[yarikoptic-gitmate]

Follow-up to #1949.

Problem

For C. elegans, nwbinspector's check_subject_sex accepts only XX (hermaphrodite) or XO (male) as Subject.sex. That check is CRITICAL under the "dandi" config, so dandi validate and dandi upload report it as an error. extract_sex, however, raised ValueError: Cannot interpret sex field: xx (or xo). No C. elegans NWB file could satisfy both:

• XX/XO passes the inspector, but asset metadata extraction fails;
• O extracts as "Other", but fails the CRITICAL inspector check.

Existing C. elegans Dandisets (e.g. 000472, 000541, 000776, 000981) record their subjects as "Other".

Change

extract_sex maps XX/XO (any case) when species designates C. elegans:

sex	identifier	name
XX	PATO_0001340	Hermaphrodite
XO	PATO_0000384	Male

C. elegans is recognized from "Caenorhabditis elegans" and "C. elegans" (the two values nwbinspector checks), "Caenorhabditis elegans - …", and the NCBITaxon IRI NCBITaxon_6239.

It also reads c_elegans_sex from the subject. The CElegansSubject type of ndx-subjects fixes sex to O and records XX/XO there, as in Dandiset 001075. When sex is O or unset, that value is used. An explicit M/F in sex still wins.

For any other or missing species, XX/XO still raise. In mammals, for example, XX would mean female, so the codes are not unambiguous outside C. elegans.

Tests

• test_extract_sex_c_elegans covers the four species spellings and both codes, in upper and lower case.
• test_extract_sex_chromosomal_other_species checks that XX/XO still raise without a species, and for mouse and fly.
• test_extract_sex_ndx_subjects_c_elegans covers c_elegans_sex, and checks that an explicit sex still wins.
• All tests are marked @pytest.mark.ai_generated.
• The sex/species tests pass (66), and pre-commit passes.

Related: NeurodataWithoutBorders/nwbinspector#353 (added the C. elegans rule), NeurodataWithoutBorders/nwb-schema#575, dandi/dandi-schema#171.

---

Generated by Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.12%. Comparing base (4e9bb5c) to head (2ab90c0).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1976      +/-   ##
==========================================
+ Coverage   91.04%   91.12%   +0.08%     
==========================================
  Files          95       95              
  Lines       14599    14645      +46     
==========================================
+ Hits        13291    13345      +54     
+ Misses       1308     1300       -8     

Flag	Coverage Δ
unittests	91.12% <100.00%> (+0.08%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[CodyCBakerPhD]

There is another place to look; this is why for something like the ndx-subjects extension, the approach was to set the global field to O and have an extended field subject.c_elegans_sex="XX" (example: https://neurosift.app/nwb?url=https%3A%2F%2Fapi.dandiarchive.org%2Fapi%2Fassets%2F66226765-86e1-40be-8d82-8ff3bf69ad37%2Fdownload%2F&dandisetId=001075&dandisetVersion=0.240930.1859&tab=%2Fgeneral%2Fsubject)

[yarikoptic-gitmate]

Thanks, I had missed that. 2ab90c0 also reads c_elegans_sex. It is now extracted with the other subject fields, and when sex is O or unset, XX/XO from c_elegans_sex maps to Hermaphrodite/Male. An explicit M/F in sex still wins. So files like those in 001075 would now get PATO_0001340 instead of plain "Other".

The two conventions conflict with each other, though. For species "Caenorhabditis elegans", nwbinspector's check_subject_sex (CRITICAL in its dandi config) only accepts XX/XO in sex. A file following ndx-subjects, with sex fixed to O, therefore fails dandi validate/upload with current nwbinspector. 001075 was published in 2024, before nwbinspector made that check CRITICAL (nwbinspector#648). Should nwbinspector accept O when c_elegans_sex is set?

---

Generated by Claude Code

[CodyCBakerPhD]

Should nwbinspector accept O when c_elegans_sex is set?

Yep, go ahead and fork that repo and make that PR

[CodyCBakerPhD]

Did you confirm that the extension field will show up in that metadata dictionary object passed to that level?

[yarikoptic-gitmate]

Not before you asked, but I have now, end to end. I wrote NWB files with ndx-subjects 0.2.0's CElegansSubject (species "Caenorhabditis elegans"). Then I uninstalled ndx-subjects, so pynwb had to use the namespace cached in the file, and ran get_metadata and nwb2asset from this branch:

file	get_metadata	wasAttributedTo[0].sex
sex="O", c_elegans_sex="XO"	sex='O', c_elegans_sex='XO'	Male, PATO_0000384
c_elegans_sex="XX", sex left unset (the "fixed" value is not written by pynwb)	sex=None, c_elegans_sex='XX'	Hermaphrodite, PATO_0001340

So the extension field reaches extract_sex through get_metadata, which getattrs every name in metadata_nwb_subject_fields on nwb.subject.

I could add this as a test. The test would need ndx-subjects, or a hand-built extension namespace, in the test environment. I left it out because that adds a test dependency, but I'm happy to add it, guarded by pytest.importorskip("ndx_subjects"), if you think it's worth it.

---

Generated by Claude Code

[CodyCBakerPhD]

that's fine thanks for checking

[yarikoptic]

Wow, thank you @CodyCBakerPhD for guiding my corporate friend to conclusion here. I just came to review its work and was pleased to see it "fully done"! ;)

[CodyCBakerPhD]

Wow, thank you @CodyCBakerPhD for guiding my corporate friend to conclusion here. I just came to review its work and was pleased to see it "fully done"! ;)

Unfortunately it did not pass my smoke test to

Yep, go ahead and fork that repo and make that PR (#1976 (comment))

So I guess 'someone' should do that eventually, if they ever want to use that extension

[yarikoptic]

sorry, do you mean this logic which is part of the PR https://github.com/dandi/dandi-cli/pull/1976/changes#diff-f1fffc52a57ec0f6b12d707f48d19c51f4dd81a4606d3535935d38e10e710bbdR317 or there is more todo?

[CodyCBakerPhD]

With respect to

A file following ndx-subjects, with sex fixed to O, therefore fails dandi validate/upload with current nwbinspector. 001075 was published in 2024, before nwbinspector made that check CRITICAL (nwbinspector#648). Should nwbinspector accept O when c_elegans_sex is set?

[github-actions]

🚀 PR was released in 0.82.0 🚀

[yarikoptic]

ah, gotcha, it didn't obey to your order!