FlowSOM#59
Conversation
…ing to multiple samples (after batch correction) and it would be helpful to have a utility function to read multiple FCS files, generage a sample ID column, and merge - implemented in accordance with the pytometry examples
* FlowSOM clustering moved to tools * FlowSOM tutorial added to docs * Some minor refactors to FlowSOM to allow modification of consensus clustering params * Some minor refactors to plotting
mbuttner
left a comment
There was a problem hiding this comment.
Looks pretty good overall! I commented on a few occasions that need to be changed. Also, consider to run pre-commit to get past the code checks. Thanks!
|
An issue I've noticed with the pre-commit config. Flake8 fails because I've used TypedDict in I've updated the pre-commit yaml to ignore TYP001, let me know if this is okay. |
Sounds good to me. |
|
🚀 Deployed on https://6549fe01f1e20777540b1dcb--pytometry.netlify.app |
mbuttner
left a comment
There was a problem hiding this comment.
One minor thing - please execute the 03_flowsom notebook to display the plots (see netlify page).
I've introduced FlowSOM under the
tools.clusteringmodule. The FlowSOM implementation is as a a functionflowsom_clusteringand is similar to scanpy functions. Dependency onconsensusclusteringa package I've put together to implement consensus cluster in Python. Other things to note:flowsom_clusteringreturns an annotated dataframe with the labels introduced under a new key (which the user can specify)flowsom_clusteringcan optionally return the SOM and ConsensusClustering objects for inspection (see tutorial for an example)Axesobject to be passed - fixed minor bugs and missing docs