main: Improve defaults for annotation workflows (#1155) Update file examples (#1067) Bump viashpy to 0.10.0 (#1178) Add `cellmapper` outputs (#1177) Fix failing filtering test after implementing obs intersection (#1175) CI - linting: pin R version to 4.5.3 (#1181) Fix broken scvelo test by excluding version 0.3.4 (#1180) Leiden: avoid making unnecessary copies of the output data and add extra arguments (#1132) Add `cellmapper` component (#1169) Bump anndata to 0.12.11 (#1174) Add optional .obs intersection (#1173) Add consensus_vote component (#1151) Add `clear_slots` component (#1171)
2 participants
Changelog
Add a
calculate_marker_genescomponent that runsscanpy.tl.rank_genes_groupsfollowed by (optionally)scannpy.tl.filter_rank_genes_groups(). Outputs a H5MU with calculated markers as well as a CSV with marker genes.Checklist before requesting a review
I have performed a self-review of my code
Conforms to the Contributor's guide
Check the correct box. Does this PR contain:
Proposed changes are described in the CHANGELOG.md
CI tests succeed!