SPKW S. pombe (Schizosaccharomyces pombe) Subproject

Parent project: SPKW.md

Overview

S. pombe is a well-curated model organism with extensive experimental annotations from PomBase. This makes it an ideal test case for identifying TRUE SPKW-unique annotations that are not subsumed by more specific experimental evidence.

Query Methodology

Used closure-based query to find TRUE SPKW-unique annotations:
- An SPKW annotation to term X is NOT unique if there exists a non-SPKW annotation to term Y where Y is_a X
- This excludes cases where SPKW provides a broad term but experimental evidence exists for a specific child term

Key Statistics (2026-01-31)

The 73% reduction demonstrates the importance of using closure-based queries. Most SPKW annotations to broad terms are subsumed by more specific experimental annotations to child terms.

Identified Over-Annotation Pattern: Meiotic Cell Cycle

207 genes have TRUE SPKW-unique "meiotic cell cycle" (GO:0051321) annotations.

Notable case: ATG autophagy genes

7 genes have BOTH "meiotic cell cycle" AND "autophagy" SPKW annotations:
- atg101, atg13, atg16, atg2, atg38, atg5, snx41

These genes have extensive experimental evidence for autophagy functions (IMP for macroautophagy, mitophagy, reticulophagy) but zero experimental evidence for meiotic cell cycle.

Root cause: UniProt assigns "Meiosis" keyword because autophagy is upregulated during sporulation (meiosis in yeast). This is the same logic error as human apoptosis annotations - a support process gets conflated with the main process.

GO subset status: GO:0051321 is in goslim_yeast and goslim_drosophila (NOT in gocheck_do_not_annotate). The specific meiotic phases (prophase I, metaphase II, etc.) ARE in do-not-annotate.

Full Gene Reviews Completed (2026-01-31)

All 7 ATG genes underwent complete annotation review with deep research. Results:

Key evidence supporting REMOVE decisions:

1. Gene naming misleading: All genes have "mug" synonyms (meiotically up-regulated gene) based on expression studies (PMID:16303567), but transcriptional upregulation ≠ functional involvement

2. Experimental evidence is for autophagy, not meiosis:

3. PMID:19778961: "autophagy-defective cells with amino acid auxotrophy were able to recover sporulation when an excess of required amino acids was supplied"

4. This proves autophagy provides nutrients during sporulation, not meiotic regulation

5. Consistent pattern: All 7 genes had the same over-annotation logic - upregulation during meiosis was conflated with functional involvement in meiotic cell cycle

Lessons Learned from S. pombe Analysis

1. Expression ≠ Function: Genes upregulated during a biological process are not necessarily functional regulators of that process

2. Support processes get conflated: Autophagy supports meiosis/sporulation by providing nutrients, but ATG genes are not meiotic regulators

3. Gene naming can mislead: The "mug" (meiotically up-regulated gene) nomenclature led to inappropriate keyword assignments

4. Closure filtering is essential: 73% of naive SPKW-unique annotations were redundant with more specific experimental evidence

5. Well-curated organisms still have systematic over-annotation: Even with extensive PomBase curation, SPKW-derived annotations introduced errors

Review Files Location

genes/SCHPO/
├── [atg101](../../genes/SCHPO/atg101/atg101-ai-review.html)/atg101-ai-review.yaml
├── [atg13](../../genes/SCHPO/atg13/atg13-ai-review.html)/atg13-ai-review.yaml
├── [atg16](../../genes/SCHPO/atg16/atg16-ai-review.html)/atg16-ai-review.yaml
├── [atg2](../../genes/SCHPO/atg2/atg2-ai-review.html)/atg2-ai-review.yaml
├── [atg38](../../genes/SCHPO/atg38/atg38-ai-review.html)/atg38-ai-review.yaml
├── [atg5](../../genes/SCHPO/atg5/atg5-ai-review.html)/atg5-ai-review.yaml
└── [snx41](../../genes/SCHPO/snx41/snx41-ai-review.html)/snx41-ai-review.yaml