AATF review notes

2026-06-03 PN batch re-review

• Falcon deep research was already present at genes/human/AATF/AATF-deep-research-falcon.md; no new provider run was needed.
• Publication caching was refreshed with just fetch-gene-pmids human AATF. PMID:12450794 was newly cached as abstract-only.
• Core biology remains two-part: AATF is a nucleolar SSU processome component required for early small-subunit maturation PMID:34516797 and a nuclear RNA polymerase II transcriptional cofactor [PMID:10783144 "Here we describe Che-1, a novel human protein that interacts with hRPB11"; PMID:12847090 "Che-1 activates p21WAF1/Cip1 by displacing histone deacetylase (HDAC)1 from the Sp1 binding sites"].
• Proteostasis PN projection was evaluated conservatively. The PN candidate report projects GO:0030687 preribosome, large subunit precursor from Translation|Cytosolic translation|Ribosome biogenesis factor|pre-60S complex|ANN complex, but AATF's direct evidence and GOA rows support SSU processome / 40S maturation (GO:0032040, GO:0030490, GO:0042274) rather than pre-60S large-subunit precursor membership PMID:34516797. Do not add the projected pre-60S term for AATF without gene-level evidence.
• The existing ER unfolded protein response annotation is supported but remains non-core. Ishigaki et al. show AATF is induced by PERK-eIF2alpha during ER stress and promotes AKT1-dependent survival PMID:19911006, but this is a context-specific transcriptional regulatory output, not evidence that AATF is a chaperone, foldase, or direct protein-quality-control factor.
• Generic protein binding rows remain mostly non-core or modified to the more informative transcription coactivator term where the interaction directly supports AATF's Pol II/SP1/Rb transcriptional cofactor role. NGDN interactions are biologically coherent with SSU processome context but still do not make protein binding an informative core molecular-function annotation.