Skip to content

fix: drop MAML2-NOTCH entries from the consensus resource - #247

Merged
dbdimitrov merged 1 commit into
scverse:mainfrom
TemRevil:fix/remove-maml2-consensus-pairs
Aug 20, 2026
Merged

dbdimitrov merged 1 commit into
scverse:mainfrom
TemRevil:fix/remove-maml2-consensus-pairs

Conversation

@TemRevil

Copy link
Copy Markdown
Contributor

Fixes #207

MAML2 is a nuclear transcriptional co-activator, not a secreted or membrane-bound ligand. The interaction NOTCH signaling actually involves is between the cleaved intracellular domain of the NOTCH receptor and MAML2 inside the nucleus, regulating transcription, not a cell-surface ligand-receptor pair. The four MAML2-NOTCH1/2/3/4 rows in omni_resource.csv tagged consensus look like a curation artifact from the same process discussed in #177.

Removed the four rows, updated the test_select_resource count (4624 to 4620), and added test_consensus_resource_excludes_maml2 as a regression check.

Tested tests/test_select_resource.py and tests/test_orthology.py in a fresh Python 3.12 container: 9/9 passing.

MAML2 is a nuclear transcriptional co-activator, not a secreted or
membrane ligand. The MAML2-NOTCH1/2/3/4 rows in omni_resource.csv were
a curation artifact: the real interaction is between the cleaved
intracellular domain of NOTCH and MAML2 inside the nucleus, not a
ligand-receptor pair at the cell surface. Removes the four rows and
updates the consensus resource count test, plus adds a regression test
so MAML2 can't creep back into a future curation pass.

Fixes scverse#207
@TemRevil

Copy link
Copy Markdown
Contributor Author

Gentle nudge on this one whenever you have a moment. It only removes the four MAML2-NOTCH rows Rob flagged and adds a test to keep them out, so it should be a safe drop-in until the full resource revamp lands. Happy to rebase if it has gone stale.

@dbdimitrov dbdimitrov mentioned this pull request Aug 19, 2026
@dbdimitrov
dbdimitrov merged commit 2fb2b22 into scverse:main Aug 20, 2026
@TemRevil

Copy link
Copy Markdown
Contributor Author

Thanks for merging both, @dbdimitrov. If the resource revamp turns up more rows like the MAML2-NOTCH ones, glad to take a pass at those.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Potential database curation issue: nuclear protein MAML2 handled as intercellular interaction partner?

2 participants