This guide separates the preserved repository caches from independently accessed originals and identifies the evidence behind the material decisions. The review contains 46 imported source assertions, three official alternative products and no new annotation. The reproducible evidence analysis reconstructs database records; it does not generate biological verdicts.
| Source and public locator | Actual evidence used | Limits |
|---|---|---|
| PMID:21723285(https://pubmed.ncbi.nlm.nih.gov/21723285/) | Separately accessed author-deposited original Methods, Results and Discussion: tagged human recombinant heavy chains, dynein-2 light-intermediate-chain copurification and polarity-marked microtubule gliding. | The repository cache remains abstract-only. This is an ensemble assay with selected track lengths, not single-complex processivity or live IFT. Intermediate-chain nondetection does not exclude those subunits from native dynein-2. Figure pixels and movies were not inspected. |
| PMID:31451806 full original | Selected Methods/Results/Discussion: human recombinant subunits expressed in insect cells, asymmetric heavy-chain architecture, accessory subunits and structural docking to a Chlamydomonas IFT train. | Human protein identity is distinguished from insect expression and algal train geometry. Cooperative loading and activation at the tip are mechanistic models in this source, not directly measured human IFT events. |
| PMID:29742051(https://pubmed.ncbi.nlm.nih.gov/29742051/) | Full-cache Results/Methods and genuine supplementary construct table: reciprocal visible immunoprecipitation of human DYNC2H1 tail 1–1650 and DYNC2LI1 in HEK293T cells. | Positive association supports the named partner class without requiring a purified interface. It does not establish the complete historical clone sequence or equivalent activity of every natural product. |
| PMID:25830415(https://pubmed.ncbi.nlm.nih.gov/25830415/) | Selected full original coimmunoprecipitation Results and Methods identify endogenous human DYNC2H1 with dynein-2 subunits in ciliated hTERT-RPE cells. | Light-chain localization is not automatically a heavy-chain localization experiment. Co-complex recovery does not by itself resolve a purified binary interface. |
| PMID:8666668(https://pubmed.ncbi.nlm.nih.gov/8666668/), indexed author original | Original antibody Methods, Golgi-localization Results/Figure 6 caption and microinjection Results: human HeLa staining; separate rat NRK antibody perturbation. | The repository cache remains abstract-only. The external original text, rather than a fabricated cache upgrade, supplies these details. COS cells and NRK cells are not human. No usable PDF or figure-pixel inspection is claimed. |
| PMID:12802074(https://pubmed.ncbi.nlm.nih.gov/12802074/) | Independently recovered original Methods/Results/Discussion, corroborated by a fresh normal full-text cache: heavy-chain localization in mouse tissues and distinct canine MDCK light-intermediate-chain experiments. | The existing repository abstract cache is preserved. The fresh full-text version is separate research evidence. Mouse apical/axonemal staining does not invalidate the earlier human HeLa Golgi pool. |
| PMID:21525035(https://pubmed.ncbi.nlm.nih.gov/21525035/), IntAct target query | Complete abstract and current curated human PEX14–DYNC2H1 pull-down records, attributed to Figure 2B and supplementary tables. The literal target ROGID matches Q8NCM8-1. | Full-original publisher attempts failed. These records do not resolve the original microtubule endpoint or establish a dynein-2 peroxisome-transport mechanism. Independent human motor studies ground the accepted microtubule annotation. |
| PMID:36973253(https://pubmed.ncbi.nlm.nih.gov/36973253/), publisher original and supplementary files | Full original antibody Methods explicitly name DYNC1H1 reagent 12345-1-AP. Results direct readers to Supplementary Figure 4, whose caption explicitly quantifies DYNC1H1, DYNC1LI2 and p150 in mouse pachytene spermatocytes. | The NAS attribution is corrected on positive reagent/target identity, not the title or nuclear-envelope topology. The complete saved body, supplementary text and two workbook name surveys corroborate that identity. No figure pixels, image-only target identities or hypothetical cross-reactivity are claimed. |
| PMID:19056867(https://pubmed.ncbi.nlm.nih.gov/19056867/), original-linked NHLBI identification table | Separately obtained full-cache isolation/mass-spectrometry Methods and the linked target table identify DYNC2H1, NP_001073932, with two peptides, assigned to reference 2. | The preserved repository cache is abstract-only. The positive urinary-exosome detection is retained as non-core; no peptide-sequence, isoform, secretion-route, topology or exosome-function inference is made. |
| PMID:30065109(https://pubmed.ncbi.nlm.nih.gov/30065109/) | Selected full review sections on dynein classes, ciliary beating, IFT cargo transport, subunit identity and protozoan gliding. | Review synthesis is distinct from primary experiments. Its gliding discussion is retained as an explicit limit on the ciliary-motion interpretation below. |
| PMID:32229580(https://pubmed.ncbi.nlm.nih.gov/32229580/) | Authentic complete abstract describing dynein-2 cargo transport within cilia. | Abstract-only access; detailed experiments in the full article are not claimed to have been read. |
| PMID:8832411(https://pubmed.ncbi.nlm.nih.gov/8832411/) and PMID:12432068(https://pubmed.ncbi.nlm.nih.gov/12432068/) | Authentic complete abstracts and the rat Q9JJ79 donor record provide ciliated-cell/cytoplasmic localization context. | Both new normal caches are abstract-only. No full-Methods claim or inferred species for every construct is made. |
| PMID:19442771(https://pubmed.ncbi.nlm.nih.gov/19442771/) | The preserved authentic abstract supports the inherited skeletal-ciliopathy and ciliary-maintenance context. | No newly inspected full-original experiment is claimed. |
UniProt Q8NCM8 identifies three official products. The primary sequence is Q8NCM8-1 (4,307 residues); Q8NCM8-2 has the Q3273-to-QIIGLKSW change and is 4,314 residues. Q8NCM8-3 deletes residues 736–4122, retaining the N-terminal 735 and C-terminal 185 residues. It is 920 residues long and lacks the AAA motor ring and stalk. The short product is not assigned full-length motor competence. Names, accessions and inherited sequence notes are preserved.
The official PANTHER gene-information service resolves the target to PTHR10676:SF352. Its tree and alignment, together with the literal PAINT IBD source, are saved in the replay bundle with exact hashes. The 787-node tree/alignment reproduce Q8NCM8-1 and confirm target descent through PTN002472914 and PTN000743491. The initially retrieved UniProt-linked PTHR46532 topology lacked the target; that lookup remains protected provenance rather than evidence against the IBA. The sibling dynein-1 IRDs are not exclusions on the target path.
The GO:0060294 decision challenges functional placement rather than node validity. The current term is is_a cilium movement, and part_of both cilium-dependent cell motility and cell motility. Dynein-2 moves cargo along the axoneme; axonemal dyneins generate ciliary bending. The review also explicitly recognizes IFT-powered protozoan surface gliding, where movement along a cilium transmits traction. Thus this decision does not deny all IFT involvement in cell locomotion: it distinguishes cargo movement along a cilium from the ontology's motion of the cilium itself. Motile-cilium/9+2-cilium locations and positive assembly/retrograde-transport assertions remain intact.
The Reactome IFT entries are traced through explicit complex/dimer membership to Q8NCM8 residues 1–4307 in the cilium or ciliary-tip compartment. The replay contains six source reaction roots and twelve target paths. These are model assertions, not twelve independent localization experiments. Current InterPro rules and the ARBA evidence are assessed at their own scope; the cytoplasmic ARBA output is not described as a fully reconstructed historical rule execution when its current conditions do not match the retrieved target cross-references.
Run the saved-input replay to verify the raw record counts, target identities, ancestry and exosome table row. Its compressed bundle contains genuine unmodified database source payloads plus explicitly described derived query selectors and product reconstructions. It contains no complete publication body. The large supplementary workbooks used for the separate target-name corroboration are available through the publisher source above; their archived hashes and inspected scope remain in the audit receipt rather than being republished here.
All existing machine sources, publication-cache bytes and availability flags are preserved. Actual original access is reported separately; an abstract-only repository cache does not imply that an independently recovered original was unavailable. The configured research-provider failures are documented in the appended notes. Existing notes and valid row-level excerpts remain unchanged, while four superseded core excerpts were removed when the two old cores were consolidated. No new quotation or new GO annotation was introduced.