This guide accompanies the review, biological journal and source evidence. The review covers 118 original human GATA2 contexts, with no proposed NEW annotation. It retains 33 ACCEPT, 19 KEEP_AS_NON_CORE, 13 MODIFY, 36 REMOVE, 13 UNDECIDED and 4 TOO_GENERAL decisions. All 63 references have individual assessments. Genuine repository acquisition and the untouched native seed have now been reconciled. Target validation passed with two advisories: the distinct GO:0010628 target-context A/U group and the manual synthesis not being cited by an annotation. Strict matching passed for all 118 original GOA contexts. The current computed status is DRAFT.
The single core describes nuclear, RNA polymerase II-specific DNA-binding transcription-factor activity. Its regulatory and cell-fate facets reflect the work GATA2 performs at regulatory DNA and in context-dependent transcriptional complexes. Human endothelial reporter and DNA-binding evidence is explicit in the authentic PMID:1370462 abstract; the historical 470-aa clone is not silently assigned to either current natural product.
For PMID:15016828(https://pubmed.ncbi.nlm.nih.gov/15016828/), an independent source reader inspected selected original PDF pages 12–18, including Methods/Results, and visually inspected pages 13, 15 and 17. The coordinator read that reader's complete report and verified its source pins; this is not a claim that the coordinator independently read the original PDF or images. Human GATA2 was tested separately after the initial GATA6-bait discovery. Reciprocal COS7 and endogenous HUVEC association, GATA-dependent KDR reporters, motif controls and antibody-sensitive DNA complexes support the bounded regulatory claims. Human Hex plasmids and a separate rat Hex adenovirus experiment are distinguished. The final bibliographic page was unavailable, and a complete-paper read is not claimed.
Selected original PMID:21788589(https://pubmed.ncbi.nlm.nih.gov/21788589/) Methods/Results and the retained original-plus-supplement PDF support endogenous human GATA2 occupancy at HMOX1, GATA2 gain/loss and rescue experiments, endothelial tube formation and apoptosis. The PAK4/BAD branch is distinguished from GATA2/HMOX1. Selected main-text PMID:22499991(https://pmc.ncbi.nlm.nih.gov/articles/PMC3399424/) provides GATA2-specific loss and VEGFR2 rescue evidence beyond the abstract. Its detailed supplementary migration experiment was not recovered, so the source-specific migration context remains unresolved.
The authentic abstract and independently read original PMID:12045236(https://pmc.ncbi.nlm.nih.gov/articles/PMC2193540/) support GATA2's instructive eosinophil capacity. Human cord-blood recipients do not establish that every expression construct was human: the Methods explicitly describe mouse GATA1 mutant cDNAs, and separate mouse experiments are retained as such. Selected original PMID:21430145(https://pubmed.ncbi.nlm.nih.gov/21430145/) concerns dorsal-midbrain/superior-colliculus neuronal migration, not cortical electroporation. Two shRNAs, human GATA2 rescue and cell-death controls support that bounded phenotype. The PC12 neurite assay is a separate context.
PMID:27780851(https://pubmed.ncbi.nlm.nih.gov/27780851/) was available as an authentic abstract; attempted original-access routes did not recover the full paper. Its individual detailed genetic-interaction, migration, proliferation, localization and target-expression contexts are not inferred wholesale from the abstract. The reference assessments distinguish unavailable originals from available material that was not inspected.
Selected original PMID:23887938(https://pubmed.ncbi.nlm.nih.gov/23887938/) Methods/Results and Figure 5 captions distinguish GATA2–AR association from CCAR1's adaptor activity. GATA-motif-dependent reporter synergy, chromatin occupancy and biochemical association support the two informative transcription-factor-binding refinements. AR ligand recognition and CCAR1's molecular role are not transferred to GATA2.
The 36 bare GO:0005515 removals retain positive partner evidence. Ninety exact current IntAct records were checked for the original PMID, participant identities and negative flag. The source evidence stores compact assay/host/interaction locators. The target query was capped at 1,000 records; missing matches are not evidence of absence. In particular, the GATA2–ADH1A context in PMID:16153155 remains UNDECIDED because the authentic abstract describes DNA assays and the relevant full interaction experiment was not recovered.
The original HI-II-14 Table S2G and selected original Methods support human ORF yeast-fusion assays with individual retesting and identity checks. Orthogonal validation covered a subset, not every pair. HuRI original records retain P23769-1/ORF7593; current GOA target identity is gene-level. Literal partner isoform suffixes remain unchanged.
Two historical identity distinctions are explicit. Original S2G row 1883 names PRR20A/Entrez122183; current IntAct names P86479/PRR20C and carries the same Entrez122183 identity cross-reference. Original HuRI B8A4K4/ORF13662 and current A8MQ03 share the versioned ENSG00000197191.4 identity. These are retained gene-identity joins, not independent demonstrations of unchanged clone sequence.
For PMID:28473536(https://pmc.ncbi.nlm.nih.gov/articles/PMC8009048/), selected original main Results and HT/methyl-SELEX Methods were read. This is a recombinant transcription-factor DNA-selection study, not a GATA2 structure. The authentic retained workbook now supplies GATA2 eDBD in S1 row294 and positive methylated/unmethylated target motifs in S2 rows8188/8193. These original rows were replayed. No natural full-length product or GATA2-specific S3 class is inferred. The cis-regulatory refinement uses separately identified human promoter/ChIP evidence; random-library SELEX alone is not an endogenous cis-regulatory localization assay.
The actual PTHR10071 tree and MSA contain GATA2 at AN219/PTN002467679. Its ungapped sequence is exactly the 480-aa official product. All eight incoming IBA contexts descend from their actual IBD nodes: AN0/Eukaryota for five activities/locations, AN3/Eumetazoa for cell fate, and AN209 for angiogenesis and miRNA transcription. No negative record was found on the inspected target lineage. The compact evidence retains literal IBD records and both donor sets where raw89 differs by one dated/current ZFIN donor. Neither donor count nor self-donation is used to discount an IBA.
Nine distinct official typed paths support the thirteen refinements. Four further official is_a/part_of paths establish retained descendants for TOO_GENERAL judgments; morphogenesis retains its broad non-core claim because no such entailment was established. The source evidence includes complete selected is_a paths, rather than treating functional similarity as ancestry. Incoming term identifiers, qualifiers and evidence codes remain literal.
Official Reactome participant pages establish GATA2 through event R-HSA-8956568 → input 8956524 → component 8956530 → member 996735 and event R-HSA-996755 → input 996770 → member 996735. Entity 996735 explicitly identifies human UniProt P23769 residues 1–480 in the nucleoplasm. These are alternative-member sets; the model does not independently demonstrate every possible paralog combination or every cited experiment. Exact source URLs and typed chain relations are retained in the evidence.
The evidence's original_contexts.columns and rows reconstruct every original QuickGO field, including null values, assignedBy, dates, relationships, donor alternatives and extension logic. Its join table preserves all 118 raw-to-forecast indices and TSV rows. Actual native indices now match the genuine 118-row seed, and all 16 exported columns agree with the original projection.
Sixteen originals contain 31 extension atom occurrences. Fifteen conjunctive contexts permit 23 flat structured atoms. Raw76 contains four OR alternatives, each a two-atom conjunction. The current schema has no OR-group class, so those eight atom occurrences remain in the exact logical archive and the row reason; they are not flattened or split into invented original rows. This is a representation limit, not an allegation against the evidence source. Official RO/BFO relation identifiers and ontology labels were checked. Protein/RNA extension labels retain literal CURIEs where no separate official display name was checked.
Direct official FASTAs and the feature reconstruction verify P23769-1 at 480 aa and P23769-2 at 466 aa. The second lacks canonical residues 340–353, SAARRAGTCCANCQ. This establishes a sequence difference, not a distinct physiological activity or universal loss of binding. No experimentally unsupported functional isoform class is created.
The qualified source plan contains 58 paths. Actual closure at base 7b4e9568a48418cc8f0c23631f340bcbbc2fb649 contains 56 genuine files: 50 publications, three Reactome entries and three existing PTHR10071 source files. All 26 prior source objects remain byte-exact and 30 missing publications were acquired natively. Added pathway-bibliography entries R-HSA-9616222 and R-HSA-983231 remain absent after native ContentService HTTP 521 responses; this service outage is not evidence of pathway absence. Their previously verified official titles/model context remain distinct from native cache retrieval. No incoming GATA2 annotation cites either absent entry. Qualification uses the retained dated inventory and bounded namespace/owner comparisons, not a fresh global census. The separately owned family-review file is excluded. Shared PMID12045236 and PMID12483298 bytes have been pinned for immutable reuse by the GATA1 owner; R-HSA983231 remains reserved to this single acquisition owner until a genuine cache is available.
Six short primary excerpts occur once each in the review. Two additional companion occurrences are recorded separately: the manual synthesis repeats the short 1370462 anchor, and the appended journal adds one 20705609 anchor. The public ledger references their locations and hashes without repeating the excerpts. All six now match actual repository cache bytes, with coherent portable path/hash/byte-offset locators. Lossless native whitespace differences preserve every quoted word. Exact source documents remain unchanged. This new target has no inherited review, journal or historical quotation archive at the qualified base. Native acquisition, mapping and targeted validation outcomes are recorded. Genuine CREATE history and private render/link/whitespace checks accompany the DRAFT candidate. The term cache and all 56 actual source files remain unchanged.
Newly acquired full PMID:20705609(https://pmc.ncbi.nlm.nih.gov/articles/PMC2952242/) Methods, Results and Discussion were read and reassessed. Human GATA2 and human FOG2 constructs in mouse 3T3-L1 cells, with contact mutants and expression controls, support GO:0001221 transcription coregulator binding. The literal Q8WW38 IPI is preserved. Partial FOG2-mutant repression and the distinct GATA2-V296M proliferation phenotype are not reduced to complete loss of function. No new purified-affinity measurement, modern isoform mapping, figure-image or supplement reading is claimed.
The newly acquired full original PMID:10938104(https://pubmed.ncbi.nlm.nih.gov/10938104/) resolves the earlier uninspected GATA2–PML pair. Reciprocal human-protein association and site-dependent reporter potentiation with expression/DNA-binding controls support raw19/native16 MODIFY to GO:0001223 transcription coactivator binding. The literal P29590-4 partner is preserved without claiming a modern sequence crosswalk. GST/cellular-extract mapping is not purified-protein affinity. Unsuccessful initial endogenous hematopoietic assays, engineered Ba/F3 findings and positive endogenous NB4 leukemia association remain separate; fusion-oncoprotein results are not generalized to wild-type physiology.
The manual biological synthesis describes the inspected literature and source limits without claiming a provider run. The earlier three-advisory outcome remains historical evidence for published head 70e9232. The follow-up now passes with two advisories and a matching DRAFT status. Primary references remain the annotation evidence; the manual synthesis is a readable companion, not an independent source introduced to silence a warning.
For PMID:22499991, the native full-text flag overstates the extracted main-text coverage: main Methods/Results were read from the separately retained authentic original, while the cache contains the abstract and Discussion. Human GATA2 perturbation/rescue and VEGFR2 expression support contextual vascular repair. Mouse miRNA/TNF wound interventions are separate. The particular migration readout is unresolved because main Results measure vessel formation and the Discussion migration statement cites earlier papers; the relevant supplementary assay was not recovered.
The ADH1A positive transcriptional claim is accepted using its positive-element reporter and GATA2 EMSA findings together with independent transcriptional evidence, while the distinct ADH1A protein-interaction row remains unresolved. The miR-126 IDA and IGI rows share the exact same miRNA/endothelial conjunction and now accept the same biological claim; the unverified Q96SQ7 experiment is a source caveat. The four alternative target-expression contexts differ from the VEGFR2 context and remain UNDECIDED. No stale consistency exception is introduced for a pair absent from the current hard rule table.
The additional native85 anatomical wording is corrected to dorsal midbrain/superior colliculus. Earlier journal entries remain historical. The single nuclear RNA polymerase II transcription-factor core is retained; DNA contact and partner recruitment are facets of that activity, and the fate claim now also links the neuronal primary source. The cochlear source remains unresolved at the inspected abstract/IEP scope. The existing natural-product question and matched-comparison experiment remain the explicit functional gap.
Genuine EDIT history records this follow-up separately from the preserved CREATE session. Private rendering and relative-link checks passed. All 56 actual sources, term-cache bytes, original machine/context tables and natural products remain unchanged. The six original review excerpts and both companion occurrences have verified cache offsets; aggregate quoted text remains at most 16 words per source. Publication and integration are not claimed.