ATF6

Summary: Retain supported Golgi membrane.

Reason: The original PMID:12782636 abstract distinguishes stress-induced ER-to-Golgi movement from subsequent proteolysis. Golgi membrane is a supported precursor compartment; no protease activity is attributed to ATF6.

Summary: Retain supported Golgi membrane.

Reason: SL-0134 maps the shipped UniProt Golgi-membrane location, independently corroborated by PMID:12782636 trafficking and the PMID:16236796 time course. This is a precursor state.

Summary: Retain supported Golgi membrane.

Reason: Complete human Reactome R-HSA-381135 summary explicitly places the ATF6 substrate at the Golgi membrane for S1P cleavage. The location is sound even though ATF6 does not catalyze the reaction.

Summary: Retain supported Golgi membrane.

Reason: Complete human Reactome R-HSA-381186 summary describes ATF6 movement from ER to Golgi; retain the destination membrane location. No cargo-transport activity is inferred.

Summary: Retain supported Golgi membrane.

Reason: Complete human Reactome R-HSA-420818 summary places the membrane-associated substrate before S2P release of its cytosolic fragment; retain Golgi membrane location.

Summary: Retain supported chromatin.

Reason: TFClass 1.1.7 is a structural classification rather than a direct chromatin experiment. Sequence-specific promoter binding and nuclear activation independently support this broad chromatin association; no TFClass subfamily reconstruction is claimed.

Summary: Retain supported transcription cis-regulatory region binding.

Reason: PMID:16469704 Figure 7C, examined in the primary-source consultation, captures HA-ATF6p50 from COS1 extracts with WT human CRP regulatory DNA but not the mutant probe. This supports the existing regulatory-region binding assertion, alongside independent ERSE evidence. The p50 construct species was not resolved; no purified human binding assay is claimed. The figure-specific evidence comes from the original publisher Results/legends at https://www.sciencedirect.com/science/article/pii/S0092867406000043, with the read scope recorded in ATF6-notes.md. The unchanged canonical cache is abstract-only; its availability flag does not describe the external reading.

Summary: Retain supported RNA polymerase II transcription regulatory region sequence-specific DNA binding.

Reason: PMID:11779464 reports ATF6-dependent XBP1 transcription and is explicitly cited for ATF6 DNA binding in UniProt. Retain the original regulatory-region binding specificity; do not transfer IRE1 RNA-splicing activity.

Summary: Retain supported RNA polymerase II cis-regulatory region sequence-specific DNA binding.

Reason: PTN005284906 phylogenetic cis-regulatory DNA-binding inference agrees with direct human ATF6 evidence. The target accession in WITH/FROM is legitimate experimental grounding, not circularity. The PAINT tree/MSA itself was not reconstructed.

Summary: Retain supported RNA polymerase II cis-regulatory region sequence-specific DNA binding.

Reason: PMID:24269637 explicitly connects ATF6 to XBP1 promoter activation in osteoarthritis context; independently established sequence-specific DNA binding corroborates the IDA. Full source-specific promoter assay details remain unread, not contradicted.

Summary: Retain supported DNA-binding transcription factor activity, RNA polymerase II-specific.

Reason: PTN005284906 supports inherited polymerase-II TF activity, independently grounded on human ATF6. ATF6B in the source set is a separate paralog, not a target splice isoform or a reason to reject the conserved activity.

Summary: Retain supported DNA-binding transcription factor activity, RNA polymerase II-specific.

Reason: The TFClass bZIP assignment agrees with the observed human DNA-binding transcriptional activity. Its broad polymerase-II TF term remains correct; no new family identifier is inferred.

Summary: Retain supported DNA-binding transcription activator activity, RNA polymerase II-specific.

Reason: PMID:14973138 directly reports that the human ATF6alpha N-terminal region activates ER-stress response transcription in HeLa cells. ATF6beta inhibition in this assay is context-specific, not a universal relationship.

Summary: Retain fibrillar center as a contextual annotation.

Reason: Current HPA subcellular table and HPA071169 ICC assessment support nucleoli fibrillar-center staining, specifically listed in SK-MEL-30. Retain this contextual location without inferring ribosomal transcription or assigning the nuclear TF core to the fibrillar center; image pixels and cleavage state were not resolved.

Summary: Retain eye development as a contextual annotation.

Reason: PMID:26029869 links human biallelic ATF6 variants to foveal hypoplasia and cone-layer defects, with impaired transcriptional activation in patient fibroblasts/controlled human constructs. Retain the existing developmental IMP as contextual; do not invent a direct morphogenetic mechanism or add a new process.

Summary: Retain supported DNA-binding transcription factor activity.

Reason: PMID:11163209 couples proteolytic liberation of ATF6 to nuclear transcription-factor function. The broad MF is correct, while the core synthesis uses its established activator specificity.

Summary: Retain supported DNA-binding transcription factor activity.

Reason: InterPro bZIP-domain associations agree with direct ATF6 DNA binding and transcriptional activation. Retain the supported broad TF activity without claiming the electronic mapping supplies an assay.

Summary: Retain supported DNA-binding transcription factor activity.

Reason: PMID:10564271 establishes the ER precursor-to-nuclear transcription-factor transition; broad TF activity accurately captures the processed product.

Summary: Retain supported DNA-binding transcription factor activity.

Reason: PMID:9271374 reports ATF6 DNA-binding/transcriptional context and SRF cooperation, including human HeLa-derived cDNA. Retain TF activity without converting an ATF6-VP16 fusion into proof of every native activation mechanism.

Summary: Retain protein binding as a contextual annotation.

Reason: Original PMID:16469704 Figure 6C detects endogenous human ATF6 in reciprocal co-IP with FLAG-CREBH in 293T, enhanced by tunicamycin, as examined in the primary-source consultation. Retain this contextual association. Co-IP, reporter synergy and the authors heterodimer interpretation are distinguished from purified binary affinity or a measured stoichiometry. The figure-specific evidence comes from the original publisher Results/legends at https://www.sciencedirect.com/science/article/pii/S0092867406000043, with the read scope recorded in ATF6-notes.md. The unchanged canonical cache is abstract-only; its availability flag does not describe the external reading.

Summary: Retain protein binding as a contextual annotation.

Reason: Actual PMID:19543265 selected Results and Methods show CRTC2 association with the transcriptionally active ATF6 N-terminal region and endogenous/tagged interaction work in mouse primary hepatocytes. Preserve the Q3U182 xenogeneic partner exactly; no human physiological gluconeogenesis function follows from the mouse context. Retain a contextual association, not a separate generic-binding core.

Summary: Retain protein binding as a contextual annotation.

Reason: Actual PMID:19543265 Methods independently identify endogenous/epitope-tagged CRTC2 IP-MS discovery in human HEK293T, and Results report ATF6 recovery and association. This supports the Q53ET0 human cofactor association, with mouse hepatocyte validation kept separate; a co-IP does not establish purified binary affinity.

Summary: Retain the established ATF6–XBP1 association with source-specific limits.

Reason: Retain the curated human ATF6–XBP1 association as contextual binding, with explicit curator deference. The actual human-human zipper-array design of PMID:20102225 was inspected, while its exact ATF6–XBP1 supplementary cell remains unread. The independently established interaction in the UniProt SUBUNIT record and the mouse ATF6alpha/XBP1 study PMID:17765680 corroborate the biological association; the latter is not relabeled as a human binding assay. No target affinity, full-length cellular assay or source-specific pair measurement is reconstructed from those corroborating records.

Summary: Retain protein binding as a contextual annotation.

Reason: The PMID:22682248 abstract, shipped UniProt THBS4 statement, and targeted original Figure 5/Methods reading support luminal-domain association. Purified binding fragments are mouse; human ATF6 GAL4 constructs provide separate cellular support, with HEK293 versus COS text/caption discrepancy preserved. Retain contextual human association without claiming purified human binary binding or vesicular catalytic activity.

Summary: Retain the established ATF6–XBP1 association with source-specific limits.

Reason: Retain the curated ATF6–XBP1 association as contextual binding, deferring to the original human IPI and the independently corroborated biological interaction. PMID:23661758 measures bZIP dimerization by solution FRET, but the precise human pair value and construct sequence remain unread. PMID:17765680 supplies mouse mechanistic corroboration, not a replacement human assay. No affinity threshold, cellular occurrence or particular source measurement is invented.

Summary: Retain contextual ATF6–ATF6B association with source-specific limits.

Reason: The UniProt processed-product SUBUNIT record independently describes heterodimerization with ATF6B, consistent with the existing curated human IPI. Retain that contextual association with curator deference. The exact human ATF6–ATF6B numerical FRET entry and construct sequence in PMID:23661758 remain unread, so this retention does not certify a target affinity or physiological stoichiometry. ATF6B is a distinct paralog, not a splice product of ATF6, and its regulatory effect is context dependent.

Summary: Retain contextual ATF6–HSPA5 association with source-specific limits.

Reason: ATF6–HSPA5/BiP association is established independently of the compound-focused title and is corroborated by the immutable UniProt interaction record and primary BiP–ATF6 work discussed in the notes. Retain the original curated association as non-core binding with explicit curator deference. The exact PMID:27238082 target experiment and its listed erratum were not reconstructed; no HA15-specific mechanism, purified binary affinity or melanoma phenotype is inferred for this annotation. This judgment concerns the established association, not certification of every source-level experiment.

Summary: Retain supported nucleus.

Reason: PMID:12782636 distinguishes inhibited nuclear appearance from preserved ER-to-Golgi trafficking; nucleus is the destination of processed ATF6.

Summary: Retain supported nucleus.

Reason: The PTN005284906 nuclear-location inference is supported directly for processed human ATF6. Target self-inclusion is valid; ortholog nodes and other paralogs are not a claim of ATF6B identity.

Summary: Retain supported nucleus.

Reason: PMID:11163209 explicitly links S1P/S2P processing to release of the nuclear transcriptionally active ATF6 fragment. Retain nucleus for that molecular state.

Summary: Retain supported nucleus.

Reason: PMID:11256944 explicitly includes ATF6alpha along with G13/ATF6beta and reports stress-dependent nuclear translocation. The paralog-focused title is not misattribution evidence.

Summary: Retain supported nucleus.

Reason: The actual PMID:16236796 ATF6 comparator Results report nuclear signal after 120 min DTT, distinct from the CREB4 behavior. Methods identify FLAG-ATF6 constructs. This supports the location despite the paper’s CREB4 title.

Summary: Retain supported nucleus.

Reason: The combined ARBA/donor/location mapping is corroborated by direct human precursor-processing and nuclear-translocation evidence. Preserve all original donor identifiers; no uninspected donor-species reconstruction is claimed.

Summary: Retain supported nucleus.

Reason: The PMID:23661758 bZIP-network paper is only NAS for localization here. Nuclear action is independently established for processed human ATF6, so retain the location while avoiding a false nuclear-imaging claim for that study.

Summary: Retain supported nucleus.

Reason: PMID:10564271 identifies the soluble nuclear p50 product of membrane precursor proteolysis; retain the nuclear compartment for that state.

Summary: Retain nuclear envelope as a contextual annotation.

Reason: PMID:10866666 describes stress-dependent redistribution of nuclear-envelope-associated ATF6 to the nucleoplasm. The envelope is a plausible observed membrane pool, ancillary to nuclear transcriptional activity; no inference that every envelope pool is transcriptionally active.

Summary: Retain supported nucleoplasm.

Reason: Current HPA table lists supported nucleoplasm across several human cell lines, agreeing with the independently established nuclear TF product. The antibody data do not resolve precursor versus fragment on their own.

Summary: Retain supported nucleoplasm.

Reason: PMID:10866666 supports stress-dependent nucleoplasmic accumulation and transcriptional activation with NF-Y/YY1. Retain the activity-associated location.

Summary: Retain supported nucleoplasm.

Reason: Complete Reactome R-HSA-381026 summary explicitly describes entry of processed ATF6 into the nucleus, supporting the curated nucleoplasm location.

Summary: Retain nucleolus as a contextual annotation.

Reason: Current HPA supports nucleolar staining, with hTERT-RPE1 serum-starved cells explicitly listed. Retain the observed context without inventing a nucleolar MF, ribosome-biogenesis role or cleavage-state assignment.

Summary: Retain supported endoplasmic reticulum.

Reason: Current HPA supports ER localization and lists U2OS staining, concordant with independent membrane precursor evidence. Its broad ER term is not a claim of soluble lumenal TF activity.

Summary: Retain supported endoplasmic reticulum.

Reason: PMID:11256944 explicitly includes ER-associated ATF6alpha and stress-triggered movement. Retain this independently corroborated precursor location.

Summary: Retain supported endoplasmic reticulum.

Reason: Actual PMID:16236796 comparator Results show transfected ATF6 initially in the ER and then Golgi/nucleus after DTT; retain the starting precursor compartment.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: PMID:12782636 documents ER-to-Golgi transport of the membrane-bound transcription-factor precursor; ER membrane is supported.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: PMID:10564271 directly identifies ATF6 as an ER transmembrane precursor. This is membrane anchoring before activation, distinct from the soluble nuclear TF state.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: ARBA00028685 and SL-0097 agree with the reviewed UniProt ER-membrane statement and primary PMID:10564271 evidence. No electronic-source assay is invented.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: The NAS PMID:11256944 ER-membrane assignment is concordant with that paper’s alpha/beta precursor context and independent PMID:10564271 membrane evidence.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: Complete Reactome R-HSA-381158 summary places the ATF6–BiP precursor complex at the ER membrane before dissociation. ATF6 is not inferred to perform BiP’s unfolded-protein chaperoning.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: Reactome R-HSA-381186 begins with membrane-anchored ER ATF6 and ends at the Golgi; retain the source compartment, without assigning transport motor activity.

Summary: Retain supported endoplasmic reticulum membrane.

Reason: The complete R-HSA-9921802 summary explicitly describes ATF6 as an ER-membrane transducer whose luminal domain associates with BiP. Retain its corroborated location although this event foregrounds IFI6; do not infer direct ATF6–IFI6 binding or antiviral activity.

Summary: Retain supported Golgi apparatus.

Reason: Current HPA supports Golgi apparatus staining in SK-MEL-30 and SiHa, consistent with independent stress-dependent trafficking. No claim that the antibody distinguishes precursor fragments.

Summary: Retain supported Golgi apparatus.

Reason: Actual PMID:16236796 ATF6 comparator Results describe Golgi localization after 30 min DTT. The target-specific observation supports this location independently of CREB4.

Summary: Retain cytosol as a contextual annotation.

Reason: Reactome R-HSA-381026 describes the released ATF6 fragment in cytosol before nuclear entry. Retain that transit pool, but the synthesized DNA-binding activity is localized to the nucleus.

Summary: Retain cytosol as a contextual annotation.

Reason: Reactome R-HSA-420818 describes S2P release of an ATF6 cytosolic domain. Retain the transient released state; ATF6 is the cleavage substrate, not S2P.

Summary: Retain supported regulation of DNA-templated transcription.

Reason: The InterPro-derived transcription-regulation annotation agrees with direct promoter activation by ATF6. Broad but accurate regulation is retained; specific activation appears in the core.

Summary: Retain supported regulation of transcription by RNA polymerase II.

Reason: PTN005284906 polymerase-II transcription regulation is consistent with human ATF6 activation data and conserved UPR biology. No PAINT topology or loss-of-function claim is invented.

Summary: Retain supported regulation of transcription by RNA polymerase II.

Reason: The NAS bZIP-network reference does not itself establish all transcriptional output, but polymerase-II regulation is independently demonstrated by ATF6 promoter/activation studies. Retain the clearly correct function with that read boundary.

Summary: Retain supported regulation of transcription by RNA polymerase II.

Reason: PMID:10564271 directly links nuclear processed ATF6 to ER-stress gene induction. This is its core transcription-regulatory role.

Summary: Refine protein folding to positive regulation of protein folding.

Reason: ATF6 increases expression of ER folding machinery rather than folding client polypeptides itself. PMID:10564271 establishes activation of the endogenous GRP78/BiP gene by the released transcriptional region; the folding-capacity interpretation is also concordant with the chaperone-induction findings in PMID:17765680. GO:1903334 preserves that regulatory role more precisely than replacing this folding assertion with the entire ATF6 signaling pathway. No direct foldase activity or measured folding-rate assay is claimed; the original TAS and source remain unchanged.

Proposed replacements: positive regulation of protein folding

Summary: Retain supported signal transduction.

Reason: PMID:9271374 places ATF6 in stimulus-responsive SRF/transcription signaling. The broad signal-transduction assertion is compatible with its independently established ER-nucleus pathway; preserve the source rather than falsely ascribe ER-specific experiments to this early SRF paper.

Summary: Retain visual perception as a contextual annotation.

Reason: PMID:26029869 establishes human cone dysfunction and impaired visual responses associated with ATF6 variants, with age-dependent mouse retinal effects kept separate. Retain this existing phenotype-supported contextual process; ATF6 is not a phototransduction enzyme.

Summary: The source-specific evidence for positive regulation of autophagy remains unresolved.

Reason: PMID:28747345 abstract describes ORMDL3 perturbation, mouse splenic B-cell survival and an ATF6–Beclin1 pathway; PubMed also indexes human material. The exact ATF6-specific intervention, human construct and autophagic-flux controls remain unread. These access limits support uncertainty, not a wrong-species/removal claim.

Summary: Retain supported membrane.

Reason: PMID:11163209 directly studies membrane-bound ATF6 processing. Retain the broad membrane location, with ER/Golgi specificity captured in existing location rows and no claim that the processed nuclear fragment remains membrane anchored.

Summary: Retain enzyme binding as a contextual annotation.

Reason: PMID:11163209 establishes ATF6 processing by S1P/S2P and the original IPI names enzyme partner O54862. Retain the curated enzyme association as activation context; it does not establish that ATF6 is a protease or its enzymatic regulator. Exact partner-species/reagent reconstruction was not recovered in the source review.

Summary: Retain supported endoplasmic reticulum unfolded protein response.

Reason: The PTN005284906 UPR inference is independently corroborated by direct human ATF6 stress activation and transcriptional output. A short donor set is not a reason to reject a PAINT inference; its topology remains unread.

Summary: Retain supported endoplasmic reticulum unfolded protein response.

Reason: PMID:19816510 primarily assays Purkinje-cell GRP78 loss in mouse; its actual Introduction describes ATF6 among ER UPR transducers. The TAS UPR role is independently well established for human ATF6. Do not represent the source as direct human ATF6 knockout evidence.

Summary: Retain supported response to endoplasmic reticulum stress.

Reason: PMID:11163209 directly connects ER stress to ATF6 proteolysis, nuclear entry and transcriptional output, supporting the core stress-response role.

Summary: Retain supported response to endoplasmic reticulum stress.

Reason: PMID:11256944 includes ATF6alpha ER-stress activation as well as beta, supporting the broad stress-response term despite its G13 title.

Summary: Retain supported response to endoplasmic reticulum stress.

Reason: Original PMID:16469704 includes ATF6p90/p50 ERSE reporter controls, stress-enhanced ATF6 association and ATF6-dependent output, as examined in the primary-source consultation. These agree with the established ATF6 stress response. Preserve the unresolved p50 construct species boundary; the CREBH foreground is not a misattribution argument. The figure-specific evidence comes from the original publisher Results/legends at https://www.sciencedirect.com/science/article/pii/S0092867406000043, with the read scope recorded in ATF6-notes.md. The unchanged canonical cache is abstract-only; its availability flag does not describe the external reading.

Summary: Retain supported response to endoplasmic reticulum stress.

Reason: PMID:22013210 is a review centered on IRE1 within the integrated UPR. The broad ATF6 stress response is independently established; retain TAS without pretending the review is an original target assay.

Summary: Retain supported ATF6-mediated unfolded protein response.

Reason: ARBA00091677 agrees with the demonstrated ATF6 activation-to-transcription pathway. The term includes the cleavage-dependent signaling sequence and downstream activation; no catalytic cleavage is assigned to ATF6.

Summary: Retain supported ATF6-mediated unfolded protein response.

Reason: PMID:32046286 explicitly reviews ATF6 activation and transcriptional response. NAS is consistent with that secondary source and the primary evidence; do not call it a new independent assay.

Summary: Retain supported ATF6-mediated unfolded protein response.

Reason: PMID:22013210 supplies pathway-level review context, and direct ATF6 primary evidence supports the specific signaling branch. Preserve TAS and distinguish IRE1 RNA splicing from ATF6 transcription.

Summary: Refine ERAD quality control pathway to positive regulation of ERAD pathway.

Reason: PMID:17765680 explicitly assigns ATF6alpha/XBP1 transcriptional induction of ERAD components, not the recognition/retrotranslocation/ubiquitination/proteasomal steps specified in GO:0036503. Refine to positive regulation of ERAD pathway; the original NAS remains NAS, with mouse knockout/MEF evidence and human corroboration distinguished.

Proposed replacements: positive regulation of ERAD pathway

Summary: Retain identical protein binding as a contextual annotation.

Reason: Retain established ATF6 homodimerization with explicit curator deference and independent UniProt corroboration. The primary-source consultation confirmed that PMID:20102225 measured human-human purified zipper arrays, but its exact ATF6 self-pair supplementary cell was inaccessible. This decision does not claim that the individual measurement or affinity was read, nor infer dimerization from bait recovery alone.

Summary: Retain identical protein binding as a contextual annotation.

Reason: Retain established ATF6 homodimerization as contextual assembly, deferring to the original IPI and independently corroborating UniProt statement. The primary-source consultation examined the actual PMID:23661758 FRET/network Results, but the human ATF6 self-pair numerical entry and construct sequence remained inaccessible. No target affinity or in-vivo complex occurrence is invented.

Summary: Retain established ATF6 self-association with an explicit TAP/MS limitation.

Reason: Retain independently established ATF6 homodimerization as contextual assembly, using the same curator-deference boundary as the other existing self-binding annotations. PMID:25609649 includes human ATF6 among HEK293T TAP/MS baits. Its Methods state that bait self-identifications were temporarily removed before SAINT and added back after filtration. Bait recovery alone is not evidence of a second interacting molecule; the precise curated self-association evidence was not reconstructed. This caveat remains explicit without treating it as evidence against the known biological association.

Summary: The source-specific evidence for positive regulation of apoptotic process remains unresolved.

Reason: PMID:14752510 abstract directly concerns CHOP/Bax apoptosis and hsp70–DnaJ protection. The specific ATF6 perturbation/control and direct pro-apoptotic claim remain inaccessible in the original full text. A different foreground gene does not justify rejection; do not infer ATF6 apoptosis merely from CHOP induction.

Summary: Retain supported sequence-specific DNA binding.

Reason: PMID:28473536 is an actual human-TF methylation-sensitive SELEX study. Its ATF6-specific supplemental matrix was not read, but the curated sequence-specific binding assertion is independently established for ATF6. No methyl-C preference is inferred.

Summary: Retain supported positive regulation of transcription by RNA polymerase II.

Reason: PMID:11163209 links proteolytic release of ATF6 to increased transcription; positive polymerase-II regulation is core, with no protease activity assigned to the substrate.

Summary: Retain supported positive regulation of transcription by RNA polymerase II.

Reason: PMID:11256944 explicitly includes alpha-driven transcription in the mammalian UPR. Retain positive transcriptional regulation while keeping beta as a separate paralog.

Summary: Retain supported positive regulation of transcription by RNA polymerase II.

Reason: PMID:14973138 directly demonstrates potent HeLa transcriptional activation by the alpha N-terminal construct; retain the specific positive regulatory role without universalizing the beta comparison.

Summary: Retain supported positive regulation of transcription by RNA polymerase II.

Reason: Original PMID:16469704 Figures 6D–F show ATF6p50 activation of murine SAP and human CRP reporters with CREBH synergy, examined in the primary-source consultation. This supports transcriptional activation alongside direct human ATF6 evidence from other sources. Reporter species, H2.35 cellular context and unresolved p50 construct species are kept separate. The figure-specific evidence comes from the original publisher Results/legends at https://www.sciencedirect.com/science/article/pii/S0092867406000043, with the read scope recorded in ATF6-notes.md. The unchanged canonical cache is abstract-only; its availability flag does not describe the external reading.

Summary: Retain protein heterodimerization activity as a contextual annotation.

Reason: Original PMID:16469704 Figure 6C reciprocal co-IP and Figures 6D–F functional synergy support the authors contextual CREBH–ATF6 heterodimer interpretation. Retain the existing curated heterodimerization assertion with those bounds; endogenous ATF6 in the 293T association experiment is human. This is not a claim of purified binary binding, exact stoichiometry or resolved p50 construct species. The figure-specific evidence comes from the original publisher Results/legends at https://www.sciencedirect.com/science/article/pii/S0092867406000043, with the read scope recorded in ATF6-notes.md. The unchanged canonical cache is abstract-only; its availability flag does not describe the external reading.

Summary: Retain supported RNA polymerase II transcription regulator complex.

Reason: Retain a transcription-regulatory bZIP complex based on independent ATF6 promoter activity and corroborated subunit interactions, with explicit curator deference for the original PMID:23661758 IPI. The primary-source consultation examined the FRET main Results but not the exact target supplementary complex edge. GO:0090575 does not imply physical RNA polymerase II membership; no target stoichiometry is invented.

Summary: Refine pathway regulation to participation in the ATF6-mediated unfolded protein response.

Reason: The demonstrated activity is execution of the ATF6 response through ERSE-dependent chaperone-gene transcription. The human ATF6 overexpression result in PMID:9837962 establishes this transcriptional output, but does not by itself distinguish an additional regulatory step acting on the ATF6 pathway. GO:0036500 explicitly includes signaling through its downstream transcriptional endpoint and is the better-supported term. This precision change does not assert that a pathway component can never regulate its own pathway or that an unread part of the original paper contains no such mechanism; the original evidence code and source are preserved.

Proposed replacements: ATF6-mediated unfolded protein response

Summary: Refine pathway regulation to participation in the ATF6-mediated unfolded protein response.

Reason: The demonstrated activity is execution of the ATF6 response through ERSE-dependent chaperone-gene transcription. The human ATF6 overexpression result in PMID:9837962 establishes this transcriptional output, but does not by itself distinguish an additional regulatory step acting on the ATF6 pathway. GO:0036500 explicitly includes signaling through its downstream transcriptional endpoint and is the better-supported term. This precision change does not assert that a pathway component can never regulate its own pathway or that an unread part of the original paper contains no such mechanism; the original evidence code and source are preserved.

Proposed replacements: ATF6-mediated unfolded protein response

Summary: Retain supported sequence-specific double-stranded DNA binding.

Reason: PMID:28473536 assays human TF binding to defined DNA libraries; ATF6-specific supplemental binding values remain unread but independently established ERSE DNA binding supports the curated double-stranded-DNA binding assertion. No methylation preference is inferred.