Research journal for the AI GO-annotation review. Provenance is recorded inline as
[PMID:xxxx "verbatim quote"]. This is a dark gene: the deliverable is a precise,
defensible statement of what is and is not known.
Bottom line on identity: despite the surrounding project framing, ppgn-1 is NOT an
epidermal/innate-immune/wound gene. It is the C. elegans paraplegin-subfamily
subunit of the mitochondrial m-AAA protease — an ATP-dependent, membrane-embedded
zinc metalloprotease of the inner mitochondrial membrane. All curation below is against
that identity.
The family is well characterised in yeast, mouse and human; ppgn-1's GO annotations are
IBA/IEA propagations from these orthologues.
To guard against propagating enzyme annotations onto a possible pseudoenzyme, I scanned
the ppgn-1 sequence itself (see ppgn-1-bioinformatics/, reproducible
analyze_motifs.py). All three catalytic motifs are intact and correctly ordered:
Interpretation: ppgn-1 is a catalytically competent m-AAA protease at the sequence
level, not a degenerate pseudoenzyme. This supports the enzymatic GO terms for this
specific protein, not merely by family membership. See
file:worm/ppgn-1/ppgn-1-bioinformatics/RESULTS.md.
KNOWN (high confidence, from orthology + domain/motif evidence):
- Encodes a paraplegin-subfamily m-AAA protease subunit (AAA+ ATPase + M41 Zn-peptidase).
- Retains intact ATP-binding (Walker A), ATP-hydrolysis (Walker B) and Zn-peptidase
(HExxH) catalytic motifs → catalytically competent in principle.
- Predicted mitochondrial, integral inner-membrane (two TM helices), matrix-facing
catalytic domains — the canonical m-AAA topology.
- Family function: ATP-dependent proteolytic quality control + regulated processing of
inner-membrane / matrix substrates.
NOT KNOWN (the real gaps for ppgn-1 specifically):
- No C. elegans experimental characterisation of ppgn-1: no mutant/RNAi phenotype,
no localisation, no biochemistry has been reported for this gene (all GO evidence is
IBA/IEA; the only UniProt reference is an unpublished sequence submission).
- Substrates unknown: which inner-membrane proteins ppgn-1 degrades or matures in the
worm is undetermined (no worm equivalent of the MrpL32 result attributed to ppgn-1).
- Complex composition unknown: whether ppgn-1 assembles with the AFG3L2-type subunit
spg-7 into a hetero-hexamer, forms a homo-oligomer, or is catalytically redundant with
spg-7 is untested.
- Physiological role unknown: whether ppgn-1 (as opposed to spg-7) contributes to
UPR^mt, respiration, mitochondrial morphology, fertility, longevity, stress/pathogen
responses, or is functionally dispensable, is unknown.
- Regulation unknown: expression, tissue distribution (Bgee: "Expressed in germ line
and 4 other cell types" — coarse, electronic), and conditions under which it acts are
uncharacterised.
All 10 GOA annotations are IBA/IEA and are consistent with a catalytically competent
paraplegin-subfamily m-AAA protease; none is contradicted, so none is REMOVEd.
- Core, well-supported family propagations → ACCEPT: mitochondrial protein processing
(IBA), m-AAA complex (IBA), metalloendopeptidase activity (IBA + IEA), ATP-dependent
peptidase activity (IEA), ATP binding (IEA), ATP hydrolysis activity (IEA).
- Correct but non-specific / non-core → KEEP_AS_NON_CORE with a note: proteolysis
(GO:0006508, generic parent of the specific processing/QC role), mitochondrion
(GO:0005739, generic parent of inner membrane), membrane (GO:0016020, generic parent
of inner membrane).
- The persistent uncertainty (no worm data) is captured in knowledge_gaps, not by
down-grading correct family propagations.
just deep-research-falcon worm ppgn-1 --fallback perplexity-lite was run FOREGROUND and
retried. Falcon timed out at the 600 s server limit on both attempts; the perplexity-lite
fallback returned HTTP 401 (account quota exhausted). No -deep-research-*.md file was
produced. Per repo policy I did not fabricate one. This review therefore rests on the
UniProt/InterPro record, the GOA evidence codes, the primary m-AAA/paraplegin literature
cited above (PubMed-verified, cached in publications/), and the reproducible sequence
analysis in ppgn-1-bioinformatics/.