This guide identifies the original evidence used for the annotation decisions. It is an authored locator and interpretation guide, not an independent primary source. The source measurement files are documented separately in DSP-evidence.
| Source | Operative evidence | Boundary |
|---|---|---|
| 9348293 | Genuine original body and embedded figure captions; human DSP N-terminal residues 1–584, JUP association and plaque clustering | The fragment arose from a stop and is not full-length IF evidence; no figure-pixel interpretation |
| 29878302 | Bounded publisher-original WT/mutant IF-overlay and localization text | Incomplete letter access; no full-figure or all-isoform claim |
| 21062920 | Human DSC2a/b cytoplasmic tails and DSP-NTP interaction experiments | Tail-specific result, not equivalent binding for every cadherin/DSP product |
| 25225338 | Results/Figures 1–7 and Methods: HeLa-library binary screen, human endogenous keratinocyte assays, EB1-binding-deficient mutants, Cx43 localization and communication | Human, rat and mouse experiments distinguished; lysate capture is not a two-purified-protein assay; DSP direct MT binding not demonstrated |
| 9864371 | Mouse Dsp deletion, early IF-network collapse and desmosomal defects | Conserved loss-of-function evidence, not a human knockout |
| 15479741 | Mouse immunoelectron Results and captions place DSP in desmosomal and fascia-adhaerens-like mixed plaques | Pkp2 perturbation; not universal epithelial adherens-junction residence |
| 16406610 | Complete abstract explicitly includes human, bovine, rat and mouse cardiac immunoelectron microscopy | Full source-specific body unavailable; no invented human assay details |
| 17227889 | Mouse ninein recruitment, cortical MT organization and full/head-plus-rod rescue | Biochemical DSP–ninein interface unresolved |
| 29959233 | Rat/mouse Cx43 targeting, turnover and rescue; human engineered DPII used in rat cardiomyocytes | Neither direct protease inhibition nor modern natural-product equivalence |
The selected core processes—desmosome organization, intermediate-filament organization and positive regulation of junctional protein localization—were compared with current typed ancestry in both directions. No pair is redundant. Functional cytoplasmic and cardiac mixed-junction pools are represented alongside desmosomes. Existing IF co-localization qualifiers are preserved; DSP is not asserted to be an IF polymer subunit.
PMID:11790773(https://doi.org/10.1074/jbc.M108765200) was recovered as an indexed original through ResearchGate, while its normal cache remains abstract-only. Methods 478–810 identify the human library/cDNA and DSP constructs; Results 869–1091 cover binary and cellular interactions; Figure 6/lines 1333–1380 distinguish PKP1-head and PKP2 binding from negative armadillo-repeat constructs. This original corroborates PKP1/PKP2 pair recognition without manufacturing unread details in other cited structural or phosphorylation papers.
The verified PKP2 class chain is human Q99959 GO:0005882 ISS annotation 328040827 → mouse Q9CQ73; rat F1M7L9 ISO annotation 57565728 → MGI:1914701 → the same Q9CQ73 mouse gene. Mouse annotation 357919950 is IDA from 16917092. The current GO:0005882 term falls under cytoskeleton. These curated component assertions do not imply that PKP2 polymerizes filaments or that the donor's full microscopy experiment was independently recovered. The binary DSP pair is separately grounded in the recovered human 11790773 original.
JUP is gamma-catenin; DSG1 and DSC2 are cadherins; EB1/MAPRE1 is a microtubule-associated cytoskeletal protein. Their verified partner classes clarify DSP binding without transferring the partners' own activities. The 19136012 abstract explicitly reports weak direct binding of the human DSG1-specific region to the DSP plakin domain; full Methods were unavailable.
FES/P07332 is a protein tyrosine kinase. 25852190 has the exact IntAct record EBI-10103287, anti-tag co-IP with FLAG-FES bait in a human DLD-1 context. 32707033 independently has EBI-28931656, N-terminal Strep/HA-FES in human HEK293 Flp-In T-REx AP-MS, recorded as spoke-expanded association. GO:0019901 describes binding to a protein kinase and does not require purified catalytic-site recognition. The refinement does not assert DSP kinase activity or direct contact proved by purified reconstitution.
The 18474624 original Figure 3/Results and Methods show PKP2-dependent PKC-alpha/DSP co-IP. The imported bovine PKC-alpha accession remains exact: human host cells and a separately specified human retroviral construct do not establish the species of every FLAG reagent. The COL17A1-specific assay in 12482924 remained inaccessible and is undecided, rather than inferred absent from the title.
31595153 supplies actual human nuclear imaging/fractionation, ChIP/FISH, TRF2 association/dependence and perturbation/rescue observations. These are positive contextual results. They do not establish purified DNA binding, a telomere enzyme or the proposed lamina mechanism. The 28891468 original supplies positive DSP/COPS3 association and DSP-dependent EGFR handling. A negative DSP-NTP binary test does not negate other constructs; COPS3 is noncatalytic and COP9 catalytic activity is not assigned to DSP. Much of the differentiation/rescue analysis perturbs COPS3 rather than DSP.
The 17475244 abstract directly reports DSP depletion, increased human HaCaT proliferation/G1–S progression and altered signaling/protein levels. It does not expose the full target-specific differentiation assay. That precise assertion stays undecided; positive proliferation and independently corroborated stabilization contexts are retained.
| Source | Exact measured record | Interpretation limit |
|---|---|---|
| 22658674 Table S1 | HeLa RNA-interactome Excel row 139, DSP/ENSG00000096696, crosslinked positive/control not identified; adjusted spectral-count p 0.0166297092443137 and ion-count p 0.00264221421315189. Whole-lysate row 759 is a separate control. | Actual stringent UV/oligo(dT)/RNase Methods read; no defined DSP regulatory RNA or product-specific function |
| 19056867 Table 1 | Human urinary preparation: 10 unique DSP peptides, 16 spectra | Fraction enrichment, no DSP-specific intravesicular protection or modern isoform resolution |
| 19199708 supplemental Table 2 | Paratid Exosomes sheet row 176, IPI00013933.2: 16 sequences, 16 spectra, 8.3% coverage | Single-donor human parotid preparation; historical isoform label not exact current product evidence |
| 21630459 / author thesis | Annex 8.2/Table 8.1: P15924, 2.1% coverage, three peptides | CTAB/DTT decapitation and repeated digitonin washes with microscopy-quality text; positive nuclear-fraction detection, not proof of a native mechanism |
The thesis's official catalogue issue date is 2014-09-19, while the PDF prints July 2014. The linked original PDF and its checksum are in the source manifest. Its later chromatin-study tables are not substituted for the Annex 8.2 sperm-nucleus study.
The HPA target page explicitly records supported cell-junction localization, with A-431/CAB037324 assay data. This was checked at the individual location level. No image-pixel read is claimed, and independent primary plaque evidence supplies the desmosome specificity.
All 12 Reactome event chains reach exact human P15924 entities through recorded relationships. Entity 201580 is the plasma-membrane target; the cleavage output of 201636 has unresolved boundaries; 8847736 is the cornified-envelope target. Event 6800426 includes DSP as a literal member of the ficolin-rich granule set and then at plasma membrane. Its underlying 23650620 DSP fraction table remained unavailable. The narrow granule verdict is therefore undecided, even though the model and independent plasma-membrane pool are positive. Likewise, inaccessible B-cell target data in 20458337 are not replaced by the urinary/parotid evidence.
10908733 was recovered as an indexed author original beyond its abstract cache. Actual resistant-envelope preparation and Table 2 identify a DSP-derived peptide; transglutaminases perform cross-linking. No isolated joined DSP peptide or DSP catalytic/template/cofactor step was demonstrated. The envelope and differentiation contexts remain positive, while assigning cross-link formation itself is excessive.
12373648 was recovered in full beyond its abstract cache. All operative Methods/Results/discussion and figure/table text were read, without image interpretation. The human family ECG/Holter, angiographic and genetic results are real. Exact current GO definitions additionally require His–Purkinje-cell attachment/binding or ventricular-cell action-potential modulation, which this study does not measure. These are over-specific annotation interpretations, not false patient data or a negative claim about DSP electrical roles.
Other source-specific full-text gaps are retained explicitly in the YAML reference assessments: 10852826, 16917092, 18343367, 19136012, 21756917, 22781308, 22889254, 23381804 and 23444369. Where independently recovered target evidence establishes the same gene-level claim, curator-supported assertions are retained. Where the exact claim cannot be resolved, UNDECIDED is used. Abstracts from donor studies 11781569, 12951053, 16406610, 25496840 and 35119382 are not described as full-paper reads.
The normal machine cache and its access flags remain unchanged even where a separate genuine original or supplement was recovered. The source guide is a route to that evidence, not an authored replacement for the cited publications.
Measurement records and Reactome membership paths are now explicit file references for the underlying target records. They are not independent evidence for the interpretations in this guide. Existing reproduction instructions, raw-input URLs/hashes and output bytes remain unchanged. In particular, the HeLa crosslinked-positive/control-negative DSP record remains affirmative RNA-contact evidence in that assay; no source-specific false-positive finding was established.