bchF

Summary: Right enzyme class, but at the top of the hydro-lyase hierarchy. The reaction the BchF family performs is specifically hydration of the C-3 vinyl group of (bacterio)chlorophyllide a (EC 4.2.1.165), and GO has no term for it, so the replacement has to be a new term request.

Reason: The IEA rests on family membership, which is well supported here by three independent signatures (NCBIfam TIGR02020 BchF, Pfam PF07284, InterPro IPR009905), and the chemistry of the family is established: recombinant BchF from Chlorobaculum tepidum hydrates a C3-vinyl group to a C3-1-hydroxyethyl group in vitro and runs the reaction reversibly (PMID:36822492), and the same activity was confirmed for 3-vinyl hydratases from four chlorophototroph lineages (PMID:28643169). For R. capsulatus itself, transposon-insertion mutants identified bchF as the enzyme that adds water across the vinyl group (PMID:8385667). So the hydro-lyase claim is true but far less specific than the evidence supports. No is_a child of GO:0016836 fits: a QuickGO search for "bacteriochlorophyllide" returns only GO:0036354, which is the activity of the NEXT enzyme (BchC, the 3-(1-hydroxyethyl) dehydrogenase), and no hydratase term for this reaction exists (checked 2026-11-01). The replacement is therefore the NTR filed in proposed_new_terms. Note the term is right on the chemistry and is not a neighbouring step's term: the preceding step is a reduction by BchXYZ and the following step an NAD-dependent oxidation by BchC, neither of them a hydro-lyase.

Proposed replacements: 3-vinyl bacteriochlorophyllide a hydratase activity

Summary: Wrong term, reached through a homonym. GO:0019685 "photosynthesis, dark reaction" is the carbon-fixation grouping term, defined as the reduction of carbon dioxide to sugar, and bchF has no role in carbon fixation. The InterPro2GO rule evidently read the family's "light-independent" pathway as GO's "dark reaction".

Reason: The GO definition of GO:0019685 is "A complex cycle of enzyme-mediated reactions which catalyzes the reduction of carbon dioxide to sugar... The reduced NADP (NADPH) and ATP are produced by the 'light' reactions" (QuickGO, 2026-11-01). BchF hydrates the C-3 vinyl group of a Mg-tetrapyrrole (PMID:36971575, PMID:8385667); it neither fixes carbon nor participates in a cycle that does. This is a demonstrably wrong electronic mapping rather than an over-propagated but defensible inference, which is the sanctioned ground for REMOVE on an IEA row. The cause is visible in the InterPro entry's own text, which describes IPR009905 as involved in the light-independent bacteriochlorophyll biosynthesis pathway, and the error is systematic: all four InterPro2GO rules that emit GO:0019685 are light-independent (bacterio)chlorophyll-synthesis families (IPR005969, IPR005970, IPR005971, IPR009905), and QuickGO reports 3,731 annotations to GO:0019685 of which all 3,731 cite GO_REF:0000002 - i.e. every GOA use of this Calvin-cycle term descends from this one misreading. The correct term for this family is GO:0036070, which the gene already carries from UniPathway, so nothing is lost by removing the row.

Summary: True, and the right pathway, but one level less specific than the gene's own evidence supports. The bacteriochlorophyll a pathway of R. capsulatus is light-independent, so GO:0036070 - which the gene already carries from UniPathway - is the supported term.

Reason: The family-level BP inference is sound: IPR009905 members act in bacteriochlorophyll biosynthesis, and in R. capsulatus the hydration step was genetically assigned to bchF (PMID:8385667). The refinement is warranted because protochlorophyllide reduction in this organism is carried out in the dark by BchL, BchN and BchB (PMID:8385667), making the whole branch light-independent; GO:0036070 is an is_a descendant of GO:0030494 and is already asserted on this gene from UniPathway UPA00671, so the replacement is a term the gene carries rather than a new claim. Keeping the parent as well would only restate it.

Proposed replacements: light-independent bacteriochlorophyll biosynthetic process

Summary: CACAO row. The cited paper is a genuine, cluster-wide mutational analysis of the R. capsulatus photosynthesis gene cluster and the claim it is used for is correct, so the row stands; it is refined to the light-independent child term for the same reason as the IEA row carrying this term.

Reason: What the paper shows for this gene: PMID:8158642 constructed directed mutations in twelve open reading frames of the 46 kb R. capsulatus photosynthesis gene cluster, of which seven lost bacteriochlorophyll biosynthesis with accumulation of specific Mg-tetrapyrrole intermediates. The cached record is abstract-only (full_text_available: false) and the abstract names no individual open reading frame, so bchF's inclusion cannot be read off it; it is corroborated independently, because PMID:36971575 states that deletion of bchF in R. sphaeroides and R. capsulatus causes accumulation of chlorophyllide a and 3-vinyl bacteriochlorophyllide a and cites this paper's DOI (10.1006/jmbi.1994.1260, reference 3 of its Crossref reference list) among its sources for that statement. That phenotype - loss of bacteriochlorophyll with accumulation of the unhydrated precursor - is exactly what an IMP for bacteriochlorophyll biosynthesis on a 3-vinyl hydratase should rest on, and it agrees with the independent transposon-mutant assignment of bchF in PMID:8385667. The term matches the claim; it is simply one level general, and GO:0036070 is supported because the pathway runs in the dark in this organism. GONUTS could not be consulted (see below), so the judgement is made from the literature; per the curator-deference rule no REMOVE was considered on an abstract alone, and none is warranted. On the batch-3 screen that selected this gene: the four CACAO rows citing PMID:8158642 span NCBITaxon:272942 (bchF, bchE, bchJ, all _RHOCB entries on the SB1003 genome) and NCBITaxon:1061 (bchM, the legacy species-level BCHM_RHOCA entry). Both ids denote Rhodobacter capsulatus, 1061 is the species node above the 272942 strain node, and the paper worked in R. capsulatus, so no isolate boundary is crossed and no locus mismatch exists; the split is UniProt entry granularity, not an annotation defect, and nothing is reportable against CACAO. GONUTS (gowiki.tamu.edu) was retried on 2026-11-01 and now returns HTTP 410 Gone with the notice "GONUTS has been retired"; the Wayback availability API has no snapshot of the relevant protein or paper pages, so that route is permanently closed rather than temporarily offline.

Proposed replacements: light-independent bacteriochlorophyll biosynthetic process

Summary: Correct and appropriately specific: this is the core process for bchF, and the pathway's light-independence is established genetically in this organism rather than merely assumed from the UniPathway label.

Reason: bchF performs a dedicated step of the bacteriochlorophyll a-specific branch - hydration of the C-3 vinyl group, with BchC oxidising the product to the 3-acetyl group of bacteriochlorophyllide a (PMID:36971575) - and its disruption abolishes bacteriochlorophyll synthesis with accumulation of the unhydrated intermediate (PMID:8158642, PMID:36971575). The "light-independent" qualifier is not taken from the UniPathway string alone: in R. capsulatus protochlorophyllide reduction is carried out in the dark by BchL, BchN and BchB (PMID:8385667), so the branch bchF belongs to operates without light. No more specific descendant exists, so this is as precise as the ontology allows.