Feat/issue 177 179 panel id and target type

[aditigopalan]

Summary

• Add HTAN_PANEL_ID as a required field to MultiplexMicroscopyLevel2 and ChannelMetadata schemas, enabling panel-level grouping and join keys across RecordSets (fixes Add HTAN_PANEL_ID as required field to MultiplexMicroscopy Level 2 and ChannelMetadata schemas #177)
• Add TARGET_TYPE enum and conditional field requirements to SpatialPanel to support non-human probes (e.g. microbiome, viral targets) that lack Ensembl/HGNC identifiers (fixes Add TARGET_TYPE field and conditional requirements for non-human probes #179, closes Support non-human (e.g. microbial) probes in panel metadata — GENE_ID / gene symbol validation too strict #175)
• Make PHYSICAL_SIZE_Z optional in MultiplexMicroscopyLevel2, conditionally required only when SIZE_Z > 1, to accommodate 2D acquisitions like CODEX (fixes Make PHYSICAL_SIZE_Z optional in MultiplexMicroscopy Level 2 #174)

Changes

modules/MultiplexMicroscopy/domains/level_2.yaml

• Added HTAN_PANEL_ID (required, with HTAN P-prefix pattern) before CHANNEL_METADATA_ID
• Changed PHYSICAL_SIZE_Z from required: true to required: false; added a rules block making it required when SIZE_Z >= 2

modules/MultiplexMicroscopy/domains/multiplex_microscopy_channel_metadata.yaml

• Added HTAN_PANEL_ID (required, with HTAN P-prefix pattern) before CHANNEL_ID

modules/SpatialOmics/domains/spatial_panel.yaml

• Added TargetTypeEnum with values Human Gene, Human Transcript, Other
• Added required TARGET_TYPE slot
• Added OTHER_TARGET_TYPE slot (free-text, conditionally required)
• Changed GENE_SYMBOL, HGNC_VERSION, GENE_ID from required: true to required: false
• Added two rules blocks: GENE_SYMBOL/HGNC_VERSION/GENE_ID required when TARGET_TYPE is Human Gene or Human Transcript; USER_GENE_NAME and OTHER_TARGET_TYPE required when TARGET_TYPE is Other

Tests

• Updated test_spatial_panel_class to reflect conditional rather than unconditional requirements
• Added test_physical_size_z_conditional asserting the rule structure for PHYSICAL_SIZE_Z

Test Plan

• poetry run pytest modules/MultiplexMicroscopy/ — 17 passed
• poetry run pytest modules/SpatialOmics/ — 11 passed

Note: Issue #175 is a duplicate of #179 and can be closed without separate changes.

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

Superseded by updated review on commit 80b479a

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

Superseded by updated review on commit 61789de

[github-actions]

HTAN Schema Review — ✅ Approved

Automated review by htan-claude · commit 61789de1a56dedf38735698e589f64f09e41c974

The PR is well-structured and the schema changes are solid — a few test coverage gaps and one meaningful biological modelling concern need attention before merge.

Caution

1 blocking issue must be resolved before merge

Files Changed

• modules/MultiplexMicroscopy/domains/level_2.yaml
• modules/MultiplexMicroscopy/domains/multiplex_microscopy_channel_metadata.yaml
• modules/SpatialOmics/domains/spatial_panel.yaml
• modules/MultiplexMicroscopy/tests/test_multiplex_microscopy.py
• modules/SpatialOmics/tests/test_spatial.py

Checklist Results

Check	Result	Notes
Inheritance correctness	PASS	No inheritance chains modified
Inlining of nested objects	N/A	No inlined objects introduced
Slot completeness (range, title, description)	PASS	All new/changed slots have range, title, and description
Enum integrity (alphabetical, descriptions)	PASS	TargetTypeEnum is alphabetically ordered; all values have descriptions
Generated artifacts	N/A	Explicitly excluded from this PR per PR description

Findings

Blocking

• spatial_panel.yaml — ENSEMBL_ID pattern accepts both ENSG and ENST prefixes with no per-TARGET_TYPE enforcement (biological)
  The slot-level pattern ^(ENSG\\d+|ENST\\d+)$ permits both gene-level (ENSG) and transcript-level (ENST) Ensembl identifiers in the same field. The Human Gene conditional rule requires ENSEMBL_ID but does not constrain it to ENSG-prefixed values, meaning a submitter could satisfy the Human Gene rule with a transcript ID and vice versa — a scientifically incorrect submission that the model would accept. The description text gives correct guidance but the model does not enforce it. The cleanest fix is to split into two slots (ENSEMBL_GENE_ID restricted to ^ENSG\\d+$ and ENSEMBL_TRANSCRIPT_ID restricted to ^ENST\\d+$) and reference each in the appropriate conditional rule; alternatively, add postcondition pattern constraints to the existing rules blocks if splitting slots is too disruptive.

Warnings

• test_multiplex_microscopy.py — No invalid-instance test for HTAN_PANEL_ID in MultiplexMicroscopyLevel2 (coverage)
  There is no test that attempts to instantiate a MultiplexMicroscopyLevel2 object without HTAN_PANEL_ID and asserts a ValueError is raised. Coverage guidelines require at least one invalid-instance path for each new required slot on a class. Add a test analogous to the invalid-instance pattern used for ChannelMetadata, omitting HTAN_PANEL_ID from a MultiplexMicroscopyLevel2 instance.

• test_multiplex_microscopy.py — No instance-level tests for the PHYSICAL_SIZE_Z conditional rule (coverage)
  test_physical_size_z_conditional inspects schema structure only; it does not exercise the rule through actual data instances. Following the pattern used in test_spatial_panel_conditional_rules_instances, at least one valid 2D instance (SIZE_Z=1, no PHYSICAL_SIZE_Z) and one invalid 3D instance (SIZE_Z=3, PHYSICAL_SIZE_Z absent) should be validated against the LinkML runtime or a validator to confirm the rule fires correctly.

• test_spatial.py — No valid-instance coverage for Bacterial, Viral, Control Probe, and Human Protein enum values (coverage)
  test_spatial_panel_conditional_rules_instances exercises Human Gene, Human Transcript, Other, and an invalid type, but the four remaining TargetTypeEnum values have no instance-level test path. A short parametrized test supplying only TARGET_NAME (and omitting identifier fields) for each of these four values would close the gap and confirm the absence of unintended required-slot violations.

• test_spatial.py — "Dropped fields" assertion checks OTHER_TARGET_TYPE (a slot that never existed) instead of guarding the correct slot (coverage)
  The test iterates over ["GENE_SYMBOL", "GENE_ID", "USER_GENE_NAME", "OTHER_TARGET_TYPE"] to assert these names are absent from SpatialPanel. Because OTHER_TARGET_TYPE was never defined in the schema, this assertion passes vacuously and provides no protection against accidentally dropping OTHER_TARGET_DESCRIPTION. The assertion should either be removed or replaced with a positive check that OTHER_TARGET_DESCRIPTION is present (which the conditional-slots loop already covers). The PR description also references OTHER_TARGET_TYPE in several places, confirming this is a naming artefact from development.

• spatial_panel.yaml — Human Protein enum value has no conditional rule and no enforced identifier (biological)
  TargetTypeEnum includes Human Protein but no rules block enforces any identifier requirements for this type, unlike Human Gene and Human Transcript which both require ENSEMBL_ID. A submitter selecting Human Protein receives no model-level constraint distinguishing it from Other, and there is no free-text descriptor required either. If protein-level identifiers are intentionally deferred, this should be explicitly documented in the slot or enum description as a known gap; otherwise, consider adding a rule requiring a protein identifier (e.g., UniProt accession) or folding Human Protein into Other until identifiers are standardised.

• spatial_panel.yaml — Viral and Bacterial target types have no required descriptor field (biological)
  OTHER_TARGET_DESCRIPTION is only conditionally required when TARGET_TYPE = Other, but the descriptions for Viral and Bacterial state that no standardised identifier is mandated. This means viral and bacterial probe targets are identified solely by TARGET_NAME with no additional free-text description enforced. Consider either extending the OTHER_TARGET_DESCRIPTION rule (or a renamed NON_HUMAN_TARGET_DESCRIPTION) to cover Viral and Bacterial, or clarifying in the model description that TARGET_NAME alone is the intended and sufficient identifier for these categories.

• spatial_panel.yaml — USER_GENE_NAME removed without a deprecation path (biological)
  USER_GENE_NAME was a previously defined slot in SpatialPanel and has been dropped entirely in this PR (replaced in function by OTHER_TARGET_DESCRIPTION). Any existing data submissions that populated USER_GENE_NAME will silently lose that field upon re-validation. The 2.0.0 version bump signals a breaking change, but a changelog entry or migration note in the schema-level description would reduce re-ingestion risk. Consider adding a comment or schema annotation mapping the old slot name to its replacement.

Informational

• spatial_panel.yaml — PR description references OTHER_TARGET_TYPE; schema consistently uses OTHER_TARGET_DESCRIPTION (structural + coverage)
  The PR description's "Changes" section states "USER_GENE_NAME and OTHER_TARGET_TYPE required when TARGET_TYPE is Other", but the actual slot name throughout the YAML and (corrected) tests is OTHER_TARGET_DESCRIPTION. The YAML itself is internally consistent — this is a prose artefact in the PR description only. No schema fix needed, but updating the PR description improves traceability.

• level_2.yaml / multiplex_microscopy_channel_metadata.yaml / spatial_panel.yaml — Version bumps are proportionate (coverage)
  spatial_panel.yaml bumps to 2.0.0 (appropriate given removal of previously required slots), level_2.yaml to 1.1.0, and multiplex_microscopy_channel_metadata.yaml to 1.5.0. All increments are consistent with the scope of their respective changes.

• spatial_panel.yaml — SpatialPanel has no HTAN_BIOSPECIMEN_ID provenance link (biological)
  SpatialPanel represents a reagent/probe panel definition rather than a biospecimen-derived record, so the absence of HTAN_BIOSPECIMEN_ID is scientifically defensible — panels are analogous to reagent manifests and are linked to biospecimen provenance indirectly via HTAN_PANEL_ID in image-level records. No action required given the current join-key design, but worth noting if SpatialPanel records are ever ingested as standalone RecordSets.

Verdict

REQUEST_CHANGES

---

_{Rules defined in CLAUDE.md · To update review rules, edit CLAUDE.md and open a PR to main}

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

Superseded by updated review on commit 7a5bb77

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

[adamjtaylor]

A few things to discuss, but nearly there I think

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

Superseded by updated review on commit 62ec5a6

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

[aditigopalan]

Fixed:
TARGET_TYPE missing terminal period ✅
OTHER_TARGET_DESCRIPTION example removed viral gene reference ✅

Intentionally not fixed:

• No conditional rules for Bacterial/Viral/Control Probe/Human Protein: deliberate, only TARGET_NAME required for those
• TARGET_TYPE name kept as-is (not renamed to SPATIAL_TARGET_TYPE etc.)
• CHANNEL_METADATA_ID deprecation notice not needed
• HTAN_PANEL_ID inlining concern — the structural issue is fine (no class has range: SpatialPanel without inlined), but the blocking item asked us to document in the HTAN_PANEL_ID description that it serves as the cross-RecordSet join key; we haven't added that text, which can be skipped here imo

Superseded by updated review on commit 1d0abd4

[github-actions]

HTAN Schema Review — ✅ Approved

Automated review by htan-claude · commit 1d0abd4756a56ef167de1772ba6a33442322dcb0

Solid PR overall — the conditional logic for TARGET_TYPE and PHYSICAL_SIZE_Z is well-structured, but there's one blocking provenance issue and several warnings worth addressing before merge.

Caution

1 blocking issue must be resolved before merge

Files Changed

• modules/MultiplexMicroscopy/domains/level_2.yaml
• modules/MultiplexMicroscopy/domains/multiplex_microscopy_channel_metadata.yaml
• modules/SpatialOmics/domains/spatial_panel.yaml
• modules/SpatialOmics/domains/level_3.yaml
• modules/MultiplexMicroscopy/tests/test_multiplex_microscopy.py
• modules/SpatialOmics/tests/test_spatial.py

Checklist Results

Check	Result	Notes
Inheritance correctness	PASS	No inheritance changes introduced
Inlining of nested objects	PASS	All new slots are range: string; no inlining required
Slot completeness (range, title, description)	PASS	All new/changed slots have range, title, and description
Enum integrity (alphabetical, descriptions)	PASS	TargetTypeEnum is alphabetical, all 7 values have descriptions
Generated artifacts	N/A	Generated .py and .json files excluded from diff per PR note

---

Findings

Blocking

• modules/MultiplexMicroscopy/domains/level_2.yaml — CHANNEL_METADATA_ID removed with no 1:1 replacement link to ChannelMetadata RecordSet (biological)

  The old CHANNEL_METADATA_ID (Synapse ID, ^syn\d+$ pattern) was the explicit foreign key pointing a single MultiplexMicroscopyLevel2 row at its exact ChannelMetadata RecordSet. Its replacement, HTAN_PANEL_ID, is a panel-level grouping key shared across many image rows and many channel rows — the join it creates is one-to-many, not one-to-one. After this change there is no unambiguous way to resolve which specific ChannelMetadata RecordSet belongs to a given image file record, breaking the provenance chain. Either retain CHANNEL_METADATA_ID (or an equivalent unique RecordSet pointer) alongside HTAN_PANEL_ID, or introduce a new slot such as CHANNEL_METADATA_RECORDSET_ID that uniquely identifies the ChannelMetadata RecordSet for each image.

Warnings

• modules/SpatialOmics/domains/spatial_panel.yaml — Human Protein target type has no conditional rule for identifier slots (biological)

  TargetTypeEnum includes Human Protein alongside Human Gene and Human Transcript, but only the gene/transcript values have conditional rules requiring ENSEMBL_ID and HGNC_VERSION. Human protein targets used in antibody-based spatial proteomics (e.g., CODEX, MIBI) have stable identifiers (UniProt, HGNC protein IDs) and deserve the same treatment. As written, a Human Protein record needs only TARGET_NAME, leaving it as under-annotated as a Bacterial probe. Either add a rule requiring HGNC_VERSION (or a new UNIPROT_ID slot) when TARGET_TYPE is Human Protein, or add an explicit note in the slot description documenting the intentional omission.

• modules/SpatialOmics/domains/spatial_panel.yaml — Bacterial and Viral target types have no conditional rule for taxonomic identifiers (biological)

  Bacterial and Viral are added as permissible TARGET_TYPE values, but unlike Other they do not trigger a requirement for OTHER_TARGET_DESCRIPTION or any equivalent taxonomic identifier (e.g., NCBI Taxonomy ID, GenBank accession). A bacterial or viral probe record could be submitted with only TARGET_NAME and no further annotation, which is insufficient for reproducibility in spatial metagenomics/viromics workflows. Consider extending the Other conditional rule to cover Bacterial and Viral, or adding a dedicated taxonomic identifier slot with appropriate conditionality.

• modules/SpatialOmics/domains/spatial_panel.yaml / level_2.yaml / multiplex_microscopy_channel_metadata.yaml / level_3.yaml — HTAN_PANEL_ID pattern duplicated across four files (biological + coverage)

  The regex pattern ^(?=.{1,50}$)(HTA2[0-2][0-9])_(0000|EXT[0-9]{1,18}|[0-9]{1,21})_(P[0-9]{1,20})$ is copy-pasted into all four files independently. If the pattern ever changes in one location it will silently diverge elsewhere. Consider defining HTAN_PANEL_ID once as a shared slot (e.g., in CoreFile or an appropriate base schema, consistent with how HTAN_DATA_FILE_ID is handled in core.yaml) and importing it into each module.

• modules/MultiplexMicroscopy/tests/test_multiplex_microscopy.py — test_htan_panel_id_missing_level2_raises uses a hand-rolled validator instead of the dataclass constructor (coverage)

  The test for a missing HTAN_PANEL_ID in MultiplexMicroscopyLevel2 implements a custom validate_panel_id function rather than instantiating the generated MultiplexMicroscopyLevel2 dataclass. This validates schema metadata but does not confirm that the runtime dataclass enforces the constraint — unlike the parallel test_invalid_channel_metadata_missing_htan_panel_id which correctly uses the dataclass constructor. Add a test that calls MultiplexMicroscopyLevel2(...) without HTAN_PANEL_ID and asserts a ValueError is raised.

• modules/SpatialOmics/tests/test_spatial.py — No valid-instance load test for SpatialPanel using the generated dataclass (coverage)

  test_spatial_panel_conditional_rules_instances uses a hand-rolled validator rather than loading a real SpatialPanel instance through the generated dataclass or a linkml-runtime loader. No test constructs an actual SpatialPanel object and verifies it parses cleanly end-to-end, which means enum range enforcement and pattern validation via linkml-validate are not exercised. Add at least one test that instantiates a SpatialPanel record via the generated dataclass or JSON schema validator.

Informational

• modules/SpatialOmics/domains/spatial_panel.yaml — Schema version field absent despite breaking changes (structural + coverage)

  spatial_panel.yaml has no version: field (before or after this PR), yet the changes are substantial and breaking: three required slots removed, three new slots added, one slot renamed with broadened semantics, a new enum added, and three conditional rules introduced. The level_2.yaml and channel_metadata.yaml files correctly carry version bumps; spatial_panel.yaml should follow suit (e.g., version: "2.0.0"). No action required to merge, but worth tracking.

• modules/SpatialOmics/domains/level_3.yaml — PANEL_SYNAPSE_ID rename may affect already-submitted data (biological)

  Renaming PANEL_SYNAPSE_ID to HTAN_PANEL_ID is a breaking change for any SpatialLevel3 records previously submitted with the old slot name. If any HTAN centers have live submissions using this field, a migration script or backward-compatibility alias will be needed. This is a coordination concern rather than a schema correctness issue — flagged for awareness.

• modules/SpatialOmics/domains/spatial_panel.yaml — SpatialPanel has no provenance back-link to a biospecimen or assay record (biological)

  SpatialPanel holds HTAN_PANEL_ID as its identifier but carries no HTAN_PARENT_ID or HTAN_BIOSPECIMEN_ID. The join key is held on the assay side (SpatialLevel3.HTAN_PANEL_ID), which appears intentional — panel as a lookup table. If SpatialPanel is ever submitted as a standalone RecordSet without a corresponding SpatialLevel3 row, provenance would be severed. Worth confirming this join-from-assay direction is the authoritative design.

• modules/MultiplexMicroscopy/domains/level_2.yaml / multiplex_microscopy_channel_metadata.yaml — HTAN_PANEL_ID correctly omits identifier: true (coverage)

  In SpatialPanel, HTAN_PANEL_ID carries identifier: true. In MultiplexMicroscopyLevel2 and ChannelMetadata, the slot is a foreign-key reference (no identifier: true), which is correct. Noted for completeness — no action required.

---

Verdict

REQUEST_CHANGES

---

_{Rules defined in CLAUDE.md · To update review rules, edit CLAUDE.md and open a PR to main}

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.

Superseded by updated review on commit 9633a4b

[github-actions]

HTAN Schema Review — ✅ Approved

Automated review by htan-claude · commit 9633a4ba54309f20c3850d1859fe0140a8bea686

Solid PR with clear intent — there are a handful of warnings worth addressing before merge, but no blocking issues preventing it.

Warning

4 warnings to address (no blocking issues)

Files Changed

• modules/SpatialOmics/domains/spatial_panel.yaml
• modules/SpatialOmics/domains/level_3.yaml
• modules/MultiplexMicroscopy/domains/level_2.yaml
• modules/MultiplexMicroscopy/domains/multiplex_microscopy_channel_metadata.yaml
• modules/SpatialOmics/tests/test_spatial.py
• modules/MultiplexMicroscopy/tests/test_multiplex_microscopy.py

Checklist Results

Check	Result	Notes
Inheritance correctness	PASS	No changes to inheritance chains; existing chains unaffected
Inlining of nested objects	PASS	HTAN_PANEL_ID slots in level files use range: string, not range: SpatialPanel; no inlining violation
Slot completeness (range, title, description)	PASS	All new slots carry range, title, description, and pattern where applicable
Enum integrity (alphabetical, descriptions)	PASS	TargetTypeEnum values are alphabetically ordered; all 7 values have descriptions
Generated artifacts	N/A	Generated Python and JSON files excluded from diff per scope rules

Findings

Warnings

• modules/SpatialOmics/domains/spatial_panel.yaml — Human Protein has no conditional identifier rule (structural + biological + coverage)
  TargetTypeEnum includes Human Protein (relevant for antibody-based spatial proteomics such as CODEX/PhenoCycler), but no rules block is triggered when TARGET_TYPE is Human Protein. This means only the unconditionally required TARGET_NAME string is enforced, with no stable identifier (e.g., HGNC symbol, UniProt accession, or Ensembl gene ID) required. If this is intentional — e.g., because TARGET_NAME is expected to follow a controlled vocabulary for proteins — please add a comment or annotations entry to the enum value or the schema documenting that decision, so future maintainers don't interpret it as an accidental omission. If it is not intentional, add a conditional rule requiring at least one identifier slot for Human Protein.

• modules/SpatialOmics/domains/spatial_panel.yaml — SpatialPanel has no biospecimen provenance anchor (biological)
  SpatialPanel (identified by HTAN_PANEL_ID) contains no HTAN_BIOSPECIMEN_ID, HTAN_PARENT_ID, or equivalent centre-level provenance slot. Without a provenance anchor it is impossible to trace which atlas centre or biospecimen collection a panel record belongs to across RecordSets, which is inconsistent with the HTAN cross-RecordSet traceability pattern. Consider whether at minimum an HTAN_CENTER_ID or HTAN_PARENT_BIOSPECIMEN_ID slot is appropriate here, or explicitly document in the schema why SpatialPanel is intentionally provenance-free (e.g., because panels are centre-level shared resources referenced by experiment-scoped records).

• modules/SpatialOmics/tests/test_spatial.py — No generated-dataclass instance test for SpatialPanel (coverage)
  All SpatialPanel tests use SchemaView structural assertions or hand-rolled validators; none construct a SpatialPanel instance via the generated dataclass (e.g., from htan_spatialomics.datamodel.spatial_panel import SpatialPanel) to confirm a valid instance loads without error and an invalid one (e.g., missing TARGET_TYPE) raises the expected error. Per the coverage guidelines, at least one valid and one invalid generated-class instance test is required for each changed class.

• modules/MultiplexMicroscopy/tests/test_multiplex_microscopy.py — No generated-dataclass valid-instance test for MultiplexMicroscopyLevel2 with HTAN_PANEL_ID (coverage)
  test_htan_panel_id_missing_level2_raises hand-rolls its own validation rather than exercising the generated MultiplexMicroscopyLevel2 dataclass. The ChannelMetadata tests correctly use the generated class, but MultiplexMicroscopyLevel2 does not have a corresponding valid-instance test confirming the new required HTAN_PANEL_ID slot is accepted end-to-end. Add at least one test that constructs a valid MultiplexMicroscopyLevel2 instance with HTAN_PANEL_ID present using the generated class.

Informational

• modules/SpatialOmics/domains/spatial_panel.yaml — No organism/database identifier slots for Bacterial, Viral, or Control Probe target types (biological)
  For Bacterial and Viral target types, only TARGET_NAME (free text) is required — there are no slots for NCBI Taxonomy IDs, GenBank accessions, or pathogen database references. This may be intentional for an initial implementation supporting microbiome/viral spatial panels. No action required now, but worth tracking as a future enhancement if stricter provenance is needed for non-human targets.

• modules/SpatialOmics/domains/spatial_panel.yaml — Schema version not bumped despite breaking changes (structural + coverage)
  The changes to spatial_panel.yaml are substantial: three fields removed (GENE_SYMBOL, GENE_ID, USER_GENE_NAME), multiple fields added, and a new enum introduced. If the schema header carries a version: key, this level of restructuring warrants an increment to signal a breaking change to downstream consumers. No action required if versioning is managed at the module or repository level rather than per-file.

• PR description contains minor inaccuracies relative to the diff (structural)
  The PR description references slot names (OTHER_TARGET_TYPE, USER_GENE_NAME, GENE_SYMBOL, GENE_ID) and a limited enum set (Human Gene, Human Transcript, Other) that do not match the actual schema changes. The YAML itself is internally consistent — this is purely a documentation issue. No schema changes needed, but updating the PR description would help reviewers and historians of the change.

Verdict

APPROVE

---

<sub>Rules defined in CLAUDE.md · To update review rules, edit CLAUDE.md and open a PR to main</sub>

[aditigopalan]

Disregarding previous claude review, plan to keep this recordset going for all of HTAN Phase 2, similar to clinical and biospecimen.

[github-actions]

✅ Python classes auto-updated!

The Python classes have been automatically regenerated and committed to this PR branch.

**Updated files:**
```
modules/Biospecimen/src/htan_biospecimen/datamodel/biospecimen.py

modules/Clinical/src/htan_clinical/datamodel/clinical.py
modules/DigitalPathology/src/htan_digitalpathology/datamodel/digital_pathology.py
modules/Imaging/src/htan_imaging/datamodel/imaging.py
modules/MultiplexMicroscopy/src/htan_multiplexmicroscopy/datamodel/multiplex_microscopy.py
modules/Sequencing/src/htan_sequencing/datamodel/sequencing.py
modules/SpatialOmics/src/htan_spatial/datamodel/spatial.py
modules/WES/src/htan_wes/datamodel/wes.py
modules/scRNA-seq/src/htan_scrna_seq/datamodel/scrna_seq.py
```

The generated classes are now up to date with the schema changes.