Apoptosis Project

Bottom line: scoped and underway as a review campaign. This project will
separate mammalian apoptosis into mechanistic modules, audit annotations parked
at the broad GO:0006915 apoptotic process term, and look for over-annotation
from caspase, TUNEL, annexin V, viability, or generic cell-death readouts. Two
draft modules already exist for intrinsic apoptotic
signaling and death receptor
apoptotic signaling; both
should be reviewed before adding execution-phase and BCL2/MOMP submodules. The
first review slice should stay mammal-focused, with human and mouse rows handled
together, while using conserved opisthokont components as comparators rather
than blindly transferring every mammalian branch to flies, worms, or fungi.

Overview

Apoptosis is a caspase-centered programmed cell-death program in metazoans. In
mammals, the best-defined inputs are:

• Extrinsic/death-receptor signaling, where FAS, TRAIL, or TNF-family
  receptor complexes recruit FADD and initiator caspases such as CASP8.
• Intrinsic/mitochondrial signaling, where BCL2-family decisions at the
  outer mitochondrial membrane lead to cytochrome c release, apoptosome
  assembly by APAF1, and CASP9 activation.
• Execution, where effector caspases such as CASP3, CASP6, and CASP7 cleave
  substrates that dismantle nuclear, cytoskeletal, and repair systems.
• Restraint and checkpointing, where anti-apoptotic BCL2 proteins, IAPs,
  c-FLIP/CFLAR, and survival kinases decide whether a stress or receptor signal
  is amplified into cell death.

Those arms should not all collapse to the same GO term. The top-level
GO:0006915 term is often correct for a core executioner but too imprecise for
mechanistic curation; positive regulation of apoptotic process and negative regulation of apoptotic process are often correct for direct regulators, but
are noisy when a paper only measured final cell death after a broad
perturbation.

Immediate Goals

1. Module boundaries
2. Review the existing draft intrinsic_apoptotic_signaling and
   death_receptor_apoptotic_signaling modules.
3. Split reusable stages where useful: receptor/DISC assembly, BCL2-family
   MOMP control, apoptosome assembly, initiator-caspase activation, and the
   effector-caspase execution phase.

4. Keep direct apoptotic execution distinct from upstream survival signaling,
   generic stress responses, and regulation of neighboring death programs
   such as necroptosis.

5. Top-level apoptosis annotations

6. Find existing annotations to GO:0006915 apoptotic process and ask whether
   the source supports a more specific stage such as intrinsic signaling,
   extrinsic death-receptor signaling, apoptosome assembly, caspase activation,
   or execution phase.
7. Use the exact human and mouse GO:0006915 snapshot in
   GO0006915 as the first broad-row triage list.
8. Prioritize generic mammalian rows and caspase-family IBA/IEA propagation.

9. Defer trivial downcasts such as neuron apoptotic process unless the row
   also hides a mechanistic problem.

10. Over-annotation from assays

11. Treat caspase activity, TUNEL, annexin V/PI, mitochondrial depolarization,
    and viability assays as high-convergence phenotypic readouts unless the
    paper connects the gene product to a proximal apoptotic step.
12. Use the ASSAY_TO_FUNCTION readout rubric for
    evidence from cell-death assays.
13. Track APOPTOSIS-specific cases in
    apoptosis readout over-annotation.
14. Add recurring failure modes to OVER_ANNOTATION_PATTERNS
    when a mouse or mammalian genetics paper licenses only a non-core survival
    phenotype but GOA asserts generic apoptosis.

Scope

Primary: human and mouse, with a zebrafish exact-term audit as a first
non-mammalian vertebrate extension. Review human and mouse orthologs together
for the first mammalian slice so MGI phenotypes, GOA propagation, and human
biochemistry can be compared rather than curated in isolation; use the ZFIN
slice to find DANRE rows where embryonic cell-death assays have been lifted to
generic apoptosis, and use the InterPro slice to find broad domain-to-GO
mappings that should be challenged upstream.

Comparators: Drosophila, Caenorhabditis, budding/fission yeasts, and other
opisthokonts. Use these to understand conserved caspase/metacaspase,
BCL2-like, apoptosome-like, nuclease, and AIF/EndoG biology. Treat absence or
rewiring as a biological finding: metazoan intrinsic apoptosis is not simply
the mammalian CASP9/cytochrome-c pathway in every opisthokont, and fungal
programmed cell death should not be promoted to apoptotic process without
evidence that the GO definition is satisfied.

Draft Modules

Module	Status	Next check
Intrinsic apoptotic signaling	AUDITED DRAFT	Mammal-scope the CYCS/APAF1/CASP9 apoptosome and link out to the BCL2/MOMP and execution companion modules.
Death receptor apoptotic signaling	AUDITED DRAFT	Add TRAIL-R1/TRAIL-R2, TNFR1, CFLAR, and CASP10 variants around the FAS/FADD/CASP8 seed.
Execution phase of apoptosis	DRAFT	Expand beyond the CASP3/CASP6/CASP7, XIAP, DIABLO, DFFA-DFFB/CAD, ACIN1, and AIFM1 seed to lamins, PARP branches, and apoptotic surface exposure only after substrate reviews.
BCL2-family MOMP control	DRAFT	Expand the BAX/BAK, BCL2/BCL-xL/MCL1, and BID/BIM/PUMA/NOXA seed to the remaining mammalian BCL2-family paralogs.
Conserved non-mammalian apoptosis	TODO	Compare worm CED-3/CED-4/CED-9/EGL-1, Drosophila Ark/caspase/IAP wiring, and fungal metacaspase/AIF/EndoG systems before proposing any cross-opisthokont model.

Candidate Review Set

Priority 1: Mammalian Core Machinery

Species	Genes	Why first
human	CASP3, CASP6, CASP7, CASP8, CASP9, APAF1, CYCS	Core initiator and effector caspases plus apoptosome components; these define whether GO:0006915 should be replaced by signaling- or execution-phase descendants.
human	BAX, BAK1, BCL2, BCL2L1, MCL1, BID, BCL2L11, BBC3, PMAIP1	Direct BCL2-family control of MOMP; central to distinguishing intrinsic signaling from generic positive/negative regulation.
human	FAS, FASLG, FADD, TNFRSF1A, TNFSF10, TNFRSF10A, TNFRSF10B, CFLAR, CASP10	Death-receptor/DISC branch and the CASP8 checkpoint that toggles apoptosis, survival, and necroptosis.
human	XIAP, DIABLO, DFFA, DFFB, ENDOG, AIFM1, ACIN1	IAP control and downstream dismantling of the cell after effector-caspase activation.

Priority 2: Mouse Over-Annotation Slice

Start with reviewed mouse genes that already have dense apoptosis or survival
phenotype rows, then expand by querying mouse GOA for GO:0006915 and
regulation descendants.

Mouse gene	Review question
Bcl2	Use the existing human and mouse BCL2 pair as the worked example for separating direct anti-apoptotic mitochondrial activity from a large tissue-development phenotyping tail.
Casp3	Check whether generic apoptotic process and tissue-specific descendants can be tightened to execution-phase activity or kept as non-core developmental outcomes.
Tnfrsf1a	Distinguish TNF death-receptor apoptosis from NF-kappaB survival and necroptosis checkpoint biology.
Akt1, Mapk1	Test survival-kinase rows for the common pattern where altered final cell death was curated as apoptosis regulation.

Priority 3: Conserved Opisthokont Comparators

• C. elegans: ced-3, ced-4, ced-9, egl-1, and CEP-1 for the
  compact core apoptosis circuit.
• Drosophila: initiator/effector caspases, Ark/Apaf-1, DIAP/IAP regulators,
  and Reaper/Hid/Grim-like IAP antagonists.
• Fungi: metacaspases, AIF/EndoG-like nucleases, and ER-stress-sensor death
  assertions where mammalian ER-stress apoptosis may have been propagated too
  far.

Curation Heuristics

• Tighten broad rows only with proximal evidence. A CASP3 row to
  GO:0006915 can often move to execution phase; an upstream kinase row usually
  needs evidence for a named BCL2, death-receptor, or apoptosome step before it
  should move deeper than generic regulation.
• Separate direct regulators from outcome reporters. Knockout survival,
  fewer TUNEL-positive cells, or less cleaved caspase-3 can justify a cautious
  process row, often non-core, but does not by itself name the substrate or
  complex the gene product acts on.
• Do not conflate apoptosis with cell death. A viability defect, necrosis
  assay, or mitochondrial depolarization readout needs orthogonal evidence
  before it can support apoptosis rather than another regulated or accidental
  death path.
• Check taxon-specific wiring before accepting IBA. Caspase-family and
  cytochrome c annotations are exactly where a PAINT node can be biologically
  meaningful in one animal clade and misleading in another branch.
• Treat regulation terms as first-class when the mechanism is restraint.
  BCL2, XIAP, and CFLAR can be direct apoptosis regulators even though their
  function is to prevent execution; the problem is unsupported distance, not
  regulation per se.

Project Status

• [x] Audit the two existing draft apoptosis modules
• [x] Add missing execution/BCL2/MOMP module pieces
• [x] Query GOA for GO:0006915 rows in human and mouse
• [x] Review the Priority 1 mammalian core genes
• [x] Review a first mouse over-annotation slice
• [x] Compare worm, fly, and fungal conserved components
• [x] Query GOA for GO:0006915 rows in zebrafish and triage the ZFIN-authored exact rows

---

STATUS

• [x] Project scope created
• [x] Mammal-first worklist prioritized
• [x] Existing draft modules audited
• [x] BCL2-family MOMP and execution-phase draft modules created and linked from the intrinsic seed
• [x] First gene review started: human CASP3 generic protein-binding rows re-reviewed; no validation warnings remain
• [x] CASP6 reviewed manually from cached primary literature, Reactome, UniProt, GOA, and GO-CAM evidence
• [x] CASP7 reviewed manually from cached primary literature, Reactome, UniProt, GOA, and GO-CAM evidence
• [x] CASP8 reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM evidence, and annotation-reviewer follow-up
• [x] CASP9 reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM evidence, and annotation-reviewer follow-up
• [x] APAF1 reviewed manually from cached primary literature, Reactome, UniProt, GOA, and annotation-reviewer follow-up
• [x] CYCS reviewed manually from cached primary literature, Reactome, UniProt, GOA, and annotation-reviewer follow-up
• [x] BAX reviewed manually from cached primary literature, Reactome, UniProt, GOA, and annotation-reviewer follow-up
• [x] BAK1 reviewed manually from cached primary literature, Reactome, UniProt, GO-CAM evidence, and annotation-reviewer follow-up
• [x] Existing human BCL2 review revalidated for the BCL2-family MOMP slice
• [x] Existing human BCL2L1 review revalidated for the BCL2-family MOMP slice
• [x] MCL1 reviewed manually from cached primary literature, UniProt, PAINT/PANTHER evidence, and annotation-reviewer follow-up
• [x] BID reviewed manually from cached primary literature, Reactome, UniProt, GO-CAM, PAINT/PANTHER evidence, and annotation-reviewer follow-up
• [x] BCL2L11 reviewed manually from cached primary literature, Reactome, UniProt, PAINT/PANTHER evidence, and annotation-reviewer follow-up
• [x] BBC3 reviewed manually from cached primary literature, Reactome, UniProt, GOA, and annotation-reviewer follow-up
• [x] PMAIP1 reviewed manually from cached primary literature, Reactome, UniProt, PAINT/PANTHER evidence, and annotation-reviewer follow-up
• [x] Existing human FAS review revalidated and generic partner-specific protein-binding imports resolved with annotation-reviewer follow-up
• [x] FASLG reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM, PANTHER evidence, and annotation-reviewer follow-up
• [x] FADD reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM, PANTHER evidence, and annotation-reviewer follow-up
• [x] Existing human TNFRSF1A review refreshed after GOA update; 12 new protein-binding imports resolved with annotation-reviewer follow-up
• [x] TNFSF10 reviewed manually from cached primary literature, Reactome, UniProt, GO-CAM, and annotation-reviewer follow-up
• [x] TNFRSF10A reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM, PANTHER evidence, and annotation-reviewer follow-up
• [x] TNFRSF10B reviewed manually from cached primary literature, Reactome, UniProt, GOA, GO-CAM, PANTHER evidence, and annotation-reviewer follow-up
• [x] CFLAR reviewed manually from cached primary literature, Reactome, UniProt, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] CASP10 reviewed manually from cached primary literature, Reactome, UniProt, GO-CAM, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] XIAP reviewed manually from cached primary literature, UniProt, GOA, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] DIABLO reviewed manually from cached primary literature, Reactome, UniProt, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] DFFA reviewed manually from cached primary literature, Reactome, UniProt, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] DFFB reviewed manually from cached primary literature, Reactome, UniProt, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] ENDOG reviewed manually from cached primary literature, UniProt, PANTHER/PAINT evidence, and annotation-reviewer follow-up
• [x] Existing human AIFM1 review revalidated after a GOA/PANTHER refresh and annotation-reviewer follow-up
• [x] Existing human ACIN1 review refreshed after GOA update and annotation-reviewer follow-up
• [x] Existing mouse Bcl2 review refreshed after GOA update and annotation-reviewer follow-up
• [x] Existing mouse Casp3 review refreshed after GOA update and annotation-reviewer follow-up
• [x] Existing mouse Tnfrsf1a review refreshed after GOA update and annotation-reviewer follow-up
• [x] Existing mouse Akt1 review refreshed after GOA update, manual cached-publication research, GO-CAM comparison, and annotation-reviewer follow-up
• [x] Existing mouse Mapk1 review refreshed after GOA update, manual cached-publication research, GO-CAM comparison, and annotation-reviewer follow-up
• [x] Mouse Casp9 reviewed manually as the intrinsic initiator-caspase extension of the GO:0006915 audit
• [x] Mouse Casp8 reviewed manually as the death-receptor initiator-caspase extension of the GO:0006915 audit
• [x] Worm ced-3 reviewed as the first conserved CED pathway comparator
• [x] Worm ced-4 reviewed as the conserved CED apoptosome adaptor comparator
• [x] Worm ced-9 reviewed as the conserved CED BCL2-family apoptosis inhibitor comparator
• [x] Worm egl-1 reviewed as the conserved CED BH3-only apoptosis activator comparator
• [x] Worm CEP-1 reviewed as the conserved p53-family DNA-damage apoptosis comparator
• [x] Zebrafish exact GO:0006915 rows fetched from QuickGO and ZFIN-authored rows triaged for assay-readout over-annotation
• [x] Drosophila Dronc reviewed as the conserved Dark-apoptosome initiator-caspase comparator
• [x] Drosophila Dark reviewed as the conserved Apaf-1/CED-4 apoptosome adaptor comparator
• [x] Drosophila DrICE reviewed as the conserved effector-caspase execution comparator
• [x] Drosophila Dcp-1 reviewed as the companion conserved effector-caspase execution comparator
• [x] Drosophila Diap1 reviewed as the conserved IAP caspase-inhibition, ubiquitin-ligase, and NEDD8-ligase comparator
• [x] Drosophila Diap2 reviewed as the conserved IAP/Imd-pathway ubiquitin-ligase comparator
• [x] Drosophila rpr reviewed as the conserved Reaper/RHG IAP-antagonist comparator
• [x] Drosophila hid reviewed as the conserved Hid/RHG IAP-antagonist comparator
• [x] Drosophila grim reviewed as the conserved Grim/RHG IAP-antagonist comparator
• [x] Drosophila skl reviewed as the conserved Sickle/RHG IAP-antagonist comparator
• [x] Budding yeast MCA1 reviewed as the fungal metacaspase programmed-cell-death and protein-quality-control comparator
• [x] Budding yeast NUC1 reviewed as the fungal EndoG-like mitochondrial nuclease, apoptotic-DNA-fragmentation, and meiotic viral-defense comparator
• [x] Budding yeast IRE1 reviewed as the fungal ER-stress UPR comparator and stale ER-stress-apoptosis IBA audit
• [x] Hypocrea/Trichoderma IRE1 reviewed as the filamentous-fungal ER-stress UPR comparator and stale TRAF2/ASK1 death-arm TreeGrafter audit
• [x] Exact human and mouse GO:0006915 QuickGO rows fetched, normalized, and summarized

NOTES

2026-10-01

• Extended the mouse exact-GO:0006915 audit to the two top-ranked initiator
  caspases left after the first over-annotation slice. Casp9 was centered on
  APAF1-apoptosome recruitment and effector-procaspase maturation: inherited
  broad apoptosis rows were tightened to intrinsic apoptotic signaling, while
  stimulus-, stress-, tissue-, and c-Abl-binding transfers that could not be
  checked from cached abstracts were left non-core, over-annotated, or
  unresolved.
• Completed the new mouse Casp8 review with annotation-reviewer follow-up.
  Generic GO:0006915 apoptotic process imports were scoped to
  GO:0008625 extrinsic apoptotic signaling pathway via death domain receptors;
  the direct PIDD row was marked over-annotated because the cited study does not
  detect procaspase-8 processing in PIDD-expressing MEFs; execution-phase rows
  were replaced by GO:0051604 protein maturation; and the anti-necroptotic
  RIPK1/CYLD/gasdermin/N4BP1 cleavage branches were kept as direct but
  apoptosis-adjacent functions.

2026-09-30

• Queried QuickGO for exact GO:0006915 rows in human and mouse with
  descendant usage disabled. The snapshot found 825 human rows across 345
  displayed symbols and 557 mouse rows across 238 displayed symbols; the
  normalized TSVs, raw JSON, and symbol rollups are under
  APOPTOSIS/go0006915, with a concise summary in
  APOPTOSIS/GO0006915.md.
• Added companion draft modules for BCL2-family MOMP control and the mammalian
  apoptotic execution phase. The BCL2/MOMP module expands the BAX/BCL2
  placeholder into BAX/BAK effectors, BCL2/BCL2L1/MCL1 guardians, and
  BID/BCL2L11/BBC3/PMAIP1 BH3-only activator/sensitizer leaves; the execution
  module separates CASP3/CASP6/CASP7 proteolysis, XIAP/DIABLO restraint,
  DFFA-DFFB/CAD DNA fragmentation, and ACIN1/AIFM1 nuclear remodeling from
  upstream death-receptor and apoptosome signaling.
• Audited the two existing draft apoptosis modules against the completed gene
  reviews and cached GO-CAM models. The intrinsic draft now records explicit
  BCL2/MOMP, mammalian CYCS/APAF1/CASP9, and execution-phase handoff gaps; the
  death-receptor draft now records TRAIL-R1/TRAIL-R2, TNFR1, and CFLAR/CASP10
  expansion gaps and grounds its FAS/FADD/CASP8 seed in the production
  FASL-FAS GO-CAM.
• Re-audited the human CASP6 review after a current GOA refresh and
  annotation-reviewer follow-up. The source set was unchanged at 256 GOA rows;
  the review still removes the high-throughput GO:0005515 rows and keeps the
  ZBP1/RIPK3 molecular-adaptor assertion, and the two accepted top-level
  GO:0006915 rows were tightened to GO:0097194 execution phase of apoptosis.
• Completed the budding yeast MCA1 review by centering
  Mca1 on Ca(2+)-stimulated Arg/Lys metacaspase activity, prion-domain
  recruitment to cytosolic insoluble protein aggregates, core protein quality
  control, and fungal apoptotic programmed cell death under H2O2,
  chronological-aging, and mRNA-decay stress. The broad GO:0006915 rows were
  accepted as metacaspase-dependent yeast apoptosis rather than narrowed to
  metazoan execution-phase terms.
• Completed the budding yeast NUC1 review by centering
  Nuc1 on mitochondrial inner-membrane DNA/RNA nuclease activity, fungal
  EndoG-like apoptotic DNA fragmentation, and meiotic cytosolic defense against
  L-A/Killer dsRNA mycoviruses. The 2026 mitophagy and regulation of mitochondrial DNA metabolic process rows were narrowed to
  GO:0032043 mitochondrial DNA catabolic process, because Nuc1 degrades
  mtDNA exposed by a mitophagy-vacuolar escape route rather than operating the
  Atg11/Atg32 mitophagy machinery.
• Completed the budding yeast IRE1 review by keeping the
  adaptive Ire1-Hac1 unfolded-protein-response kinase/RNase activities as core
  and marking the GO_Central GO:0070059 intrinsic apoptotic signaling pathway in response to endoplasmic reticulum stress IBA as over-annotated. The
  current PANTHER cache places the death term on a mouse ERN-supported node,
  while the yeast protein remains on the conserved fungal/eukaryotic UPR sensor node;
  broad catalytic, kinase, RNase, nucleotide-binding, metal-binding, and generic
  DCR2 protein binding imports were cleaned before marking the review
  complete.
• Completed the Hypocrea/Trichoderma IRE1 review by
  keeping the adaptive fungal ER-stress kinase/RNase UPR functions as core,
  preserving the obsolete GO:0051082 unfolded protein binding rows as
  MODIFY to GO:0002235 detection of unfolded protein, and tightening broad
  catalytic, kinase, RNase, nucleotide-binding, and metal-binding IEAs. The
  TreeGrafter ER-stress apoptosis row remains over-annotated as a mammalian
  ERN-branch import to fungal SF6, while the GO:1990604 TRAF2/ASK1 complex
  row is removed outright because Trichoderma lacks the required TRAF2 and
  ASK1/MAP3K5 partners.
• Completed the Drosophila Dronc review by separating
  core Dark-apoptosome signaling and DrICE activation from DIAP1 substrate-side binding,
  local nonapoptotic caspase proteolysis, Eiger-specific apoptosis, and atypical programmed necrosis.
• Completed the Drosophila Dark review by centering its
  cytosolic NB-ARC/CARD/WD40 apoptosome scaffold activity, narrowing broad
  apoptosis rows to Dronc-activating apoptotic signaling, keeping direct
  spermatid, chaeta, glial, gamma-radiation, and salivary histolysis outputs as
  non-core, and marking systemic necrosis/inflammation/metabolism phenotypes as
  over-annotated rather than direct Dark functions.
• Completed the Drosophila DrICE review by centering its
  cytoplasmic C14 cysteine-endopeptidase activity in the execution phase of
  apoptosis, narrowing top-level apoptosis and programmed-cell-death rows to
  execution phase, converting generic DIAP interaction rows to BIR-domain
  binding, removing substrate-side Dronc protein-binding rows, keeping
  spermatid, nurse-cell, SOP-patterning, and innate-immune cleavage outputs as
  non-core, and marking a PAINT neuron-apoptosis propagation as over-specific.
• Completed the Drosophila Dcp-1 review by centering its
  short-prodomain C14 cysteine-endopeptidase activity in apoptotic execution,
  narrowing top-level apoptosis and programmed-cell-death rows to execution
  phase, moving broad ovarian transport, starvation, and oogenesis rows toward
  nurse-cell apoptosis or macroautophagy, keeping SesB/mitochondrial
  localization, Toll/NF-kappaB repression, and neuromuscular degeneration as
  non-core outputs, removing neuronal RNP granule rows caused by
  decapping-protein Dcp1 symbol confusion, and leaving abstract-only
  BIR-domain and neuron-remodeling rows unresolved.
• Completed the Drosophila Diap1 review by centering
  DIAP1/thread on BIR-domain caspase inhibition plus RING-dependent ubiquitin
  and NEDD8 ligase activities. The curation narrows generic Dronc, DrICE, and
  Dcp-1 binding rows to caspase binding where the partner was visible; tightens
  broad catalytic and top-level apoptosis rows; keeps RHG, Jafrac2, dOmi/HtrA2,
  border-cell, SOP, Grim, DTRAF1/JNK, and Wg/Wnt branches as non-core; and marks
  spermatid nuclear differentiation, Toll signaling, Eiger/JNK, and a
  survivin-like cell-cycle PAINT transfer as over-annotated or incorrect for
  DIAP1.
• Completed the Drosophila Diap2 review by centering
  its RING-dependent K63 ubiquitin ligase role in the peptidoglycan/Imd innate
  immune pathway. The curation keeps direct DrICE binding and inhibition as a
  secondary apoptotic-threshold branch, removes generic Hid/Reaper/Grim
  GO:0005515 interactions, narrows broad catalytic imports to ubiquitin
  protein ligase activity, downgrades organismal Gram-negative defense outputs
  to non-core, and leaves abstract-only DIAP1-degradation and Grim-ubiquitination
  rows unresolved rather than flattening them into the better-supported DREDD,
  IMD, and Kenny K63-linked ubiquitination mechanism.
• Completed the Drosophila rpr review by centering
  Reaper on DIAP1 IAP antagonism, GH3/phospholipid-dependent mitochondrial
  outer-membrane association, and UbcD1-dependent DIAP1 autoubiquitination and
  turnover. The curation narrows broad apoptosis rows to apoptotic signaling,
  keeps ecdysone, midgut, salivary-gland, neuroblast, JNK, and DNA-damage rows
  as contextual non-core branches, narrows generic Reaper-DIAP2 binding to
  ubiquitin protein ligase binding, removes the substrate-side dBruce row, and
  leaves several abstract-only Sickle, dOmi, DmIKKepsilon, and CNS-remodeling
  rows unresolved.
• Completed the Drosophila hid review by centering
  Head involution defective on mitochondrial RHG-family antagonism of DIAP1 and
  DIAP2 BIR domains. The curation narrows broad apoptosis/programmed-cell-death
  rows to apoptotic signaling, narrows visible DIAP protein-binding rows to
  ubiquitin-protein-ligase or BIR-domain binding, keeps supported tracheal,
  midgut, macroautophagy, DNA-damage, and Ack-buffered developmental contexts as
  non-core or accepted outputs, marks generic ecdysone, organ-growth, radiation,
  and larval/pupal-development rows as over-annotated, and leaves abstract-only
  sex-differentiation, circadian, autophagic-cell-death, and localization rows
  unresolved.
• Completed the Drosophila grim review by centering
  Grim on RHG-family BIR-domain binding to DIAP1/DIAP2-family IAPs plus a
  separable GH3-dependent mitochondrial route. The curation narrows top-level
  apoptosis and programmed-cell-death rows to apoptotic signaling, replaces the
  DIAP2 GO:0005515 row with BIR domain binding, broadens mitochondrial
  matrix localization to mitochondrion, broadens cytochrome-c release to
  apoptotic mitochondrial changes because Drosophila Grim redistributes
  cytochrome c rather than detectably releasing it to cytosol, keeps supported
  Malpighian and neuronal/glial developmental contexts as non-core or accepted
  outputs, and leaves abstract-only Sickle, lifespan, melanization, larval CNS
  remodeling, and localization rows unresolved.
• Completed the Drosophila skl review by centering
  Sickle on N-terminal IAP-binding-motif binding to DIAP1 and DIAP2 BIR
  domains. The curation adds a missing GO:1990525 BIR domain binding
  molecular-function row, narrows six top-level apoptosis rows to
  GO:0097190 apoptotic signaling pathway, keeps the corazonergic-neuron
  developmental PCD row as non-core, and leaves the Trends in Genetics
  DOI-only NAS row plus the abstract-only ionizing-radiation row unresolved.

• Completed the worm CEP-1 review by centering its nuclear
  p53-family transcription activator role upstream of egl-1 and ced-13.
  Generic DNA-binding and transcription rows were narrowed toward Pol
  II-specific transcription activator terms; the core DNA-damage rows to
  intrinsic apoptotic signaling pathway in response to DNA damage by p53 class mediator were accepted or narrowed there from broader parents; PRMT-5/CBP-1
  repression and IRE-1/ER-stress apoptosis were retained as secondary contexts;
  and hypoxia, starvation, lifespan, meiotic segregation, paraquat/SMG-1, old
  RAD-51 IGI, zinc-binding, and nucleolar rows were kept non-core,
  over-annotated, or unresolved according to how directly the cached evidence
  supported direct action by the protein.

• Completed the worm egl-1 review by centering EGL-1 on BH3-mediated
  inhibition of the BCL2-family survival protein CED-9. The five generic
  GO:0005515 protein binding rows to CED-9 were narrowed to a proposed
  BH3-ligand-side pro-survival BCL2-family inhibitor activity rather than the
  directionally wrong CED-9-side BH3 domain binding; broad
  positive regulation of programmed cell death rows were tightened to
  apoptosis; and the generic CED-4-complex assembly row was moved to apoptotic
  signaling. The local EGL-1/CED-9/CED-4/CED-3 synapse-pruning branch and
  EGL-1-dependent mitochondrial fragmentation were retained as non-core, while
  Salmonella defense and PIG-1 cell-extrusion endpoints were kept out of the
  core apoptotic-signaling model.

• Completed the worm ced-9 review by centering CED-9 on mitochondrial
  sequestration of CED-4 and EGL-1 BH3-domain recognition. Generic
  GO:0005515 rows to EGL-1 were narrowed to BH3 domain binding; CED-4
  binding rows were narrowed to protein sequestering activity; broad
  apoptotic process rows were moved to negative regulation of apoptotic process; and BCL2-family PAINT transfers for cytochrome c release,
  channel/transmembrane transport, and ligandless extrinsic apoptotic signaling
  were removed as mammalian or unsupported leak-throughs. Synapse-pruning and
  DRP-1/FZO-1 mitochondrial-dynamics rows were retained as real but non-core
  branches, while Salmonella defense, PIG-1 cell shedding, DCT-1/ceBNIP3
  binding, and CSP-2/CSP-3 substrate-control rows were kept out of the core
  CED-9 model.

• Completed the worm ced-4 review by centering CED-4 on ATP/Mg-bound
  apoptosome assembly and CED-3 zymogen activation rather than top-level
  apoptosis. The curation narrows generic GO:0006915,
  positive regulation of apoptotic process, and peptidase/protein-processing
  rows to apoptotic signaling or apoptotic cysteine-endopeptidase activator
  activity; replaces supported CED-3 GO:0005515 protein binding rows with
  caspase binding; removes most generic CED-9, SUN-1, MAC-1, FEM-1, and viral
  E1B interaction rows; keeps viral E1B BH-domain binding as non-core; accepts
  mitochondrial and apoptotic perinuclear localization; and keeps local
  CED-pathway synapse pruning plus CED-4-dependent cell-size control separate
  from canonical whole-cell apoptosis.

• Completed the worm ced-3 review as the first conserved non-mammalian
  comparator. The curation centers CED-3 on CED-4-apoptosome-mediated caspase
  zymogen activation and execution-phase substrate cleavage; narrows broad
  GO:0006915, programmed cell death, generic peptidase, and protein
  processing rows; removes nearly all generic GO:0005515 protein binding
  interactor rows except the early CED-3/CED-4 CARD-domain interaction; removes
  CSP-2/CSP-3/NPP-14 inhibitory interactions from activator and
  negative-regulation terms that belong on the inhibitors; keeps the GSNL-1
  synapse-pruning and LIN-14/LIN-28/DISL-2 heterochronic cleavage branches as
  direct but non-core proteolytic outputs; and marks Salmonella defense, PIG-1
  cell-shedding, manganese-response, and adult muscle-homeostasis assertions as
  over-annotated or unresolved phenotype endpoints.

• Refreshed the existing mouse Mapk1 review after a GOA update. The curation
  resolves 18 newly seeded rows: it accepts refreshed nucleoplasm, cytoplasm,
  cytosol, and ERK1/ERK2-cascade source rows; keeps regulation of ossification
  and myelination as non-core developmental outputs; removes four generic
  GO:0005515 protein binding partner rows; narrows the DUSP6/MKP-3 row to
  GO:0019902 phosphatase binding; marks transferred cholesterol-biosynthesis
  and Smoothened-signaling rows as over-annotated endpoint assertions; narrows
  three GO:0106310 protein serine kinase activity rows to
  GO:0004707 MAP kinase activity; and drops seven stale ARBA/IEA source
  assertions that were superseded by the refreshed GOA snapshot.

• Refreshed the existing mouse Akt1 review after a GOA update. The curation
  resolves 32 newly seeded rows: it accepts refreshed PAINT and orthology
  support for core serine/threonine kinase activity and cytosol/plasma-membrane
  localization; keeps osteoblast differentiation, Cntnap2/Akt-mTOR
  inflammatory/pain responses, TSC/IRS PI3K-AKT signaling, TORC1, cilium,
  eNOS, and mitochondrial-localization assertions as non-core pathway context;
  accepts UniProt EXP GO:0106310 protein serine kinase activity rows for
  GSK3B, cGAS, and MICU1 phosphorylation; removes three generic
  GO:0005515 protein binding rows; and drops three stale PAINT/IEA source
  assertions that were superseded by the refreshed GOA snapshot.

• Refreshed the existing mouse Tnfrsf1a review after a GOA update. The curation
  is now complete and validates cleanly after 18 rat/human ISO rows received
  structured propagation-review metadata. It accepts newly surfaced TNF
  receptor activity, TNF binding, TNF-mediated signaling, TNF receptor
  superfamily complex, and extracellular-region source rows; removes six new
  generic GO:0005515 protein binding partner rows and migrates legacy
  Tnfrsf1a generic-binding rows to REMOVE; narrows a broad ARBA response to lipopolysaccharide row to GO:0071222 cellular response to lipopolysaccharide; marks an ARBA cellular response to lipid row as
  over-annotated; and drops nine stale source assertions no longer present in
  the refreshed GOA snapshot.

• Refreshed the existing mouse Casp3 review after a GOA update. The curation
  is now complete after adding propagation-review metadata to 21 rat/human ISO
  transfers. It accepts refreshed cysteine-type endopeptidase,
  cytoplasm/cytosol, proteolysis, broad apoptotic-process, and execution-phase
  source rows; keeps newly surfaced human-ortholog nucleus, GSDME-linked
  pyroptotic inflammatory response, protein maturation,
  regulation-of-protein-localization, and PIDD apoptotic-signaling rows as
  non-core; modifies broad GO:0008233 peptidase activity to GO:0004197 cysteine-type endopeptidase activity; removes the new RPS18 GO:0005515 protein binding IntAct row from a substrate screen; drops nine stale source
  assertions no longer present in the refreshed GOA snapshot; and preserves
  three intentional mixed-action warnings for abstract-only evidence rows.
• Refreshed the existing mouse Bcl2 review after a GOA update. The curation now
  has no pending rows and validates cleanly, accepts newly surfaced
  PAINT/ortholog mitochondrial outer-membrane and BCL-2 complex transfers,
  removes the new Gimap3 GO:0005515 protein binding row as generic IntAct
  evidence, accepts the Fnip1-rescue apoptosis row from full PMID:22709692 text,
  leaves all seven PMID:15613488 ER-calcium/IP3R rows UNDECIDED because the
  cached source is abstract-only, narrows a broad transferred molecular-adaptor
  row to BH3-domain binding, narrows an ARBA neuron-apoptosis parent to negative
  regulation, and drops five stale source assertions no longer present in the
  refreshed GOA snapshot.
• Refreshed the existing human ACIN1 review after a GOA update. The curation
  preserves ACIN1's core nuclear RNA-binding/ASAP/EJC-associated splicing role
  and caspase-activated apoptotic chromatin-condensation role; refreshes UniProt
  and GOA source snapshots; reviews five newly imported RNPS1/PNN generic
  protein binding rows; migrates older SF3A2, PCBD1, SRPK2, PNN, and RBM5
  generic interaction rows from the legacy over-annotation action to REMOVE;
  and removes generic enzyme binding because CASP3 cleavage makes ACIN1 a
  substrate, not a separate enzyme-binding factor. The Drosophila Acinus
  autophagy signal remains a human follow-up question rather than a new human
  autophagy annotation.
• Revalidated the existing human AIFM1 review. The completed curation preserves
  77 reviewed GOA rows and separates three core functions: mitochondrial
  FAD/NADH oxidoreduction, NADH-dependent AIF anchoring and activation of
  CHCHD4/MIA40 for mitochondrial disulfide-relay import, and released nuclear
  DNA/PAR-dependent death-effector activity. It keeps broad apoptosis and
  programmed-cell-death rows where AIF does direct prodeath work, removes or
  tightens generic protein-binding and oxidoreductase rows, and intentionally
  leaves rat stimulus transfers, peroxide-specific NAD(P)H oxidase chemistry,
  respiratory-chain complex assembly placement, and the necroptosis boundary as
  curator follow-up questions.
• Completed the human ENDOG review. The curation centers ENDOG on
  Mg-dependent DNA/RNA endonuclease activity in mitochondria and the nucleus,
  accepts direct human evidence for 5hmC-modified DNA cleavage/recombination,
  mitochondrial DNA catabolism with compensatory mtDNA replication, and
  starvation autophagy, adds a conservative 14-3-3 protein binding row for the
  YWHAG-dependent mTORC1-suppression mechanism, narrows broad nuclease,
  hydrolase, metal-binding, DNA-catabolism, and TOR rows, keeps apoptotic DNA
  fragmentation as a plausible non-core ENDOG branch, removes generic
  DNAJA4/ITLN2 interactome protein binding rows, and marks rat-derived
  stimulus and neuron-localization transfers as over-annotation.
• Completed the human DFFB/CAD review. The curation centers DFFB on its
  caspase-released DNA endonuclease activity in apoptotic oligonucleosomal DNA
  fragmentation, narrows broad nuclease/hydrolase rows to DNA endonuclease activity, narrows generic DNA catabolic process and top-level
  apoptotic process rows to apoptotic DNA fragmentation, removes the
  DFFA-specific negative regulation of apoptotic DNA fragmentation assertion,
  removes generic DFFA protein binding rows, keeps DNA binding, TOP2A enzyme
  binding, DFF complex membership, homodimerization, and HPA nucleolus rows as
  non-core, and retains chromatin, nuclear/nucleoplasmic, and cytosolic
  locations needed for the DFF40:DFF45 activation model.
• Completed the human DFFA/ICAD review. The curation keeps DFFA as the
  DFFB/CAD-specific folding chaperone and deoxyribonuclease inhibitor in the DFF
  complex, explicitly avoiding any CAD nuclease assertion on DFFA itself; adds
  protein folding chaperone and protein folding as conservative new
  annotations for CAD folding; accepts negative regulation of apoptotic DNA fragmentation, deoxyribonuclease inhibitor activity, and cytosolic,
  nucleoplasmic, and chromatin localization; narrows broad top-level apoptosis
  rows to CAD-dependent apoptotic DNA-fragmentation inhibition; removes
  high-throughput or otherwise generic DFFB, HSPB1, and TSPYL4 protein binding
  rows; and keeps generic protein-containing complex only as a non-core
  placeholder pending a DFF-specific complex term.
• Completed the human DIABLO/SMAC review. The curation centers DIABLO on
  mitochondrial IAP antagonism after PARL maturation and MOMP-dependent cytosolic
  release; removes most protein binding rows from survivin, ARTS-context,
  AREL1, PKCdelta, Livin degradation, and large interactome evidence; narrows
  core XIAP, cIAP1, ML-IAP/Livin, and BIRC6/BRUCE edges to broad
  molecular function inhibitor activity until GO has a specific IAP-antagonist
  MF; tightens top-level apoptosis rows to intrinsic apoptotic signaling pathway; keeps mitochondrion, mitochondrial intermembrane space, and cytosol
  locations; removes mouse-transferred CD40-complex localization; and drops an
  unsupported survivin complex row.
• Completed the human XIAP review. The curation centers XIAP on direct BIR2/BIR3
  inhibition of apoptotic caspases and on RING-domain E3 ligase activity;
  narrows broad ubiquitin-transfer rows to ubiquitin protein ligase activity;
  keeps RIPK2/NOD, TLE/Groucho Wnt, COMMD1/CCS copper, and IRF7 antiviral
  ubiquitination branches with non-apoptotic context where appropriate; accepts
  negative regulation of apoptosis rows supported by direct caspase inhibition;
  removes high-throughput generic protein-binding rows; and removes or demotes
  rows for SMAC/DIABLO, HTRA2/3/4, ARTS, SIAH1, XAF1, TRIM32, and USP19 where
  the cited interactor antagonizes or degrades XIAP rather than representing an
  XIAP molecular function.
• Completed the human CASP10 review. The curation keeps CASP10 as a
  catalytically active tandem-DED initiator caspase recruited through FADD to
  Fas/CD95 and TRAIL death-inducing signaling complexes; narrows top-level
  apoptosis and generic DED regulation imports to the death-domain-receptor
  extrinsic pathway; retains CD95 DISC, ripoptosome, cytosolic Reactome, CARP
  E3-binding, and prodomain NF-kappaB rows with non-core context where needed;
  converts the c-FLIP(L) myeloma interaction to negative regulation of
  autophagic cell death; and removes high-throughput or generic RIOK3, XIAP,
  SERPINB9, and RIPK1 protein-binding rows.
• Completed the human CFLAR/c-FLIP review. The curation centers c-FLIP
  isoforms on tandem-death-effector-domain assembly with FADD and
  CASP8/CASP10, explicitly separates FLIP(L) pseudo-caspase heterodimers from
  DED-only FLIP(S) caps, removes PANTHER and InterPro catalytic-caspase
  propagation to the inactive pseudoenzyme, keeps DISC/CD95 DISC/ripoptosome
  component rows, proposes the missing CASP8 inhibitor and negative-necroptosis
  child terms, and marks the large rat/mouse Compara tissue and stimulus tail
  as over-annotation from downstream phenotypes.
• Completed the human TNFRSF10B/TRAIL-R2 review. The curation centers DR5 on
  GO:0036463 TRAIL receptor activity at the plasma membrane in
  TRAIL-activated apoptotic signaling pathway; narrows generic TRAIL
  ligand-binding imports, broad death-receptor and signaling ancestors, and
  direct TRAIL ligand rows to the receptor activity or TRAIL pathway; removes
  HuRI partner edges as uninformative DR5 molecular-function rows; leaves
  abstract-only sensitization and anti-apoptotic-complex imports undecided;
  marks ER-stress, mechanical-stimulus, TP53, and c-FLIP/CASP8 downstream
  assertions as over-annotations; and keeps the DR5-specific Reactome
  procaspase-8 dimerization event that was intentionally rejected for
  TNFRSF10A.
• Completed the human TNFRSF10A/TRAIL-R1 review. The curation centers DR4 on
  GO:0036463 TRAIL receptor activity at plasma-membrane rafts in
  TRAIL-activated apoptotic signaling pathway; narrows generic death
  receptor, signaling receptor, TRAIL-binding, self-oligomerization, and
  DISC co-complex rows to that activity or the TRAIL pathway; keeps
  GODZ/ZDHHC3 and ARAP1 evidence as trafficking context; marks DR5-only,
  c-FLIP/CASP8, and TP53-transcription Reactome exports as too indirect for a
  mature DR4 plasma-membrane assertion; and leaves abstract-only DR5-peptide
  and interferon AP-MS imports undecided rather than overruling the curator from
  incomplete local evidence.
• Completed the human TNFSF10/TRAIL review. The curation centered TRAIL on a
  membrane-bound or soluble TNF-family homotrimer that binds TNFRSF10A/DR4 and
  TNFRSF10B/DR5 to initiate TRAIL-activated apoptotic signaling pathway;
  narrowed generic receptor-binding, protein binding, top-level apoptosis,
  cell-communication, and signal-transduction rows to TNF receptor binding or
  the TRAIL-specific pathway; removed the downstream CUL3/caspase-8 interaction
  import; kept zinc coordination, identical subunit binding, NF-kappaB
  activation, immune response, and exosome detection as non-core contexts; and
  marked cytochrome-c release as a downstream over-annotation through the
  BID/BAX/BAK mitochondrial arm.
• Refreshed the existing human TNFRSF1A review after QuickGO added 12 new
  generic protein binding rows. New TRADD, RIPK1, RFK, and SH3RF2 rows were
  folded into tumor necrosis factor receptor activity; GRN/progranulin,
  UBB, MON2, and BioPlex-only rows were removed rather than pointed at TNF
  binding or kept as bare interaction edges; and the core-function synthesis
  was retargeted from unreviewed parent processes to the already accepted
  positive regulation of canonical NF-kappaB signal transduction and
  extrinsic apoptotic signaling pathway via death domain receptors terms.
• Completed the human FADD review. The curation centered FADD on its
  DD-to-DED adaptor role bridging activated FAS/TRAIL/TNFR death receptors to
  CASP8/CASP10-containing death-inducing signaling complexes; narrowed broad
  protein binding, protease binding, generic apoptosis, complex, and plasma
  membrane rows to caspase binding, death effector domain binding, death
  receptor binding, DISC assembly, or death-domain-receptor extrinsic signaling;
  kept ripoptosome and MAVS/RIG-I-like receptor cytokine/antiviral outputs as
  non-core; removed NleB-substrate and high-throughput interactome rows; and
  marked broad mouse lymphoid-organ, T-cell, kidney, cocaine, NF-kappaB, and
  host-defense projections as over-annotation.
• Completed the human FASLG review. The curation centered FasL on
  TNF-family cytokine/death-receptor binding activity upstream of the FAS/CD95
  DISC; narrowed generic receptor-binding, top-level apoptosis, and
  DISC-context protein binding imports to the FAS death-receptor branch;
  removed high-throughput SH3/HuRI/BioPlex/U2OS interactome rows; separated the
  ADAM10/SPPL2A-released nuclear FasL ICD, exosome FasL, endothelial apoptosis,
  and necroptotic signaling as non-core contexts; and removed Reactome FASLG
  transcription events that were being exported as extracellular FASLG protein
  localization.
• Completed a targeted human FAS cleanup after GOA refresh added 17
  partner-specific protein binding rows to the existing isoform-aware
  review. Canonical DISC-context FADD/CASP8 rows were tightened to
  FAS/TNFRSF6 receptor activity; exact-partner rows hidden in abstract-only
  papers were left UNDECIDED; and non-core GSK3B/BIRC2/PMLRARalpha or
  proximity-ligation/BioPlex generic PPI rows were removed rather than being
  allowed to stand as uninformative molecular functions.
• Completed the human PMAIP1/NOXA review. The curation centered
  NOXA on BH3-mediated inhibition of pro-survival MCL1 and BCL2A1,
  with lower-affinity contextual BCL2 and BCL2L10 binding retained
  under the same proposed NOXA-side inhibitor activity. Generic
  protein binding, top-level apoptosis, DNA-damage, permeability,
  cytochrome-c release, and intrinsic-apoptosis rows were tightened to
  MCL1/MOMP biology where the source was proximal; BioPlex and
  expression-module rows were removed; dsRNA, glucose, hypoxia,
  T-cell-homeostasis, ER-stress, and ROS rows were either kept as
  contextual over-annotations or left unresolved when the cache was
  abstract-only.
• Completed the human BBC3/PUMA review. The curation centered PUMA on
  BH3-only derepression of pro-survival BCL2-family proteins, plus weak
  direct BAX/BAK activation, at the mitochondrial MOMP checkpoint. It
  narrows generic protein binding rows to proposed PUMA/BIM/BID-side
  molecular functions; removes high-throughput interactome edges and EGFR
  sequestration; separates upstream BBC3 transcription events from PUMA
  protein localization; replaces broad DNA-damage, ER-stress, and apoptosis
  rows with p53, ER-stress, or MOMP descendants where possible; and removes
  hypoxia and ERN1/UPR over-annotations.
• Completed the human BCL2L11/BIM review. The curation centered BIM on
  two complementary BH3-only activities at the mitochondrial BAX/BAK
  checkpoint: direct effector activation and inhibition of pro-survival
  BCL2-family grooves. The review avoids using BH3 domain binding for
  BIM-side GO:0005515 rows, removes high-throughput and viral sequestration
  interaction noise, separates transcriptional induction from BIM protein
  localization, tightens broad apoptosis/complex-assembly rows to MOMP or
  intrinsic signaling where evidence is proximal, and keeps abstract-only
  source rows unresolved when the exact curated interaction is not locally
  recoverable.
• Completed the human BID review. The curation centered cleaved tBID
  as an activator BH3-only relay from CASP8, CASP6, and other cleavage
  events to mitochondrial BAX/BAK outer-membrane permeabilization;
  removed the mis-scoped death receptor binding row; avoided
  directionally wrong BH3 domain binding replacements for BID-side
  GO:0005515 interaction imports; tightened broad release,
  mitochondrial-change, complex-assembly, and extrinsic/intrinsic
  signaling rows to direct MOMP or death-receptor descendants; and
  proposed a BID/BIM/PUMA-side BCL2-family effector activator molecular
  function term.
• Revalidated the existing complete human BCL2 and BCL2L1 reviews before
  moving deeper into the BCL2-family MOMP slice. BCL2 already captured the
  direct mitochondrial/ER anti-apoptotic role, conditional converted BCL2
  pro-apoptotic activity, BECN1/AMBRA1 autophagy inhibition, and remaining
  source-level uncertainties; BCL2L1 already captured the Bcl-xL/Bcl-xS
  isoform split and anti-apoptotic mitochondrial function.
• Completed the human MCL1 review. The curation centered canonical MCL-1L
  on BH3-domain binding in OMM BCL2-family complexes that restrain
  BAK/BAX-dependent MOMP; split the pro-apoptotic MCL-1S and caspase-cleaved
  fragment evidence away from the long isoform; removed high-throughput
  and substrate/stabilizer GO:0005515 rows; tightened broad or wrong-sign
  apoptosis terms to negative regulation of MOMP or cytochrome-c release;
  and kept autophagy, anoikis, mitochondrial fusion, and DNA-damage rows as
  secondary contexts.
• Completed the human BAK1 review. The curation centered BAK1 on its resident
  mitochondrial outer-membrane BCL2-family effector role in MOMP, cytochrome-c
  release, and BAK-complex pore assembly; removed the generic GO:0005515
  interaction import set; tightened broad apoptotic-process and membrane-
  permeability rows to MOMP; removed negative-regulation rows where BAK is the
  restrained target; and kept ER calcium, UPR, UV, and chaperone rows as
  secondary.
• Completed the human BAX review. The curation centered BAX on BH3-triggered
  mitochondrial outer-membrane permeabilization, cytochrome-c release, BAX
  oligomerization, and BCL2-family complex assembly; removed all generic
  GO:0005515 interaction rows; tightened broad apoptotic process and execution
  rows to MOMP or intrinsic signaling; fixed mitochondrial-fusion polarity; and
  removed the historical permeability-transition-pore complex assertion.
• Completed the human CYCS review. The curation kept CYCS's mitochondrial
  electron-transfer and apoptosome functions distinct, tightened broad apoptosis
  rows to intrinsic apoptotic signaling, added a proposed apoptosome complex
  assertion, moved inner-membrane Reactome projections to mitochondrial
  intermembrane space, removed cytosolic precursor translation as a functional
  localization, and removed high-throughput GO:0005515 interaction rows.
• Completed the human APAF1 review. The curation kept APAF1 centered on
  cytochrome-c-dependent apoptosome assembly in intrinsic apoptotic signaling,
  replaced CASP9 GO:0005515 rows with CARD-domain binding, removed generic
  regulator and cytochrome-c binding rows lacking a precise GO MF replacement,
  and marked stress, nutrient, tissue, and developmental outcome terms as
  over-annotations of the core apoptosome scaffold role.
• Completed the human CASP9 review. The curation tightened top-level apoptosis
  rows to intrinsic apoptotic signaling or effector-procaspase maturation,
  replaced APaf-1/procaspase-9 GO:0005515 assertions with CARD-domain binding
  or apoptotic caspase activator terms, removed high-throughput generic binding
  rows, and retained an isoform-specific Caspase-9S negative-regulation proposal.
• Completed the human CASP8 review. The curation kept CASP8's FADD/DED DISC
  assembly, death-domain receptor initiation, RIPK1-dependent necroptosis
  checkpoint, and gasdermin/procaspase maturation branches, while removing the
  generic GO:0005515 interaction set and tightening or deferring broad
  apoptotic-process, execution-phase, immune-differentiation, and readout-based
  rows.
• Completed the human CASP7 review from local cached evidence, covering its
  core executioner-caspase activity, RNA-assisted PARP1/RNA-binding-protein
  cleavage, and SMPD1 maturation in extracellular pore-repair contexts. The
  review removed uninformative high-throughput GO:0005515 rows, tightened
  generic GO:0006915 assertions to execution-phase apoptosis, and left
  abstract-only CASP7-specificity or localization gaps as UNDECIDED.
• Completed the human CASP6 review without provider-backed deep research, using
  local cached publications and pathway/model evidence. The review removed the
  high-throughput GO:0005515 interactome flood, kept direct CASP6 protease and
  ZBP1/RIPK3 PANoptosis biology, added a missing molecular-adaptor assertion for
  the catalysis-independent RIPK3/ZBP1 scaffold role, and left only evidence
  gaps unresolved.

2026-09-28

• Created the project as a mammal-first apoptosis campaign with three workstreams:
  module boundary review, generic GO:0006915 triage, and assay-derived
  over-annotation checks.
• Seeded the project with the existing draft intrinsic and death-receptor
  apoptosis modules rather than treating those pathways as unmodeled.
• Cross-linked the readout-quality and over-annotation projects because this
  project should reuse their rubric for caspase and viability readouts.
• Started the first Priority 1 gene by migrating all human CASP3 generic
  GO:0005515 rows away from the legacy MARK_AS_OVER_ANNOTATED action,
  adding missing PAINT propagation metadata, and marking the CASP3 review
  COMPLETE after clean validation.
• Seeded human CASP6 with just fetch-gene human CASP6; Falcon deep research
  could not run because no provider API keys were configured.