The genuine normal fetch produced 76 source assertions, two human products and 25 reference identities. All source terms, evidence codes, relations, supporting entities and existing product tags are retained verbatim. The gene was absent at the authenticated campaign base, so there is no inherited notes prefix. The final review is a biological audit; a normal computed status is not a campaign completion or merge receipt.
The description is a standalone biological summary. One integrated ATP-dependent chromatin-remodeling core connects the experimentally demonstrated motor to its DNA/histone engagement and context-dependent transcriptional outputs. No new process or molecular-function assertion is added. Existing broad ATP hydrolysis, substrate binding and core-compartment assertions are retained alongside the integrated activity; their overlap does not require removal of an original source assertion.
PMID:18378692 establishes human CHD8 motor activity, beta-catenin/WDR5 association and target-gene repression. Its exact selected cache is abstract-only; authentic indexed original Methods, Results and Discussion were additionally read. PMID:28533432 independently tests full-length human CHD8 cloned from HeLa cDNA and purified from Sf9 cells, with calf-thymus histones and defined DNA substrates. ATP hydrolysis, DNA EMSA, nucleosome sliding and accessibility are distinct assays. These establish remodeling rather than an inferred DNA strand-unwinding reaction. The electronic helicase and broad ATP-dependent DNA-activity assertions are refined to GO:0140658; this is not a universal negative statement about every untested CHD8 condition.
The two unchanged current human products are Q9HCK8-1 and Q9HCK8-2. The older NP_065971.2 fragment experiments identify CHD8-2, and their original source tags remain intact. Mouse short CHD8/Duplin lacks the motor region and must not be equated with either current human long product. PMID:19151705 mixes mouse short-form constructs, a human long construct and endogenous human-cell experiments. PMID:22083958 likewise combines endogenous human association/occupancy with detailed mouse short-form beta-catenin/H1 bridging experiments. Findings are stated at their actual scope, without claiming all human products were tested. In particular, H1 recruitment by tested mouse motorless short CHD8 can repress Wnt output without its ATPase domain; the integrated core does not imply that every repression mechanism is generated by the motor.
Primary sources directly resolve each of the original binding assertions even though the bounded current IntAct queries return no exact joins for these older publications. Query absence is not negative evidence. PMID:17938208 identifies endogenous ZNF143/hStaf-associated CHD8 and tests U6 and IRF3 regulation. PMID:26816381 explicitly validates the LRF/ZBTB7A–CHD8 interaction in its full main Results; the detailed supplementary construct panel was not independently recovered. Their DNA-binding transcription-factor class is distinguished from CHD8's own DNA engagement. PMID:20453063 uses CHD8-2 C-terminal fragments in Y2H, tagged co-IP and nuclear BiFC, not full-length CHD8; CHD7 missense effects differ across assay types. PMID:26949739 validates endogenous ACOT7 association in differentiated human SH-SY5Y cells. Class-binding refinements do not assign partner catalysis, a native transcriptional role for every pair, or purified binary contact where only complex recovery was measured.
RBBP5 is refined to transcription-coregulator binding by analogy to WDR5 because the PMID:20085832 abstract names CHD8 in a complex including PMID:20085832; ASH2L is removed instead because its own winged-helix and AT-hook contacts make that partner-class term less defensible. Both FAM124B alternatives associate with the tested CHD8-2 region PMID:23285124, but FAM124B's molecular activity remains unresolved. HNRNPL association was also validated PMID:36537238; the interaction is partly RNA-sensitive and is removed rather than converted to CHD8 RNA-binding activity. Exact original partners are preserved where the terms are modified, and bare protein binding rows are removed where no informative partner class is supported.
PMID:15960975 Fig.1B positively detects CHD8 in WDR5-purified bands. ROOT inspected the genuine author-hosted PDF and figure pixels; mixed subfractions leave particle composition incompletely resolved. PMID:18378692 independently recovers CHD8 with WDR5 but no MLL proteins. The newer PMID:42801681 adds reciprocal CHD8–KMT2A association. Its human co-IP Methods use HTR8 nuclear extracts; hTSC CUT&Tag and accessibility experiments are separate. These results justify context-scoped retention of existing MLL1 membership as non-core. GO part_of does not assert membership in every molecule or cell state. No invariant stoichiometry, purified intact complex, recruitment order or CHD8 methyltransferase chemistry is claimed. This synthesis supersedes the provisional unresolved recommendation in ROOT's initial bounded reader report; the signed follow-up records that change.
The H3K4me3-reader assertion remains supported at the original PMID:17938208 scope: methylated-peptide recovery from a partially purified hStaf-associated fraction. An RNase control addresses an RNA bridge but cannot exclude another protein. The review preserves the positive curated reader observation without claiming that an isolated CHD8 chromodomain was tested. Histone recognition is integrated into the remodeling core rather than presented as a separate fully resolved molecular operation.
PMID:36537238 measures alternative-splicing changes after partial CHD8 suppression in three human neural-progenitor lines, with ITSN1 isoform-PCR validation. This supports refinement from broad mRNA processing to regulation of RNA splicing. Chromatin engagement and HNRNPL association provide CHD8's performed regulatory context, while the proposed H3K36me3/HNRNPL link remains a model. Correlation does not prove direct CHD8 methyltransferase or splicease activity.
PMID:24998929 combines human genetic observations with zebrafish brain, enteric-neuron and transit phenotypes. Brain/digestive development remains non-core organismal context, not a direct CHD8 motility or digestive enzyme activity. p53-dependent apoptosis protection is likewise retained at its tested context. Human nuclear fractionation, chromatin occupancy and two visually inspected HPA images support the core nuclear compartment. Authentic Reactome events and the Q9HCK8 physical entity support nucleoplasmic location; candidate-set or co-complex membership does not transfer another protein's chemistry.
The default Falcon provider was genuinely attempted with a bounded timeout and the normal fallback returned a quota error. No deep-research report was fabricated. Manual primary research, original database records and independent bounded peers supply the evidence. Existing main publication bytes are protected; genuine fresh variants and externally recovered originals are separately archived. In particular, the selected PMID:22083958 cache is marked full but contains only Introduction/Discussion, while authentic external full text was recovered. The selected PMID:19151705 cache is partial, and the separate untouched normal full fetch supplies the additional read scope. Access flags are not mechanically inferred from cache completeness.
The review relies on primary PMIDs, GO_REF and Reactome identifiers, GOA source fields, and explicit reviewer judgment rather than self-authored extraction paraphrases as biological support. Historical PAINT placement is explicitly uninspected; target self-evidence and a small donor list are not failures. Original PTN nodes are retained as unresolved historical sources while direct target biology supports the retained functions.
All primary quotations introduced in YAML and notes are short exact anchors, counted cumulatively per source with repetitions and hyphen splitting. A separate audit records the totals, source hashes and literal checks. Normal validation, computed-status review, rendering, source guards and independent whole-peer receipts are recorded in the final handoff; ROOT adds normal history and handles publication.
The complete incoming notes above and its history event are preserved. This follow-up supersedes the earlier paragraph's RBBP5 class refinement and four generic-binding removals. The standing ClinGen project instruction retains supported generic binding as non-core when no evidence-backed informative replacement has been established. The positive ASH2L, FAM124B product-specific and HNRNPL assays therefore remain visible. The HNRNPL association remains partly RNA-sensitive; no CHD8 RNA-binding activity or native binary interface is invented. The incoming deletion of the self-quoted extraction artifact and its citations is retained. Its historical bytes and generating/extraction scripts remain in the earlier authenticated audit; it is not restored as independent biological evidence.
RBBP5 and WDR5 do not have interchangeable DNA-contact evidence. Human purified full-length RBBP5 (2–538), CTD3 and lysine mutants were tested by fluorescence polarization against 167-bp DNA. The actual result includes PMID:31544921. Human ASH2L assays likewise report PMID:21660059. Thus the explicit non-DNA-binding clause in GO:0001221 is a poor literal fit for these two partners; no DNA-binding domain prediction or lack thereof is used as a substitute for an assay. WDR5's histone-tail binding is a different molecular contact, and its supported coregulator-binding refinements remain. Nonspecific DNA association does not establish a sequence-specific transcription-factor class. These are partner-class checks, not new CHD8 activities.
Direct human motor evidence remains central: PMID:28533432. The positive H3K4me3 recognition assay used PMID:17938208, with an RNase control; it did not exclude another protein intermediary. The curated reader assertion is retained as non-core source context, without treating purification to a binary system as a universal prerequisite. DNA/ATP binding, recruitment and histone contacts are integrated features of the one motor core. The broad protein-containing-complex term is also retained as non-core context; it does not establish an invariant MLL1 particle. The existing positive-transcription parent is retained alongside its Pol II/III children as source provenance, not added coverage. No NEW term is introduced.
The genuine normal source projection contains 76 reviewed objects from 78 raw GOA records. Two pairs explain the difference, using zero-based raw record indices (header excluded). Records42/43 are GO:0016887 EXP PMID:28533432, differing in assigner/date (UniProt/Reactome) but representing the same activity assertion. Records35/62 are GO:0005515 IPI PMID:20453063 with CHD7/Q9P2D1: the first explicitly names Q9HCK8-2 and IntAct, while the second names canonical Q9HCK8 and UniProt. The normal seeder retained one product-scoped Q9HCK8-2 assertion at review17. This second pair is a product-scope projection, not a byte-identical duplicate. The raw GOA and all76 original review source objects remain unchanged; this follow-up does not claim a separately reviewed canonical-product object or rewrite the seeder.
Original publication caches remain byte-exact. The two partner papers were genuinely fetched normally after confirming their absence at the source head and current main. Authored quotes are counted across the entire current YAML and complete incoming-plus-appended notes, including repeated snippets; no inherited campaign paragraph is excluded from that cap. The external description cleanup, source-artifact deletion and incoming history are preserved. ROOT handles a new normal history record and final publication.