flvcr2a review notes

• Curated in DANRE batch 04. Flvcr2a should no longer be reviewed as primarily a heme transporter: the current UniProt entry and 2024 paper support MFSD7c as a choline transporter at the BBB PMID:38302740. I retained ethanolamine/heme transport as non-core inferred capabilities.
• Follow-up: changed generic membrane CC from cross-aspect MODIFY to REMOVE; specific membrane locations remain reviewed separately.

Re-review 2026-09-28

Rewrote every review block against primary literature. Previously the file rested on
49 quotes from the UniProt flat file (the same FUNCTION line pasted under localization
and heme terms) plus deep-research paraphrases; it now cites four cached papers, with
UniProt lines used only where they are the right line (SUBCELLULAR LOCATION for the
ER/mitochondrion rows, CATALYTIC ACTIVITY for the reaction rows).

Newly cached and used: PMID:38778100 (cryo-EM/transport of human FLVCR1/FLVCR2),
PMID:20823265 (the 2010 FLVCR2 heme-import report), PMID:32369449 (Mfsd7c-KO
vasculopathy, abstract-only cache).

Key evidence now anchored in the review:

• The zebrafish protein itself was assayed, not just inferred: PMID:38302740. This is the support for the IDA row on GO:0015220 and for the core function.
• Mechanism: PMID:38302740, PMID:38302740, and PMID:38778100.
• In vivo barrier flux for the donor of the GO:0150104 ISS row: PMID:38302740 with PMID:32369449. The gene product performs the barrier-crossing
  step itself, so the participation test is met; the zebrafish in vivo test has not been
  done, which is flagged with [file:DANRE/flvcr2a/flvcr2a-deep-research-falcon.md "there
  is no direct in vivo zebrafish mutant phenotype, tissue expression map, or subcellular
  localization imaging for flvcr2a"].

Action changes:

• GO:0097037 heme export (IBA): KEEP_AS_NON_CORE -> REMOVE. The WITH/FROM is the ancestral
  node plus human FLVCR1 alone, i.e. only the exporting paralog, and the FLVCR2 literature
  states the opposite direction: PMID:20823265. Added propagation_review
  (PROPAGATION_BAD; WRONG_ORTHOLOG_OR_PARALOG, FUNCTIONAL_DIVERGENCE) with both sources
  marked SUPPORTS_SOURCE_BUT_NOT_TARGET. Donor identities were checked rather than assumed:
  MGI:2444881 = mouse Flvcr1, MGI:2384974 = mouse Flvcr2, Q9Y5Y0 = human FLVCR1,
  Q9UPI3 = human FLVCR2, Q91X85 = mouse Flvcr2, Q96SL1/Q6GNV7 = SLC49A4/DIRC2 orthologs.
• GO:0016020 membrane (IBA): REMOVE -> MODIFY to GO:0005886 plasma membrane. The node
  PTN000858822 also contains lysosomal SLC49A4/DIRC2 members, so the generic parent is the
  safe node-level call, but the site of action for this gene is resolved
  PMID:38778100. A same-aspect MODIFY avoids discarding a true localization.
• GO:0005886 plasma membrane, is_active_in ISS: PENDING -> ACCEPT (the last GOA row added on
  refresh). Distinguished in reason from the located_in row: the is_active_in form is the
  more accurate claim, since translocation is measured across the plasma membrane in both
  directions PMID:38778100.
• Heme rows retained as non-core (GO:0015232, GO:0020037) but re-argued from the actual
  assays: PMID:20823265 against PMID:38302740 and PMID:38302740.
• Ethanolamine rows (GO:0034228 ISS, GO:0034229 IEA) kept non-core with the conditionality
  made explicit: PMID:38302740 versus
  PMID:38302740 and PMID:38778100.
• Unchanged actions but rewritten reasons: GO:0015220 (IDA, IEA) ACCEPT; GO:0015871 ACCEPT;
  GO:0022857 MODIFY to GO:0015220; GO:0055085 MODIFY to GO:0015871; GO:0005886 (IEA, ISS)
  ACCEPT; GO:0005789 and GO:0031966 (IEA and ISS each) KEEP_AS_NON_CORE.

Other changes: description rewritten as standalone biology (uniporter mechanism,
localization, BBB choline supply, ethanolamine as secondary substrate, Fowler-syndrome
phenotype) with no curation commentary; reference_review added to all four PMIDs
(PMID:20823265 marked DISPUTED - correctly cited but its central import claim is
uncorroborated; PMID:32369449 marked full_text_unavailable and cited only for abstract
statements); core_functions reduced to the single choline-uniport entry with primary
quotes; questions and experiments rewritten around the flvcr2a/flvcr2b co-ortholog
redundancy and the untested zebrafish in vivo role. Validation: zero errors, zero warnings.