DOLK

Summary: ER membrane localization is supported by phylogeny and target evidence.

Reason: The current family tree places Q9UPQ8/PTN002491113 below the literal IBD at PTN000324722; the target alignment sequence matches all 538 canonical residues. This is an ancestral-node inference, not a count of pairwise donors. Target experimental evidence among the descendants is legitimate, not circular. The literal location IBD includes SEC59, Arabidopsis EVN and human DOLK; official human topology/localization and the curated target GO-CAM agree.

Summary: Dolichol kinase is the experimentally grounded ancestral activity.

Reason: The current family tree places Q9UPQ8/PTN002491113 below the literal IBD at PTN000324722; the target alignment sequence matches all 538 canonical residues. This is an ancestral-node inference, not a count of pairwise donors. Target experimental evidence among the descendants is legitimate, not circular. The kinase IBD contains human DOLK, yeast SEC59 and Trypanosoma Q38DD5, whose current record carries an experimental kinase annotation. Human disease-cell assays directly confirm CTP-dependent dolichol phosphorylation.

Summary: DOLK directly performs the terminal phosphorylation of de novo Dol-P synthesis.

Reason: The current family tree places Q9UPQ8/PTN002491113 below the literal IBD at PTN000324722; the target alignment sequence matches all 538 canonical residues. This is an ancestral-node inference, not a count of pairwise donors. Target experimental evidence among the descendants is legitimate, not circular. Its chemical product is dolichyl monophosphate. The IBD placement is independently consistent with human kinase assays and the target GO-CAM process.

Summary: The electronic EC/Rhea mapping specifies the correct kinase reaction.

Reason: The current raw source names EC 2.7.1.108 and RHEA:13133. The official Q9UPQ8 reaction is CTP-dependent phosphorylation of dolichol, matching the current GO:0004168 definition and independent human assays. The failed standalone Rhea retrieval is not substituted for a claimed Rhea-page read.

Summary: The UniProt subcellular mapping correctly identifies ER membrane.

Reason: The saved official SL-0097 mapping resolves exactly to GO:0005789. Q9UPQ8 carries experimental ER-membrane curation from PMID:16923818; this specific membrane assignment does not imply that all observed cellular pools have been exhaustively characterized.

Summary: The EDA binary pair supports a cytokine-partner binding refinement.

Reason: Three PMID:25416956 records map DOLK/Q9UPQ8 and EDA/Q92838, including validated two-hybrid record EBI-10597898. Their reference ROGIDs match the saved sequences; this is not independent sequencing of the experimental inserts. The recorded 391-residue EDA reference contains its TNF-homology domain and is an established cytokine. GO:0019955 describes the binding partner class, without requiring receptor activity or an extracellular binding module in DOLK. The positive retested pair supports that bounded refinement. The assay does not identify the contact surface or processed EDA species, establish receptor competition, or demonstrate a native DOLK signaling interaction. Membrane topology alone does not negate the reported binding.

Proposed replacements: cytokine binding

Summary: The verified KCNA3 pair supports transmembrane transporter binding.

Reason: The source participant maps to DOLK/Q9UPQ8, and three PMID:26871637 records identify binding to the KCNA3/P22001 53–575 region, including validated two-hybrid retesting. That region retains the channel transmembrane segments. Original Figure 6D distinguishes the positive DOLK_1–KCNA3 clone from DOLK_2, which has different UBQLN1-clone partners and does not rescue the tested yeast defect. GO:0044325 describes the partner class, not channel gating modulation. The gene-level refinement is confined to the tested DOLK_1 result; it does not establish that the current canonical 538-residue product or every splice product was assayed. Reference-sequence matching and partial nucleotide identification do not resolve the unavailable historical clone sequences.

Proposed replacements: transmembrane transporter binding

Summary: BioPlex associations are retained as contextual binding evidence.

Reason: Two positive tagged-bait AP-MS associations recover DOLK with KCNA6/P17658 or LRRC4C/Q9HCJ2 in human cells. KCNA6 is a transporter-class partner. The historical aggregate additionally includes the LRRC4C association and is retained as non-core under the ClinGen project instruction, without treating LRRC4C as a transporter. These controlled complex recoveries support the imported association; their native contacting interfaces and any effect on channel gating remain unresolved. No neurite function is assigned to DOLK from the LRRC4C association.

Summary: The heterogeneous HuRI pair collection is valid non-core binding evidence.

Reason: All 19 current partner contexts join positive HuRI records, including one validated two-hybrid record per partner. EDA and KCNA3 support the same bounded cytokine-partner and transporter-partner interpretations used for the pair-specific rows from PMID:25416956 and PMID:26871637; keeping this aggregate non-core does not reject those conclusions. The aggregate additionally retains the other heterogeneous positive pairings under the ClinGen project instruction, without assigning them all to either class or inferring downstream partner functions. Repeated screening stages are not independent confirmations. The EDA validated record has no mutation feature; a separate array record specifies a DOLK residue-441 perturbation, so not every record is described as a WT experiment.

Summary: Later BioPlex recoveries corroborate contextual protein associations.

Reason: Three positive tagged-bait AP-MS records recover DOLK with KCNA6/P17658 in two source contexts and LRRC4C/Q9HCJ2 in one. KCNA6 retains its transporter-class interpretation; the heterogeneous aggregate also preserves the LRRC4C association as non-core under the ClinGen project instruction. The source supports contextual association in the assayed human cells, while native contacting interfaces and channel-gating effects remain unresolved. Shared partners with the earlier BioPlex study do not turn repeated recovery into independent mechanistic demonstrations.

Summary: Reactome places DOLK in direct Dol-P synthesis.

Reason: The pathway R-HSA-446199 contains the normal DOLK reaction R-HSA-446195. Its catalyst 449273 resolves human Q9UPQ8 residues 1–538 and converts dolichol plus CTP to Dol-P plus CDP. This is direct reaction participation, separate from downstream glycosylation dependencies.

Summary: The normal Reactome reaction represents DOLK kinase activity.

Reason: R-HSA-446195 explicitly links catalyst 449273, GO:0004168 and the Q9UPQ8 ER-membrane entity 449274. Its documented CTP/dolichol reaction agrees with human biochemical evidence.

Summary: The disease reaction preserves a valid gene-level kinase assignment.

Reason: R-HSA-4755600 is a failed reaction for a defined set of DOLK variants, not an experiment demonstrating normal activity of every mutant. Its normalReaction link returns R-HSA-446195 and its catalyst uses the same kinase class. Independent WT and patient-cell evidence establishes the gene-level function; the impaired-variant context remains explicit.

Summary: HPA nucleolar detection is retained as a contextual location.

Reason: Positive target-resolved HPA ICC/IF records support nucleolar localization: approved HPA066767 detects nucleoli in MCF-7, and approved HPA068706 detects them in HaCaT, MCF-7 and U2OS. The target summary rates the localization supported. These observations justify retaining the contextual location as non-core alongside the independently established ER pool. They do not establish nucleolar catalysis or identify a processed nuclear species. Image pixels and genetic specificity controls were not independently assessed; two antibodies are not presented as orthogonal validation.

Summary: Human DOLK complements the yeast Dol-P synthesis defect.

Reason: The curated IGI source is cross-species complementation with yeast SEC59, retained as that experimental context. The kinase reaction directly produces Dol-P; later independent human variants and yeast-rescue assays corroborate the pathway role without turning downstream CPY rescue into proof of a separate catalytic activity.

Summary: Human fibroblast assays directly support dolichol kinase activity.

Reason: The original Methods and Figure 4 assay crude human fibroblast extracts with radiolabeled CTP and dolichol, including a no-dolichol background and lipid-product extraction. Patient alleles markedly reduce activity, while WT human DOLK rescues the yeast temperature-sensitive defect. This is enzyme activity in cellular extracts, not purified-protein kinetics.

Summary: Independent patient-cell kinase measurements support the activity.

Reason: The study measures incorporation from gamma-32P-CTP into Dol-P using dolichol-19 and patient/control fibroblast material. Human WT and mutant ORFs were sequence-verified before yeast complementation. Loss-of-function phenotypes are supported by the direct chemical assay, rather than being used alone to infer catalysis.

Summary: The measured reaction supports direct Dol-P biosynthesis.

Reason: The chemical assay measures Dol-P formation, and human DOLK variants impair that reaction. Restoration of yeast growth and CPY glycosylation provides functional context; those downstream endpoints do not add an unmeasured glycosyltransferase activity.

Summary: The disease Reactome model retains ER-membrane placement.

Reason: The defined DOLK variant set in R-HSA-4755600 explicitly occupies ER membrane and maps each member to Q9UPQ8. This curated localization is also independently established for human DOLK; it is not interpreted as new imaging of every pathogenic variant.

Summary: The normal Reactome catalyst is an ER-membrane protein.

Reason: The exact reaction-to-catalyst-to-target entity path places Q9UPQ8 residues 1–538 in GO:0005789. This agrees with independent human topology/localization curation.

Summary: The human cDNA establishes the defining dolichol kinase activity.

Reason: The complete source abstract reports expression of a human cDNA with increased dolichol kinase activity and rescue of the yeast sec59 defect, while separating dolichol phosphorylation from DAG kinase activity. The full original Methods remain inaccessible; later direct human cellular assays independently corroborate this gene-level assignment.

Summary: The human topology study supports dolichol kinase activity.

Reason: The curated IDA and source abstract connect the expressed human membrane protein to dolichol phosphorylation and a cytoplasmically oriented CTP-binding region. The original full Methods are not claimed when inaccessible. Independent human cellular biochemical assays establish the same activity.

Summary: The reaction product establishes Dol-P pathway participation.

Reason: CTP-dependent dolichol phosphorylation is the terminal reaction of de novo Dol-P biosynthesis. The human cDNA complementation study and subsequent human biochemical measurements support this specific process without requiring a broader downstream glycosylation annotation.

Summary: Dol-P biosynthesis is the direct pathway of the characterized kinase.

Reason: The human DOLK activity produces dolichyl monophosphate. The original curated IDA is consistent with the source abstract, independent disease-cell chemistry and the explicitly modeled GO-CAM process.

Summary: The human expression study directly supports ER-membrane localization.

Reason: The complete abstract reports a polytopic ER-membrane protein with luminal N terminus and cytoplasmic C terminus. This supports the specific imported location, while detailed topology and mutant-method claims remain limited by unavailable full Methods.

Summary: The historical additional LLO-process proposal is withdrawn pending direct scope justification.

Reason: This is an inherited authored proposal, not a current GOA assertion. DOLK directly supplies Dol-P, whereas the current GO:0006488 definition describes glycosyl-chain assembly on that carrier. CPY rescue and reduced LLO abundance establish downstream dependence without locating DOLK in a glycosyl-addition step. Comparator annotations are mixed: SRD5A3 carries this term, while DHDDS, DOLPP1 and yeast SEC59 do not; no universal exclusion is claimed. The target GO-CAM models the established kinase/Dol-P process. The conservative active review retains that performed step and withdraws this unestablished expansion without deleting its historical identity.