FAT4 is a large atypical cadherin whose extracellular region binds DCHS1 to create heterophilic cell contacts. Direct human ITC and interface-mutation/cell-boundary experiments support the adhesion activity (PMID:36797229); the original full-length human cloning and co-culture experiments are in PMID:28826487. The concise core is cadherin binding involved in cell-cell adhesion at the plasma membrane. The retained calcium-binding annotation describes the extracellular architecture, not a second independent signaling mechanism.
The intracellular region contributes to boundary-complex accumulation and internalization in engineered human-cell assays (PMID:39955614). This cellular result is not a new tissue-level planar-polarity annotation. YAP phosphorylation/localization and developmental effects differ across cortex, kidney, heart and liver contexts; the evidence guide distinguishes these rather than imposing a universal fly Fat–Dachs or mammalian Hippo-kinase chain.
The three inherited product objects remain exact. Official sequence reconstruction gives Q6V0I7-1 4,981 residues, Q6V0I7-2 3,222 residues, and Q6V0I7-3 4,980 residues. Product 2 removes canonical residues 1–1702 and therefore lacks both the canonical predicted signal peptide and the N-terminal EC1–4 region used in the direct binding experiments. Its additional sequence edits do not rescue a demonstrated localization or binding function. Product 3 removes canonical residue 4360; no functional equivalence is inferred from this small sequence difference. The canonical signal peptide at 1–38 is computationally annotated, not a measured universal cleavage boundary.
The human N-terminal constructs used for structure/ITC carry heterologous leaders and purification tags; full-length fluorescent constructs and intracellular deletion constructs are experimental reagents. Neither is silently assigned to a natural isoform. The 2025 intracellular deletion retains nine cytoplasmic residues and differs from all three inherited natural products.
Human FAT4 peptides are recognized by TNKS2; the separate modification assay used a mouse Fat4 fragment. This interaction is retained as non-core and does not establish FAT4 PARP or adaptor activity. The exact exosome proteomics target entry remains inaccessible and UNDECIDED. The verified tankyrase-paper erratum has unresolved content; no scope is guessed.
The source guide and two YAML tables connect every decision to its source and assay limits. The specific questions concern intracellular partners in a native cell context and the localization/adhesion of the N-terminally shortened product, without declaring a comprehensive literature gap. No new syndrome or clinical-mechanism assertion is introduced.
The ClinGen curation instructions require reference identifiers and precise evidence locators without adding publication quotations. The review therefore contains no new supporting_text; a successful reference validation checks titles, not unperformed quotation comparisons. The genuine research-provider attempt failed authentication. These notes and the source guide are authored literature assessments, not a generated provider report.
The two quantitative claims that need external originals can be checked directly:
| Claim | Original source and exact locator | Access boundary |
|---|---|---|
| TNKS2 recognition of two human FAT4 peptides; peptide 1 Kd 36.7 ± 11.0 μM and peptide 2 Kd 31.0 ± 5.6 μM | PMID:22153077 complete article and supplement, Figure 6 (PDF page 12) and Figure S7A (PDF page 28); Table S6 identifies the peptide sequences | Original figure pixels and Methods were read separately from the immutable abstract-only publication cache. The figure values are fit standard errors. |
| Reduced DCHS1 trans-endocytosis with the engineered FAT4 intracellular deletion | PMID:39955614 original article, Figure 1I, corresponding Results and Figure 1 caption | The two deletion groups in the actual plotted experiment and Results support reduced uptake. One caption sentence says the opposite. This is an internal inconsistency; no formal correction is asserted. |
| Distinct apicolateral membrane appositions, rather than classical adherens junctions | PMID:25355906(https://pubmed.ncbi.nlm.nih.gov/25355906/), Results describing Figure 5A–H; PMID:19506035(https://pubmed.ncbi.nlm.nih.gov/19506035/), cortical localization and electron microscopy | MDCK and mouse neuroepithelial contexts are kept separate from direct human assays. GO:0016327 describes the apical end of an epithelial cell's lateral plasma membrane. |
The numerical IntAct discrepancy is now recorded as two probable source-record queries on the binding annotation, for EBI-5333444 and EBI-5333467. Their qualitative binding evidence remains supported. No external correction report has been sent. The apicolateral replacement retains membrane-level localization, while source index 15 still records the broader apical compartment. The heterophilic-adhesion replacement is already represented by source index 7.
Other claim-level locators remain in the source guide: PMID:36797229(https://pubmed.ncbi.nlm.nih.gov/36797229/), binding Results and ITC Methods for human fragments and L379R; PMID:28826487(https://pubmed.ncbi.nlm.nih.gov/28826487/), cloning Methods and heterotypic-boundary Results for full-length human constructs; and PMID:23533145(https://pubmed.ncbi.nlm.nih.gov/23533145/), proteomics Methods for the mixed exosome preparation. The last paper's target-bearing supplementary FAT4 entry remains inaccessible and its annotation remains undecided.
Round 4 of projects/INVOLVED_IN_PRECOMPOSITION.md (2026-11-07): the NEW FAT4-DCHS1 row and the core-function molecular function now use the genus GO:0045296 cadherin binding instead of the precomposed GO:0098641; the adhesion context is carried by GO:0044331 cell-cell adhesion mediated by cadherin in the core functions. The earlier statement above that the concise core is cadherin binding involved in cell-cell adhesion is superseded.