This guide records the evidence behind the source-specific judgments. It is a locator and access record, not an independent experimental source. The review preserves 26 historical source objects, adds three genuine current-source objects and proposes no NEW annotation.
| Source | Material actually available and inspected | Scope and limitation |
|---|---|---|
| PMID:29459785(https://pubmed.ncbi.nlm.nih.gov/29459785/) | Genuine cached full original; Results on human protein/dimer and Methods on constructs, purification, radioactive product assay and TLC. | Full-length human 408-aa GPT expressed in Sf9 cells. Direct Dol-P reaction; Cys106 disulfide interpretation is finding-level disputed by the later study. |
| PMID:30388443(https://pubmed.ncbi.nlm.nih.gov/30388443/) | Genuine cached full original; human purification, product/TLC/MS assays, dimer analysis, variant kinetics and exon-trap Methods/Results. | Predominantly noncovalent dimer; some variants increase purified rate while disrupting splicing. TE671 exon traps are not native patient-muscle measurements. Model residues 7–400 and the post-TEV extra residue are construct/model details, not natural isoform definitions. |
| PMID:9451016(https://pubmed.ncbi.nlm.nih.gov/9451016/) | Complete original abstract and primary ENA sequence record; publisher HTML/PDF access failed. | Human lung-fibroblast cDNA, yeast complementation and isolated-membrane product synthesis. CAB04787.1/Z82022.1 translation is 400 aa, with 399 residues matching current Q9H3H5-2 and the explicit L25/R33 conflict. This is positive product-boundary evidence without exact sequence equivalence. |
| PMID:19549906(https://pubmed.ncbi.nlm.nih.gov/19549906/) | Genuine cached full Methods, Results and Discussion; Figures 1–4 described in the text. | Human HOK/A253 cells and oral tumor material, siRNA, EndoH/PNGaseF and junction measurements. Supports N-glycosylation pathway effects through DPAGT1's performed LLO step, not direct transfer of a glycan onto E-cadherin. |
| PMID:12872255(https://pubmed.ncbi.nlm.nih.gov/12872255/) | Complete cached abstract; publisher original inaccessible. | Explicit human patient microsomal GPT assay and Y170C expression in CHO cells. Preserve the curated activity and positive membrane-fraction evidence; finer ER assignment is independently supported. |
| PMID:8179616(https://pubmed.ncbi.nlm.nih.gov/8179616/) | Complete cached abstract; full original not recovered. | Measures Dol-P and GPT activity in fibroblasts; those CDG cases did not show reduced activity. No unseen fractionation or construct detail is claimed; current well-established function is retained. |
| PMID:6289658(https://pubmed.ncbi.nlm.nih.gov/6289658/) | Genuine author-hosted original, Methods pp.718–720, Results/Figures 1–2 pp.720–723 and Discussion pp.724–725. | Alpha-methylmannoside and glycoprotein acceptors establish different chemistry from DPAGT1's Dol-P reaction. This supports removing the misattributed GO:0003976 assertion, without treating the valid lysosomal-enzyme experiments as false. |
| PMID:32814053(https://pubmed.ncbi.nlm.nih.gov/32814053/) | Genuine institutional original PDF, STAR Methods “Automated Y2H interaction screening”; exact IntAct pair records. | Three positive human HTT–DPAGT1 records: EBI-25962194, EBI-26116455, EBI-26270586. Array, pooling and validated Y2H; no target-specific mammalian validation claimed. |
| PMID:22742743(https://pubmed.ncbi.nlm.nih.gov/22742743/) | Genuine normal full-text cache; variant/segregation, endplate and HEK293 AChR/tunicamycin experiments, especially Figure 3. | CMS context and glycosylation-dependent AChR surface expression. Tunicamycin buffering means rescue is not a clean assay of intrinsic catalytic competence for every mutant. |
| PMID:8940049(https://pubmed.ncbi.nlm.nih.gov/8940049/) | Complete genuine abstract; full Methods not recovered. | Hamster ER topology through selective permeabilization, epitope and cleavage-site approaches. Supports ortholog localization context; not a direct human isoform experiment. |
| PMID:36074821(https://pubmed.ncbi.nlm.nih.gov/36074821/) | Genuine full main cache, byte-identical to main; relevant Results and Methods summary read. Detailed supplementary files unavailable through the attempted routes. | Main text examines LYSET–GNPTAB Golgi interaction/localization. Current mouse Dpagt1 annotation is real, but target-specific supporting evidence remains unresolved, hence human Golgi UNDECIDED rather than presumed misattribution. |
No figure pixels were independently interpreted in this owner lane; the cited passages include original captions and Methods/Results text. The full-text flags remain exactly as fetched. Abstract-only caches are not silently promoted after separate original recovery.
The HTT pair records annotate an HTT sufficient-binding region of 1–546 and modifications in residues 18–38. These are participant features distinct from parent sequence identifiers/checksums. DPAGT1's partner field is not evidence that every native product was tested. The original Y2H protocol excludes reporter-autoactivating baits and uses multiple growth/reporters. This positive pair supports non-core binding; neither a specific physiological role nor a narrower functional class is established by the screen. Genericity alone is not grounds for removal or over-annotation.
PAINT node PTN000060284 carries the membrane and GPT-activity IBD assertions at the eukaryotic ancestor. The official PTHR10571 tree places human Q9H3H5 at leaf PTN002471652 below it. The human experimental descendant among the IBD sources is legitimate target grounding. The membrane claim is refined to ER membrane using target-specific evidence; the catalytic claim is accepted. No donor-count or circularity objection is made.
IPR033895 maps the specific GPT activity and LLO process. IPR000715 maps a broader membrane/phosphotransferase family encompassing other substrate chemistries; human target evidence supplies the specific reaction and compartment. The current Golgi IEA row transfers from P42867 and ENSMUSP00000056282; the literal mouse record is MGI IPI, PMID:36074821, with Q8BH26. Its unresolved source scope is explicitly separated from the independently supported ER pool.
GO:0003975 uses Dol-P as acceptor (EC 2.7.8.15). GO:0003976 uses lysosomal-enzyme D-mannose (EC 2.7.8.17). The full 1982 source and current definitions establish this difference.
The typed GO closure places LLO synthesis GO:0006488 within protein N-linked glycosylation GO:0006487. One catalytic core therefore represents the performed step, and the old broad process row is refined without introducing a duplicate NEW process. GO:0042803 is the homodimerization child of identical-protein binding GO:0042802; the refinement does not assert a particular intersubunit bond. ER membrane GO:0005789 reaches membrane GO:0016020 and intracellular membrane-bounded organelle GO:0043231 through the checked is_a/part_of closure.
Reactome R-HSA-446191 → catalyst 449231 → protein 449262 → reference entity 55898 identifies human Q9H3H5, residues 1–408, in ER membrane. R-HSA-4549334 → catalyst 9631918 → defined mutant set 4549346 links to normal reaction 446191. Its loss-of-function framing is not proof that every clinical variant has reduced purified turnover. The source summary's glycosidic-linkage shorthand is not substituted for the chemically explicit phosphotransfer reaction.
GO-CAM 65c57c3400000687, activity 65d7e4ac00000377, already models Q9H3H5 enabling the catalytic activity, part of LLO synthesis and occurring in ER membrane. No new process assertion is proposed.
The source-provenance YAML maps the 28 current seeded objects onto the 29-row augmented review while preserving every historical source dictionary. The old GOA archive is byte-exact; the live GOA is genuine normal output, not a concatenated or hand-edited file. Only historical row 5 is retired for its absent source identity, with biological judgment retained.
The historical review already contains repeated excerpts exceeding 25 words in aggregate for several papers. Those preexisting quotations are inventoried separately. Newly reproduced source text is limited to nine words from PMID:30388443, used once in the finding-level disulfide assessment; all other added support is reference-only. No copied passage is repeated here.