FDXR transfers electrons from NADPH to ferredoxins. One ferredoxin-NADP+ reductase activity accounts for electron supply to mitochondrial steroidogenic cytochrome P450 enzymes and the ferredoxin-dependent COQ6 reactions in ubiquinone biosynthesis. These are contributions of the reductase itself, rather than an inference from its being required for a downstream phenotype. The activity-centered core therefore uses GO:0004324 with the two existing biosynthetic-process terms. The mechanistic COQ6 reconstitution used reconstructed tetrapod-ancestral FDXR/FDX2 and short-chain intermediates; it is not described as a purified extant-human assay or evidence that FDXR forms a stable metabolon (PMID:2845396; PMID:38425362; PMID:38885337).
Human endogenous imaging and exact MitoCoP P22570-containing target groups support mitochondrial residence. The core uses the organelle-level location. Retained matrix and inner-membrane annotations carry their individual Reactome-model and yeast-ortholog inference boundaries. Human confocal imaging also reports nuclear/cytoplasmic signal; mitochondrial residence is not exclusive localization (PMID:29040572; PMID:34800366; PMID:9727014).
All seven curated splice-product objects are retained. The numbering in PMID:41511346 differs from UniProt: paper isoform 7 corresponds to P22570-4, while paper isoform 4 corresponds to P22570-7. In tagged MCF7 experiments the former is predominantly cytosolic, while paper isoforms 1 and 4 overlap MitoTracker. Altered electrophoretic mobility and targeting do not establish an exact endogenous cleavage site, equivalent activity of all products, or a defined cytosolic substrate. Knockout effects on DNA-damage responses do not by themselves establish direct DNA-repair activity. The exact sequence and source joins are in FDXR-isoform-source-joins.yaml.
The clinical description is limited to biallelic FDXR-associated mitochondrial disease with optic atrophy and neurological impairment, supported by the patient study. Its corrigendum changes three family genotype labels in Figure 1 and does not describe a correction to Figure 2 localization. The later adrenal-like-cell and mouse findings support steroid physiology while leaving enzyme-specific versus secondary redox effects separate (PMID:29040572; PMID:29554255; PMID:38885337).
The genuine Falcon attempt failed authentication with HTTP 403; its Perplexity-lite fallback failed with HTTP 401/insufficient quota. No provider report was produced. This authored synthesis and its source guide use identified original evidence; they are not provider-generated research. No new annotation, quotation or experimental proposal was added. The inaccessible 1991 mapping original and unrecovered PAINT tree/MSA retain their explicit source limits.
Evidence journals and full source-keyed decisions remain available in the unchanged direct notes, direct decisions, database notes, and database decisions. Their descriptions of intermediate lane scope document their provenance; the main review contains the combined decisions and core synthesis. Publication caches and machine UniProt/GOA records remain unchanged.
The production GO-CAM was read at the exact published commit. Its FDXR activity 675b862c00001526 is part of ubiquinone biosynthesis and directly positively regulates FDX2 activity 675b862c00001519. This supports the direct electron-supply core while the imported upstream qualifier stays immutable; ACCEPT is a biological claim-level decision, not an endorsement of every source metadata choice.
The two ancestral-reconstitution IDA records now include PROBABLE source-level CHECK requests. The actual experiments are direct assays of reconstructed tetrapod proteins, but the appropriate evidence-code representation for human P22570 is for UniProt to assess. No corrected ISS code, curator error as confirmed fact, or filed upstream report is asserted. Exact original Methods/Results and model locators are in the direct source guide.
The saved typed QuickGO ancestors exclude GO:0006091 and GO:0006733 from GO:0006744; the broad energy/metabolite annotation remains non-core, with no unsupported redundancy path. Repeated PAINT, yeast and matrix-model limits were moved into the database guide while each row retains its source route and biological scope. UniProt's Long = P22570-2 note is tentative inactivity with curator-inference evidence, distinct from the later 4/7 localization numbering. The FAD-containing reductase is already explicit; no new binding row or core is needed in this bounded follow-up.
All 27 imported source objects, 23 whole references/findings, seven splice products and zero quotation paths are preserved. Existing actions and core terms are unchanged. The project completion checkbox remains a separate actual-merge decision under the campaign workflow; a reviewed or approved open PR is not a merged gene credit.