Human ELF4 is the 663-residue ETS-family transcription factor Q99607. The official record has one displayed product and no alternative-products block. RUNX1 isoforms are interacting partner products; experimental MEF deletions and the disease-associated ELF4–ERG fusion are separate entities. None is added as a natural ELF4 splice product.
Human MEF DNA binding and promoter transactivation support an integrated transcription-factor mechanism. The original PMID8895518 abstract reports GM-CSF/IL3 promoter transactivation, while PMID14625302 explicitly reports an ELF4-responsive IL8 ETS site. Their full operative Methods were unavailable. PMID10207087 supplies reciprocal and domain-resolved MEF–RUNX1 binding, with cooperative IL3 reporter activation in a COS7 host. These complementary sources support precise polymerase-II transcription annotations without assigning every immune phenotype as a core activity.
The RUNX1-binding refinements use this independent mechanism, not merely the partner’s transcription-factor label. Historical AML1/AML1B constructs cannot be equated automatically with the current Q01196-1/-8 sequences. PMID14970218 genuinely includes a positive MEF–AML1 association result; its inspected BLK reporter/deletion experiments concern NERF2. The PML-body and transcription assertions citing that paper retain their original source IDs and biology, with two conditional PROBABLE/CHECK reference queries. PMID14976184 independently supplies MEF/PML association, PML-body localization and lysozyme activation. Unavailable supplemental material prevents a universal absence claim.
HPA supports nuclear-body and nucleoplasmic localization, but does not identify the bodies as PML bodies. The PML-specific evidence comes from the original localization study. Generic PML binding remains non-core under the task-specific policy retaining supported generic interactions; neither partner class nor association is promoted to a new scaffold activity.
Human and mouse inflammatory studies support contextual restraint of cytokine outputs. Human W251S, human W231R and mouse W250S/knockout experiments are kept distinct. The actual published Figure 5 and full caption of PMID35266071 support its TNF-alpha polarity; full Methods and supplements were not recovered, and the earlier preprint remains a separate version. These observations do not imply every downstream cytokine is a direct transcriptional target.
The four NK/NKT proliferation transfers remain UNDECIDED. The exact mouse donor records and authentic PMID12387738 abstract support positive development/function, but the operative division assays were not recovered. Cell abundance can reflect differentiation, survival or proliferation. This uncertainty concerns the specific division claims and does not negate the positive developmental observations.
All 42 imported source dictionaries and accepted decisions remain exact: 17 ACCEPT, 11 MODIFY, 10 KEEP_AS_NON_CORE and four UNDECIDED. The normal GOA file preserves all 43 raw contexts. Its two GO:0045944/IDA/PMID14625302 agency/date records collapse into source row34 without discarding the raw multiplicity. No NEW annotation or publication quotation is added. Existing normal caches remain unchanged; the additional PMID12387738 cache was fetched and qualified separately. The actual provider and fallback attempts failed and produced no report; no manual text is attributed to them.
The PAINT replay places all four assertions on the target ancestry and verifies the aligned canonical sequence. Target self-donation is legitimate. Two donor sets differ between frozen snapshots and remain separately recorded. Literal rule checking confirms two selected branches; an absent IPR006715 premise prevents claiming the inspected ARBA00029082 branch matches. Independent human transcription evidence is considered separately from that unresolved mapping provenance.
Source access and assay guide records exact cache flags, public routes, inspected sections and source limits. Reproducible database results and the offline replay expose the underlying database joins. Article bodies and figure images remain outside the custom public evidence archive.
The source-specific PML evidence supports a refinement of source row14 from generic non-core binding to GO:0001223 transcription coactivator binding. Reciprocal tagged/endogenous association and mammalian two-hybrid mapping accompany PML enhancement of MEF-dependent lysozyme reporter activity and endogenous lysozyme RNA without increased MEF protein abundance (PMID:14976184). This functional coactivation evidence supersedes the earlier generic-PML-binding disposition above. The tested interaction fragment is MEF 477–663; an important 477–517 region is inferred from overlapping truncations, not an isolated-peptide binding measurement. The paper leaves the downstream recruitment/modification mechanism unresolved, so this refinement does not assert a measured PML–basal-machinery bridge or autonomous ELF4 scaffold/coactivator activity.
The conditional source-reference query on row35 now addresses transcription specifically: PMID:14970218 supports MEF–AML1 association, while its inspected BLK reporter experiments concern NERF2/ELF2; PMID:14976184 supplies the MEF lysozyme-promoter activation evidence. The irrelevant PML-absence/localization clauses have been removed from that query. The original source reference, PROBABLE confidence, CHECK request and all other query fields remain intact. Row20’s separate PML-body query is unchanged.
The current action counts are 17 ACCEPT, 12 MODIFY, nine KEEP_AS_NON_CORE and four UNDECIDED across the same 42 source assertions and 43 raw contexts. The description, single product, integrated core and all references are unchanged. No new source row or publication quotation is added. Existing source-access limits remain in the source guide.