feat: add illicofor rank_genes_groups#4038
Draft
ilan-gold wants to merge 8 commits intoig/exp_post_aggfrom
Draft
feat: add illicofor rank_genes_groups#4038ilan-gold wants to merge 8 commits intoig/exp_post_aggfrom
illicofor rank_genes_groups#4038ilan-gold wants to merge 8 commits intoig/exp_post_aggfrom
Conversation
❌ 8 Tests Failed:
View the top 3 failed test(s) by shortest run time
To view more test analytics, go to the Test Analytics Dashboard |
|
Hey, good catch for the OVO column ordering. I have no idea where the bug comes from, it seems to be related to group labels for PBMC that |
|
So it seems to be due to np.argsort and np.sort not sorting the weird characters (" ", "+", "/", "#", etc) of the bulk_labels the same way. I will ship a fix sometime today or this weekend. NB: I spent a bit of time looking into the PBMC dataset and it seems rather unusual: a lot of values are negative leading to non-defined logfoldchanges, on top of which there seems to be very little value diversity in the dataset. As a result, a lot of genes end up with identical ranksum, but because illico and scanpy do not compute z-score the exact same way (mathematically equivalent but programmatically different), gene ordering is impacted. The gene ordering impacts the position of NaN values in the results, but non-NaN values do match because they are very similar. I am not sure as if this dataset is a good test case. NB2: details of the programmatic differences: U = ranksum - n_tgt * (n_tgt+ 1)/2
z = U - n_ref * n_tgt / 2scanpy does: z =
ranksum - n_tgt * (n_tgt + 1 + n_ref) / 2Those are mathematically equivalent but result in differences of the order of 1.e-9 approximately, changing gene orders when ranksums are equal. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
TODOs:
See: https://github.com/scverse/scanpy/actions/runs/24088645078/job/70268566419?pr=4038
ovo"column" in scores (cc @remydubois)illico(see thefilterwarningson the new test)ovrresults are so far offexp_post_aggfeature feat: allow exponentiation post agg for log-fold-change inrank_genes_groups#4037) toillico, we will add theillicobackend (or all changes have been documented)illico#4012