rad21 (SCHPO, P30776) — Review of GO:1990342 "heterochromatin island" IDA annotation

Focus: existing_go_annotation_decision · hypothesis slug existing-go-1990342-undecided
Annotation under review (PomBase detail 472247): rad21 (SPCC338.17c/P30776) is_active_in heterochromatin island (GO:1990342), evidence IDA (ECO:0000314), reference PMID:22144463, extension exists_during meiotic cell cycle (GO:0098762), assigned_by PomBase 2020-11-26, low throughput.
Seed hypothesis: the annotation should receive review action UNDECIDED.

Executive Judgment

Verdict: SUPPORTED. The annotation is directly confirmed by the primary paper's own ChIP data. UNDECIDED is no longer warranted — the recommended action is ACCEPT (with one minor caveat on the temporal extension).

The central knowledge gap — whether PMID:22144463 actually contains a Rad21 localization assay, which was inaccessible from the abstract — was resolved by reading the full text (obtained via PMC6338074). The paper states verbatim:

"Factors that bind H3K9me and their associated effectors localize to heterochromatin islands: Chromatin immunoprecipitation (ChIP) detected the Clr4 complex (ClrC) subunit Raf2, Swi6/HP1, as well as … RITS (Chp1 and Ago1) and SHREC (Clr3 and Mit1) … at heterochromatin islands (fig. S2). … We also found the cohesin-loading factor Mis4, which interacts with Swi6/HP1 (10), and the cohesin sub unit Rad21 enriched at meiotic heterochromatin islands (fig. S2)."

This is an explicit ChIP-based (IDA-appropriate) demonstration that Rad21 — and its loader Mis4 — are enriched at heterochromatin islands. It fully validates the PomBase IDA annotation and explains the internal curation pattern found in Iteration 2 (both the cohesin kleisin and its loader annotated to islands from this single reference). The abstract's silence on Rad21 (the entire basis for the seed's caution) is simply because the cohesin result is a supplementary-figure finding, not an abstract-level headline.

Caveats / remaining minor issue:
- The ChIP was performed in log-growing (vegetative) cells; the islands cover meiotic-gene loci ("meiotic heterochromatin islands") but the localization is a vegetative-growth phenomenon. The rad21-specific exists_during meiotic cell cycle (GO:0098762) extension is therefore likely a minor mis-extension (Rad21 is also largely replaced by Rec8 during actual meiosis; Kitajima 2003, Yokobayashi 2003). The base CC term is fully supported; the temporal qualifier should be reviewed/removed independently.
- Confidence in the base term is now high/direct; confidence that the extension is correct is low.

Evidence Matrix

(Provenance: PMC6338074 full-text retrieval with verbatim quote; PomBase REST API annotation + term records; NCBI eutils GDS esummary for GSE33404. CSVs: evidence_matrix_rad21_GO1990342.csv, go_decision_table_rad21.csv, island_genes_from_PMID22144463.csv.)

GO Curation Implications

This is a cellular component (CC) annotation. Lead (requires curator verification):

• ACCEPT / RETAIN GO:1990342 for rad21. The IDA is directly supported by ChIP data in Fig. S2 of PMID:22144463. Change the review action from UNDECIDED → ACCEPT.
• Review the exists_during meiotic cell cycle (GO:0098762) extension (present only on the rad21 annotation). The assay is in vegetative cells and islands merely cover meiotic genes; consider removing or replacing this temporal qualifier (e.g., no temporal extension, or a coincident_with meiotic-gene locus extension as used for other island genes in the same term).
• This is not a "protein binding" case — the informative content is the CC localization term, which is now well-supported.

Mechanistic Scope

• Direct molecular event (confirmed): the cohesin kleisin Rad21 (with loader Mis4) is physically enriched by ChIP at facultative heterochromatin islands — small H3K9me/Swi6 domains over meiotic genes silenced during vegetative growth.
• Upstream vs. downstream: island formation is driven by RNA elimination (Mmi1/Red1) + Clr4/H3K9me; Swi6/HP1 then binds H3K9me and recruits downstream effectors, including cohesin (Rad21/Mis4). Rad21 is thus a recruited effector at islands, not an island-forming factor — but the CC term "is_active_in heterochromatin island" is exactly a localization statement and is correct.
• The mechanistic chain (H3K9me → Swi6 → Psc3/Mis4–Swi6 interaction → cohesin) is consistent with Nonaka 2002 (cited as ref 10 in the paper).

Conflicts and Alternatives

• Temporal-extension conflict (only remaining concern): exists_during meiotic cell cycle conflicts with (i) the vegetative-cell context of the actual ChIP and (ii) Rad21→Rec8 replacement in meiosis. The base localization is unaffected.
• No paralog confusion for the base term: the paper explicitly assays Rad21 (mitotic kleisin), not Rec8; the vegetative context is where Rad21, not Rec8, is the active kleisin.
• Context-dependent cohesin role: Folco 2019 shows cohesin can antagonize heterochromatin elsewhere, but that does not affect the localization claim at islands.

Knowledge Gaps

1. Quantitative magnitude / genome-wide extent of Rad21 island enrichment. Checked: the paper shows targeted ChIP (fig. S2); GSE33404 has no cohesin arrays. Matters only for strength, not existence, of the annotation. Resolve by: intersecting published Rad21/Psc3 ChIP-seq (Mizuguchi 2014; Folco 2019) with island coordinates.
2. Intended meaning/validity of the exists_during meiotic cell cycle extension. Checked: assay is vegetative; conflicts with Rec8 replacement. Resolve by: curator review of the extension against the figure legend and Materials.
3. Functional role of cohesin at islands (vs. mere presence). Checked: not addressed in this paper. Not required for a CC localization annotation. Resolve by: rad21-ts/mis4-ts effects on island H3K9me and meiotic-gene silencing.

Discriminating Tests

• (Already done here) Full-text/Fig. S2 inspection — decisive; confirms Rad21 at islands.
• Coordinate-overlap analysis: intersect published Rad21/Psc3 ChIP-seq peaks with the ~21 island coordinates; permutation test for enrichment (quantifies the effect genome-wide).
• Perturbation: test island Rad21 enrichment in swi6Δ/clr4Δ (predicted loss, per Nonaka 2002 logic) to confirm the recruitment mechanism.
• Rec8 vs. Rad21 discrimination only needed if the meiotic-timing extension is to be retained.

Curation Leads (require curator verification)

• Action change: UNDECIDED → ACCEPT for rad21 is_active_in heterochromatin island (GO:1990342).
• Confirming reference + exact snippet to verify: PMID:22144463, Fig. S2 / Results — "the cohesin sub unit Rad21 enriched at meiotic heterochromatin islands (fig. S2)" and "We also found the cohesin-loading factor Mis4, which interacts with Swi6/HP1 (10) … enriched at meiotic heterochromatin islands." (Text obtained from PMC6338074; curator should confirm against the published Fig. S2.)
• Supporting mechanistic reference: PMID:11780129 — Swi6/HP1 recruits cohesin to heterochromatin (paper's ref 10).
• Separate flag: review/remove the exists_during meiotic cell cycle (GO:0098762) extension on this annotation (assay is vegetative; Rad21→Rec8 in meiosis).
• Suggested question to annotator (ORCID 0000-0003-4148-4606): confirm Fig. S2 as the IDA basis and clarify the intent of the meiotic-cell-cycle extension.
• Suggested experiment (optional, for functional depth): rad21/mis4 loss-of-function effect on island H3K9me and meiotic-gene silencing.

Resolution note: Iterations 1–2 flagged this as unresolved/partially supported because the full text was believed inaccessible (Science 403; Europe PMC OA XML 404). Iteration 3 successfully retrieved the author-manuscript full text from PMC6338074, which contains the direct ChIP statement, converting the verdict to SUPPORTED. The only residual issue is the temporal extension.

Artifacts

• OpenScientist evidence matrix rad21 GO1990342
• OpenScientist final report
• OpenScientist final report
• OpenScientist go decision table rad21
• OpenScientist island genes from PMID22144463