Feature/hucira by Jenniliu12 · Pull Request #901 · scverse/pertpy

Jenniliu12 · 2026-01-11T21:33:32Z

PR Checklist

Referenced issue is linked
If you've fixed a bug or added code that should be tested, add tests!
Documentation in docs is updated

Description of changes
adding new tool.

Technical details
Technically, the most interesting part is the data itself. Most of the methods that come with this tool are wrappers that aid with the enrichment analysis (e.g., help create a ranked list from query data, or help navigate the different levels of the data).
The plotting tools are seaborn heatmap wrappers that are tailored to the output format.

Only the cytokine communication methods are specific to this tool. (get_senders and receivers()). This plotting function is specific to this tool.

Additional context
new feature "hucira". Lukas wants to check if methods are too specific.
Some of the code might actually be redundant.. I had implemented the wrappers to make simple visualizations a click away for users but I can rethink about making the tool more lightweight.

The only dependencies I added (bokeh and pyCirclize) are for visualization tools

The very generic tests that I wrote are not passing. I don't know why, it looks like they're timing out.

Zethson

Thank you very much already! Here's some first initial feedback:

Please ensure that you have a descriptive pull request title and the pull request description only has the necessary detail. This can also be done later but it's important.
Your implementation should not add any dependencies. Pertpy has to minimize dependencies because it's used by a loooot of people. Every dependency that we add has the potential to break users environment and increases complexity. Therefore, please remove bokeh, pycirclize, and tqdm. The plot should be implemented with seaborn and instead of using tqdm, you can use Rich directly. Not sure what to do about pycirclize - if necessary we can talk about it. That's one of the differences of implementing something for yourself vs implementing it for many.
Please ensure that all functions that have a bit more than trivial complexity have a proper docstring. Please adhere to our docstring style.
I didn't review everything yet because I think there's still a lot to do. Let's address these first.

Zethson · 2026-01-23T12:21:32Z

pertpy/data/_datasets.py

    return adata
+
+
+def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:


Please type this.

Suggested change

def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:

def human_cytokine_dict(exclude_well_biased_genes: bool = True) -> pd.DataFrame:

Zethson · 2026-01-23T12:21:49Z

pertpy/data/_datasets.py

+
+
+def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:
+    r"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.


Suggested change

r"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.

"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.

this doesn't need to be a raw string, right?

Zethson · 2026-01-23T12:21:59Z

pertpy/data/_datasets.py

+def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:
+    r"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.
+
+    The Human Cytokine Dictionary was created from single-cell RNA-seq of 9,697,974 human peripheral blood mononuclear cells (PBMC) from 12 donors stimulated in vitro with 87 different cytokines. The object is a dataframe representing cytokine activity as differentially expressed genes after cytokine perturbation.


Please always write sentences in their own line.

Zethson · 2026-01-23T12:22:11Z

pertpy/data/_datasets.py

+
+    Returns:
+        Pandas DataFrame
+


Suggested change

Nit

Zethson · 2026-01-23T12:22:18Z

pertpy/data/_datasets.py

+    The Human Cytokine Dictionary was created from single-cell RNA-seq of 9,697,974 human peripheral blood mononuclear cells (PBMC) from 12 donors stimulated in vitro with 87 different cytokines. The object is a dataframe representing cytokine activity as differentially expressed genes after cytokine perturbation.
+
+    References:
+        Oesinghaus, Lukas and Becker, S{\"o}ren and Vornholz, Larsen


Zethson · 2026-01-23T12:36:27Z

pertpy/tools/_hucira.py

+        2. Creates ranking of query data genes contrasting condition1 vs condition2. A continuum from genes most associated with condition1 (top) to genes most associated with condition2 (bottom)
+        3. Computes enrichment of each cytokine by matching their associated gene set in the ranked list.
+
+        Parameters


Zethson · 2026-01-23T12:36:58Z

pertpy/tools/_hucira.py

+        verbose: bool = False,
+        threads: int = 6,
+    ) -> pd.DataFrame:
+        """Function wrapper: Computes cytokine enrichment activity in one celltype using GSEA scoring. Loops through several threshold value to obtain more robust gene sets.


Suggested change

"""Function wrapper: Computes cytokine enrichment activity in one celltype using GSEA scoring. Loops through several threshold value to obtain more robust gene sets.

"""Computes cytokine enrichment activity in one celltype using GSEA scoring.

Loops through several threshold value to obtain more robust gene sets.

It's a rule that the first line is always just one sentence. No exceptions.

Zethson · 2026-01-23T12:37:06Z

pertpy/tools/_hucira.py

+        weight: float = 1.0,
+        seed: int = 2025,
+        verbose: bool = False,
+        threads: int = 6,


This is a super random value

Zethson · 2026-01-23T12:38:04Z

tests/tools/test_hucira.py

+    ranked_stats, _num_cells = hucira._compute_ranking_statistic(dummy_adata, contrast_column, contrasts_combo)
+    assert isinstance(ranked_stats, pd.DataFrame)
+
+    # with pytest.raises(KeyError):


This shouldn't be commented out, right?

Zethson · 2026-01-23T12:38:31Z

tests/tools/test_hucira.py

+        ("B cell", "B_cell"),
+        ("CD8a", "CD8_T_cell"),
+        ("Mono", "CD14_Mono"),
+    ]  # can't be a list for "run_one_enrichment_test()"


Not sure what's meant with that.

Jenniliu12 added 6 commits January 6, 2026 00:46

added hucira tool. Just run_one_enrichment

20ae692

added robustness_test.py from huCIRA package

768d0f9

Modified dict loading and added pl tools and sender receiver tl.

afafba3

added function wrapper run_all_enrichment_test()

3193fdd

removed import of gseapy

ea2fbfd

added dependencies and very generic tests

df4371c

Jenniliu12 marked this pull request as draft January 12, 2026 22:20

Zethson requested changes Jan 23, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feature/hucira#901

Feature/hucira#901
Jenniliu12 wants to merge 6 commits intoscverse:mainfrom
Jenniliu12:feature/hucira

Jenniliu12 commented Jan 11, 2026 •

edited

Loading

Uh oh!

Zethson left a comment

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Zethson Jan 23, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		return adata


		def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:



		def human_cytokine_dict(exclude_well_biased_genes=True) -> pd.DataFrame:
		r"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.

	r"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.
	"""Human Cytokine Dictionary curated from PBMC allows you to infer differential cytokine activity.

Conversation

Jenniliu12 commented Jan 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Zethson left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Jenniliu12 commented Jan 11, 2026 •

edited

Loading