CAMK2A

Summary: Conserved calcium/calmodulin-regulated kinase activity is the principal activity of CAMK2A.

Reason: The cached PAINT record contains the positive IBD for GO:0004683 at PTN002389359. Human alpha-domain biochemistry and neuronal E183V experiments independently support this activity. Target Q9UQM7 among descendant evidence is expected, not circular. Calcium/calmodulin initiates activation; Thr286 autophosphorylation can maintain autonomous activity after calcium falls.

Summary: Refine a correct broad catalytic mapping to the established calcium/calmodulin-dependent kinase activity.

Reason: The domain/keyword mapping correctly recognizes a protein phosphotransferase, but human CAMK2A has experimentally established calcium/calmodulin-regulated serine/threonine kinase activity. Use GO:0004683 for the target-specific catalytic activity; the original machine term and mapping source remain preserved.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: The enzyme/family mapping agrees with independently established CAMK2A activity.

Reason: EC:2.7.11.17 and the CAMK2 family signature support the calcium/calmodulin-dependent kinase assignment. Independent human alpha-domain and E183V experiments establish the target activity; this is not accepted solely from a conserved generic kinase domain.

Summary: ATP binding is the relevant nucleotide-binding mechanism.

Reason: CAMK2A uses ATP as its phosphate donor. The human alpha kinase-domain structures in PMID:20668654 are bound to ATP-competitive inhibitors, not ATP. Independently established ATP-dependent phosphorylation supports the more informative ATP-binding term. This co-substrate interaction is part of catalysis rather than a separate core function.

Proposed replacements: ATP binding

Summary: Refine a correct broad catalytic mapping to the established calcium/calmodulin-dependent kinase activity.

Reason: The domain/keyword mapping correctly recognizes a protein phosphotransferase, but human CAMK2A has experimentally established calcium/calmodulin-regulated serine/threonine kinase activity. Use GO:0004683 for the target-specific catalytic activity; the original machine term and mapping source remain preserved.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: The experimentally tested partner is the ionotropic glutamate receptor subunit GRIN2B.

Reason: The preserved partner Q13224 is GRIN2B/NR2B. The complete PMID:19453375 abstract describes CaMKII-alpha-GFP/NR2B experiments in HEK293 cells and GST pull-down assays, including activation-state mutants. Refine generic binding to ionotropic glutamate receptor binding; this does not assign receptor channel activity to the kinase.

Proposed replacements: ionotropic glutamate receptor binding

Summary: This source partner is CAMK2D, not calmodulin.

Reason: Q13557 is the delta kinase subunit and its exact association with CAMK2A is also recorded in UniProt. PMID:20668654 examines all four human CaMKII kinase domains, while its calmodulin co-crystal is specifically delta. Retain the curated alpha/delta interaction and refine to kinase binding; the delta/calmodulin structure does not justify replacing this partner-specific assertion with calmodulin binding.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported HSP90AB1 interaction to Hsp90 protein binding.

Reason: The source records UniProtKB:P08238 (HSP90AB1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HSP90 client-network study supports a kinase/chaperone interaction. Its exact target row was not independently inspected, and no particular client-folding or stabilization mechanism is inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: Hsp90 protein binding

Summary: Refine the supported CAMK2D interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13557 (CAMK2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a kinase-interaction study coupled to a TRAIL sensitization screen; the individual CAMK2A partner experiment was not inspected. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported CAMK2B interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13554 (CAMK2B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The available Methods describe affinity purification and mass spectrometry in HEK293T cells with spoke/matrix network scoring. The specific pairwise target table was not inspected, and coassociation does not by itself establish direct contact or ciliary function. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported CAMK2B interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13554 (CAMK2B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The paper uses affinity purification in T-REx293/SH-SY5Y systems and complementary yeast two/three-hybrid assays. The exact alpha/beta target record was not inspected, so no particular orthogonal assay is assigned to this edge. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported HSP90AA1 interaction to Hsp90 protein binding.

Reason: The source records UniProtKB:P07900 (HSP90AA1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The study compares affinity-purified signaling networks in HCT116 and HKE3 cells with quantitative mass spectrometry. The exact pairwise target entry was not inspected; retain the curated association without transferring colorectal-cancer or KRAS mechanisms to CAMK2A. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: Hsp90 protein binding

Summary: Retain the curated association with ZBTB32 in its tested interaction context.

Reason: The source records UniProtKB:A0A0C4DGF1 (ZBTB32), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Refine the supported CAMK2B interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13554 (CAMK2B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The kinase interaction network uses affinity purification/mass spectrometry and controlled network filtering. The individual target-pair record was not inspected. Co-purification supports association without proving a direct interface or transferring the partner kinase activity. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with PLPP1 in its tested interaction context.

Reason: The source records UniProtKB:O14494 (PLPP1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Actual alpha-kinase manipulation supports the calcium/calmodulin-regulated enzyme.

Reason: PMID:35568036 reports calcium/CaM-CaMKII control of interferon responses and explicitly uses activated CaMKII-alpha T286D/T305A/T306A. That experiment supports the alpha enzyme rather than an inference from the bacterial title. The engineered construct bypasses aspects of normal calcium regulation; its rescue does not establish all endogenous alpha-dependent signaling contexts.

Summary: The source includes real alpha/beta kinase activity despite its cellular gamma/delta emphasis.

Reason: PMID:11972023 identifies CaMKII-gamma as the STAT1-associated kinase and tests gamma/delta cellular constructs, but also uses purified rat-brain alpha/beta CaMKII to phosphorylate STAT1 TAD and histone H3 in a calcium/calmodulin-dependent assay. Independent human alpha biochemistry corroborates the conserved kinase activity. The mixed preparation is not described as a pure human-alpha experiment.

Summary: The source includes real alpha/beta kinase activity despite its cellular gamma/delta emphasis.

Reason: PMID:11972023 identifies CaMKII-gamma as the STAT1-associated kinase and tests gamma/delta cellular constructs, but also uses purified rat-brain alpha/beta CaMKII to phosphorylate STAT1 TAD and histone H3 in a calcium/calmodulin-dependent assay. Independent human alpha biochemistry corroborates the conserved kinase activity. The mixed preparation is not described as a pure human-alpha experiment.

Summary: Retain the curated interferon/STAT1 context with an explicit isoform boundary.

Reason: PMID:11972023 provides cellular gamma/delta STAT1 interaction and signaling experiments and an alpha/beta rat-brain kinase control. The cached GO-CAM 62b4ffe300004489 explicitly assigns human Q9UQM7 activity within interferon-gamma/JAK-STAT signaling. Defer to that curated target attribution while distinguishing it from independent proof of endogenous human-alpha signaling. This contextual role is not a separate core. The historical retired source record, where present, remains retired.

Summary: Retain the curated interferon/STAT1 context with an explicit isoform boundary.

Reason: PMID:11972023 provides cellular gamma/delta STAT1 interaction and signaling experiments and an alpha/beta rat-brain kinase control. The cached GO-CAM 62b4ffe300004489 explicitly assigns human Q9UQM7 activity within interferon-gamma/JAK-STAT signaling. Defer to that curated target attribution while distinguishing it from independent proof of endogenous human-alpha signaling. This contextual role is not a separate core. The historical retired source record, where present, remains retired.

Summary: Use the established calcium/calmodulin-dependent kinase activity for this target.

Reason: The broad kinase assignment is compatible with CAMK2A. Human alpha structural/biochemical and E183V functional data establish the more specific calcium/calmodulin-regulated serine/threonine kinase. This target-level refinement does not imply that every original assay independently tested calcium dependence or that autonomous phosphorylated CAMK2A requires ongoing calcium binding.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: The preserved binding partner is metabotropic glutamate receptor 5.

Reason: The literal partner P41594 is GRM5, a G protein-coupled glutamate receptor. PMID:28130356 studies CaMKII-associated synaptic proteins; the available abstract/Discussion does not reproduce this individual target assay. Retain the curated positive association with that access limit and refine the binding class to the exact type-5 metabotropic receptor, rather than ionotropic receptor binding.

Proposed replacements: type 5 metabotropic glutamate receptor binding

Summary: Use the established calcium/calmodulin-dependent kinase activity for this target.

Reason: The broad kinase assignment is compatible with CAMK2A. Human alpha structural/biochemical and E183V functional data establish the more specific calcium/calmodulin-regulated serine/threonine kinase. This target-level refinement does not imply that every original assay independently tested calcium dependence or that autonomous phosphorylated CAMK2A requires ongoing calcium binding.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: Retain the curated BCL10 interaction in its kinase/substrate context.

Reason: O95999 is the BCL10 substrate discussed in PMID:17052756. Its complete abstract reports BCL10 Ser138 phosphorylation, inhibitor/siRNA effects and prolonged NF-kappaB activation with S138A. The original interaction experiment and alpha-specific reagent details were not accessible; defer to the curated IPI association without transferring BCL10 scaffold activity or claiming a verified endogenous alpha interface.

Summary: Postsynaptic-density association is a central location of neuronal CAMK2A action.

Reason: The PAINT/localization mappings agree with the synaptic targeting literature and human CAMK2A variant studies. Rat postsynaptic-density complexes and receptor-dependent synaptic localization provide independent support. The PAINT PSD IBD at PTN002389358 was found in the cached family data; the complete tree/alignment was not reanalyzed.

Summary: Cytoplasmic localization is consistent with the conserved intracellular kinase.

Reason: The cached PAINT family includes the cytoplasm IBD at PTN000686485. CAMK2A acts as an intracellular kinase and translocates among cytoplasmic/synaptic pools. The diverse descendant set is not a reason to dismiss this broad, valid location, and a finer source-specific compartment is not inferred for every descendant.

Summary: Regulation of neuronal synaptic plasticity is supported by both phylogenetic curation and alpha-specific experiments.

Reason: The cached PTN002389358 IBD supports the IBA, and the literal mouse P11798 donor has an IMP record from PMID:17610578. Alpha-dependent synaptic signaling and receptor-targeting work independently corroborate the neuronal plasticity role. Mouse experiments are not relabeled as human neuronal assays, and the short donor list is not treated as weak evidence.

Summary: Retain regulation of surface-protein localization as a contextual consequence of target phosphorylation.

Reason: The cached PAINT record places GO:1903076 at PTN002389359 with rat experimental descendants. Receptor-trafficking contexts are compatible with CAMK2 activity, but the exact individual rat donor assays were not re-read. There is no demonstrated target-specific loss or clade error; retain the inherited process outside the primary synaptic catalytic core.

Summary: Neuronal projections and synaptic compartments are established sites of CAMK2A function.

Reason: Human alpha constructs and neuronal E183V studies support dendritic-spine targeting and synaptic action, consistent with the preserved UniProt localization assertions. The broad neuron-projection/synapse terms are valid at their source resolution. Target self-inclusion in the IBA is expected; the precise neuron-projection IBD row was not recovered in the cached PAINT excerpt and is not claimed independently verified.

Summary: Calmodulin binding is the activation mechanism of the kinase rather than an additional core function.

Reason: Human alpha-domain biochemistry and activation studies support CaM recognition. The PMID:20668654 CaM co-crystal specifically uses delta, so it is not recast as an alpha structure. Preserve the individual P0DP23/P0DP24/P0DP25 source assertions: their canonical calmodulin proteins share the relevant sequence, but this does not identify which CALM gene supplied an assay reagent. The cached CaM-binding IBD at PTN002389359 includes target evidence legitimately.

Summary: Calmodulin binding is the activation mechanism of the kinase rather than an additional core function.

Reason: Human alpha-domain biochemistry and activation studies support CaM recognition. The PMID:20668654 CaM co-crystal specifically uses delta, so it is not recast as an alpha structure. Preserve the individual P0DP23/P0DP24/P0DP25 source assertions: their canonical calmodulin proteins share the relevant sequence, but this does not identify which CALM gene supplied an assay reagent. The cached CaM-binding IBD at PTN002389359 includes target evidence legitimately.

Summary: ATP binding supplies the kinase co-substrate.

Reason: ATP binding supplies the phosphate donor for CAMK2A-catalyzed phosphorylation. The human alpha kinase-domain structures in PMID:20668654 contain ATP-competitive inhibitors and are not direct ATP-bound structures; catalytic evidence independently supports ATP use. This valid binding annotation is subsumed by the catalytic core rather than treated as a separate biological activity.

Summary: CAMK2A functions within CaMKII homo- and heteromeric assemblies.

Reason: Human alpha structural studies and WT/E183V coassembly support CaMKII-complex membership. PMID:20668654 compares isolated kinase-domain self-association, while PMID:21884935 uses a specifically engineered alpha holoenzyme for crystallography; neither is evidence that every native holoenzyme has the same conformation or isoform composition. Human WT/mutant coassembly supplies independent target support.

Summary: Postsynaptic-density association is a central location of neuronal CAMK2A action.

Reason: The PAINT/localization mappings agree with the synaptic targeting literature and human CAMK2A variant studies. Rat postsynaptic-density complexes and receptor-dependent synaptic localization provide independent support. The PAINT PSD IBD at PTN002389358 was found in the cached family data; the complete tree/alignment was not reanalyzed.

Summary: Refine a correct broad catalytic mapping to the established calcium/calmodulin-dependent kinase activity.

Reason: The domain/keyword mapping correctly recognizes a protein phosphotransferase, but human CAMK2A has experimentally established calcium/calmodulin-regulated serine/threonine kinase activity. Use GO:0004683 for the target-specific catalytic activity; the original machine term and mapping source remain preserved.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: Refine a correct broad catalytic mapping to the established calcium/calmodulin-dependent kinase activity.

Reason: The domain/keyword mapping correctly recognizes a protein phosphotransferase, but human CAMK2A has experimentally established calcium/calmodulin-regulated serine/threonine kinase activity. Use GO:0004683 for the target-specific catalytic activity; the original machine term and mapping source remain preserved.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: Neuronal projections and synaptic compartments are established sites of CAMK2A function.

Reason: Human alpha constructs and neuronal E183V studies support dendritic-spine targeting and synaptic action, consistent with the preserved UniProt localization assertions. The broad neuron-projection/synapse terms are valid at their source resolution. Target self-inclusion in the IBA is expected; the precise neuron-projection IBD row was not recovered in the cached PAINT excerpt and is not claimed independently verified.

Summary: The precise endocytic-vesicle membrane assignment remains unresolved.

Reason: The Reactome event concerns AMPA-receptor trafficking, and CAMK2-dependent phosphorylation participates in that pathway. The available event summary does not independently localize the kinase itself to the endocytic-vesicle membrane; the exact ARBA rule premises were not inspected. This is a source-resolution gap, not evidence that the kinase cannot associate with that membrane.

Summary: Self-association is a supported assembly mechanism of the kinase.

Reason: Human alpha kinase domains self-associate in solution, and engineered alpha holoenzyme structures plus WT/E183V coassembly establish target self-interaction. Isolated-domain dimerization and full holoenzyme organization are distinguished; an engineered crystal conformation is not assumed universal. Retain the source dimer/identical-binding claims as assembly mechanisms within the catalytic core.

Summary: Neuronal projections and synaptic compartments are established sites of CAMK2A function.

Reason: Human alpha constructs and neuronal E183V studies support dendritic-spine targeting and synaptic action, consistent with the preserved UniProt localization assertions. The broad neuron-projection/synapse terms are valid at their source resolution. Target self-inclusion in the IBA is expected; the precise neuron-projection IBD row was not recovered in the cached PAINT excerpt and is not claimed independently verified.

Summary: Neuronal projections and synaptic compartments are established sites of CAMK2A function.

Reason: Human alpha constructs and neuronal E183V studies support dendritic-spine targeting and synaptic action, consistent with the preserved UniProt localization assertions. The broad neuron-projection/synapse terms are valid at their source resolution. Target self-inclusion in the IBA is expected; the precise neuron-projection IBD row was not recovered in the cached PAINT excerpt and is not claimed independently verified.

Summary: Metal binding supports phosphotransfer chemistry.

Reason: The canonical record assigns magnesium as catalytic cofactor and the kinase structures/assays are compatible with metal-assisted ATP use. The broad metal-binding assertion is valid mechanistic information, not evidence that CAMK2A is itself the calcium sensor or a distinct metal-handling core.

Summary: Long-term synaptic potentiation is a principal process served by the kinase.

Reason: The original structural paper discusses LTP rather than directly measuring it, but independent neuronal alpha-kinase and GluN2B/CaMKII-targeting studies support synaptic substrate phosphorylation and potentiation. The kinase performs phosphotransfer and target association within the process, not merely serving as a substrate consumed by it. Behavioral memory claims are kept separate.

Summary: Serine phosphorylation is one valid component of the broader kinase activity.

Reason: The Rhea mapping records serine-directed phosphotransfer, consistent with CAMK2A substrate phosphorylation. Retain this chemistry-specific assertion without confusing it with an exclusive serine specificity: the enzyme also phosphorylates threonine, including its own regulatory site. The complete catalytic activity is represented in the core.

Summary: Refine the supported CAMK2B interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13554 (CAMK2B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. BioPlex compares affinity-purification interaction networks across 293T and HCT116 cellular contexts. Available framework text and the curated target edge support association; the individual target table was not inspected, and direct physical contact or disease participation is not inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with YWHAE in its tested interaction context.

Reason: The source records UniProtKB:P62258 (YWHAE), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes the 14-3-3 interaction network. The exact target-pair experiment was not inspected. This supports the curated association without assigning the partner's chaperone-like activity to CAMK2A.

Summary: Self-association is a supported assembly mechanism of the kinase.

Reason: Human alpha kinase domains self-associate in solution, and engineered alpha holoenzyme structures plus WT/E183V coassembly establish target self-interaction. Isolated-domain dimerization and full holoenzyme organization are distinguished; an engineered crystal conformation is not assumed universal. Retain the source dimer/identical-binding claims as assembly mechanisms within the catalytic core.

Summary: Self-association is a supported assembly mechanism of the kinase.

Reason: Human alpha kinase domains self-associate in solution, and engineered alpha holoenzyme structures plus WT/E183V coassembly establish target self-interaction. Isolated-domain dimerization and full holoenzyme organization are distinguished; an engineered crystal conformation is not assumed universal. Retain the source dimer/identical-binding claims as assembly mechanisms within the catalytic core.

Summary: Use the established calcium/calmodulin-dependent kinase activity for this target.

Reason: The broad kinase assignment is compatible with CAMK2A. Human alpha structural/biochemical and E183V functional data establish the more specific calcium/calmodulin-regulated serine/threonine kinase. This target-level refinement does not imply that every original assay independently tested calcium dependence or that autonomous phosphorylated CAMK2A requires ongoing calcium binding.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: Retain the interferon/ferroptosis context with its literal process relation.

Reason: PMID:40281343 reports interferon-gamma/CaMKII-dependent PSAT1 Ser337 phosphorylation upstream of GPX4 stabilization and resistance to ferroptosis. Cached GO-CAM 680ad14200007201 explicitly assigns human Q9UQM7 catalytic activity to negative regulation of ferroptosis. The complete abstract was read but detailed isoform-specific Methods were not available. Preserve both distinct acts_upstream_of and involved_in source records without collapsing them; neither transfers PSAT1 catalysis, PHD3 hydroxylation or GPX4 peroxidase activity to CAMK2A.

Summary: Retain the interferon/ferroptosis context with its literal process relation.

Reason: PMID:40281343 reports interferon-gamma/CaMKII-dependent PSAT1 Ser337 phosphorylation upstream of GPX4 stabilization and resistance to ferroptosis. Cached GO-CAM 680ad14200007201 explicitly assigns human Q9UQM7 catalytic activity to negative regulation of ferroptosis. The complete abstract was read but detailed isoform-specific Methods were not available. Preserve both distinct acts_upstream_of and involved_in source records without collapsing them; neither transfers PSAT1 catalysis, PHD3 hydroxylation or GPX4 peroxidase activity to CAMK2A.

Summary: Long-term synaptic potentiation is a principal process served by the kinase.

Reason: The original structural paper discusses LTP rather than directly measuring it, but independent neuronal alpha-kinase and GluN2B/CaMKII-targeting studies support synaptic substrate phosphorylation and potentiation. The kinase performs phosphotransfer and target association within the process, not merely serving as a substrate consumed by it. Behavioral memory claims are kept separate.

Summary: Activated alpha kinase can support interferon-beta/JAK-STAT signaling in the tested system.

Reason: The source explicitly uses a CaMKII-alpha T286D/T305A/T306A construct that rescues interferon-response inhibition, alongside calcium/CaM and kinase-inhibitor experiments. GO-CAM 62b4ffe300004525 assigns the target within this pathway. Retain this experimentally supported contextual role while distinguishing engineered rescue from unperturbed alpha-specific signaling in all cells.

Summary: Activated alpha kinase can support interferon-beta/JAK-STAT signaling in the tested system.

Reason: The source explicitly uses a CaMKII-alpha T286D/T305A/T306A construct that rescues interferon-response inhibition, alongside calcium/CaM and kinase-inhibitor experiments. GO-CAM 62b4ffe300004525 assigns the target within this pathway. Retain this experimentally supported contextual role while distinguishing engineered rescue from unperturbed alpha-specific signaling in all cells.

Summary: Retain the curated interferon/STAT1 context with an explicit isoform boundary.

Reason: PMID:11972023 provides cellular gamma/delta STAT1 interaction and signaling experiments and an alpha/beta rat-brain kinase control. The cached GO-CAM 62b4ffe300004489 explicitly assigns human Q9UQM7 activity within interferon-gamma/JAK-STAT signaling. Defer to that curated target attribution while distinguishing it from independent proof of endogenous human-alpha signaling. This contextual role is not a separate core. The historical retired source record, where present, remains retired.

Summary: Mouse alpha kinase directly inhibits DAGLA and modulates endocannabinoid signaling.

Reason: The literal mouse P11798 donor is supported by PMID:23502535: striatal alpha/DAGLA coassociation, purified alpha-kinase phosphorylation of the DAGLA C terminus, phosphorylation-site analysis and altered 2-AG signaling. CAMK2A performs kinase work that regulates the lipase; it does not acquire DAGLA hydrolase activity. Human transfer is retained with the rodent and phosphomimetic-assay limits explicit.

Summary: Mouse alpha kinase directly inhibits DAGLA and modulates endocannabinoid signaling.

Reason: The literal mouse P11798 donor is supported by PMID:23502535: striatal alpha/DAGLA coassociation, purified alpha-kinase phosphorylation of the DAGLA C terminus, phosphorylation-site analysis and altered 2-AG signaling. CAMK2A performs kinase work that regulates the lipase; it does not acquire DAGLA hydrolase activity. Human transfer is retained with the rodent and phosphomimetic-assay limits explicit.

Summary: CAMK2A functions within CaMKII homo- and heteromeric assemblies.

Reason: Human alpha structural studies and WT/E183V coassembly support CaMKII-complex membership. PMID:20668654 compares isolated kinase-domain self-association, while PMID:21884935 uses a specifically engineered alpha holoenzyme for crystallography; neither is evidence that every native holoenzyme has the same conformation or isoform composition. Human WT/mutant coassembly supplies independent target support.

Summary: CAMK2A functions within CaMKII homo- and heteromeric assemblies.

Reason: Human alpha structural studies and WT/E183V coassembly support CaMKII-complex membership. PMID:20668654 compares isolated kinase-domain self-association, while PMID:21884935 uses a specifically engineered alpha holoenzyme for crystallography; neither is evidence that every native holoenzyme has the same conformation or isoform composition. Human WT/mutant coassembly supplies independent target support.

Summary: Human alpha constructs target dendritic spines.

Reason: PMID:28130356 compares wild-type human alpha with E183V and describes spine targeting and synaptic effects in neuronal models. The available cached abstract/Discussion supports the location, although detailed Methods and original imaging panels were not accessible. This is target localization evidence, not merely the compartment of a phosphorylated substrate.

Summary: Catalytic and interaction defects in human CAMK2A affect dendritic-spine development.

Reason: The E183V study reports reduced kinase activity, WT/mutant holoassembly and altered spine density in neuronal expression models. These findings connect CAMK2A catalytic/interaction work with spine development. Mutant overexpression and model context are preserved; this is not presented as direct imaging of a human patient's neurons or as a universal gain/loss-of-function rule.

Summary: Regulatory threonine autophosphorylation is part of CAMK2A activation.

Reason: Adjacent activated alpha subunits phosphorylate Thr286, generating autonomous activity; E183V impairs regulatory autophosphorylation. This is phosphotransfer performed by the kinase on another subunit of its assembly, not an inferred process based only on being a phosphorylation substrate.

Summary: Retain the mutation-dependent neuronal-migration phenotype without making it a universal wild-type role.

Reason: PMID:29100089 identifies human CAMK2A/CAMK2B variants and tests neuronal migration in a developmental model. Its Discussion distinguishes alpha mutant effects from the finding that wild-type CAMK2A overexpression or knockdown did not change migration at the examined stage. The existing IMP is retained for this contextual perturbation evidence; beta dosage and alpha gain-of-function constructs are not conflated.

Summary: The organelle claim for CAMK2A rests on a cardiac CaMKII signal that was not assigned to the alpha gene product; CAMK2A has no mitochondrial targeting feature and its curated location is postsynaptic.

Reason: The mouse P11798 source joins lead to PMID:23051746, which studies cardiac CaMKII inhibition, mitochondrial preparations and active T287D constructs. The inspected Methods do not establish an alpha-specific endogenous cardiac reagent for this claim. Moreover, the reported direct MCU-current mechanism was challenged in a subsequent primary communication (DOI:10.1038/nature13626); the published reply was not accessed. Examined as a synthesis row (ORGANELLES batch 2, 2026-10-24). QuickGO shows the donor's only mitochondrion annotation is a BHF-UCL IDA to PMID:23051746 on mouse Camk2a; mouse Camk2b, Camk2d and Camk2g carry none. In the full text, the mitochondrial blots and immuno-EM use an antibody listed only as "CaMKII", in heart, and the paper never names the alpha subunit, so the donor IDA assigns a family-level signal to one paralog. The Percoll-purified fractions were probed with ER markers (SERCA2, calnexin), so this is not a simple MAM artifact, but the isoform question is unresolved at the source. On the human side, CAMK2A has no transit peptide or TM segment, and UniProt's curated location is synapse, postsynaptic density, dendritic spine and dendrite, with no mitochondrion. A soluble signalling kinase transferred from a heart preparation to a neuronal paralog is the over-propagation pattern; the row is marked over-annotated rather than removed because a CaMKII pool in heart mitochondria was reported and is disputed, not refuted. No sub-compartment is proposed.

Summary: Calmodulin binding is the activation mechanism of the kinase rather than an additional core function.

Reason: Human alpha-domain biochemistry and activation studies support CaM recognition. The PMID:20668654 CaM co-crystal specifically uses delta, so it is not recast as an alpha structure. Preserve the individual P0DP23/P0DP24/P0DP25 source assertions: their canonical calmodulin proteins share the relevant sequence, but this does not identify which CALM gene supplied an assay reagent. The cached CaM-binding IBD at PTN002389359 includes target evidence legitimately.

Summary: Retain the curated angiotensin/NHE3 signaling context with source-specific limits.

Reason: PMID:20584908 supports calcium/CaMKII-dependent IRBIT/NHE3 association and surface trafficking in a proximal-tubule cell model. The available Discussion explicitly says unequivocal IRBIT phosphorylation was not established; alpha-specific Methods are absent from the cache. Defer to the existing target annotation for pathway participation while avoiding a direct IRBIT-substrate or endogenous human-alpha-specific claim.

Summary: The precise alpha-specific cell-cycle/calcium-transport transfer is not resolved by the accessible donor paper.

Reason: The mouse donor annotations cite PMID:12660151. Its complete abstract reports PMCA4, calcium handling, CaMKII inhibition and G1/S control in mouse vascular smooth-muscle cells, but does not identify the target isoform or provide the relevant alpha-specific experiment. The original full assay was not accessible. Preserve the source and use uncertainty rather than removing it for being a non-neuronal context.

Summary: Human-alpha-specific ischemic stress response remains unresolved in the donor evidence.

Reason: The mouse P11798 source joins lead to PMID:23051746, which studies cardiac CaMKII inhibition, mitochondrial preparations and active T287D constructs. The inspected Methods do not establish an alpha-specific endogenous cardiac reagent for this claim. Moreover, the reported direct MCU-current mechanism was challenged in a subsequent primary communication (DOI:10.1038/nature13626); the published reply was not accessed. This does not refute every mitochondrial or injury phenotype, but the precise human CAMK2A ischemic stress response transfer cannot be finalized from the available evidence.

Summary: The precise alpha-specific cell-cycle/calcium-transport transfer is not resolved by the accessible donor paper.

Reason: The mouse donor annotations cite PMID:12660151. Its complete abstract reports PMCA4, calcium handling, CaMKII inhibition and G1/S control in mouse vascular smooth-muscle cells, but does not identify the target isoform or provide the relevant alpha-specific experiment. The original full assay was not accessible. Preserve the source and use uncertainty rather than removing it for being a non-neuronal context.

Summary: Human-alpha-specific cardiac apoptotic regulation remains unresolved in the donor evidence.

Reason: The mouse P11798 source joins lead to PMID:23051746, which studies cardiac CaMKII inhibition, mitochondrial preparations and active T287D constructs. The inspected Methods do not establish an alpha-specific endogenous cardiac reagent for this claim. Moreover, the reported direct MCU-current mechanism was challenged in a subsequent primary communication (DOI:10.1038/nature13626); the published reply was not accessed. This does not refute every mitochondrial or injury phenotype, but the precise human CAMK2A cardiac apoptotic regulation transfer cannot be finalized from the available evidence.

Summary: The traced mouse interaction concerns an ionotropic NMDA receptor.

Reason: The donor P11798 binding assertion joins to PMID:16710293 with mouse GRIN2B. That abstract supports receptor/CaMKII signaling but does not itself reproduce the binding assay. Independent direct alpha/NR2B interaction experiments in PMID:19453375 support the more specific ionotropic glutamate receptor binding term, while preserving the original orthology evidence and access distinction.

Proposed replacements: ionotropic glutamate receptor binding

Summary: Regulation of neuronal synaptic plasticity is supported by both phylogenetic curation and alpha-specific experiments.

Reason: The cached PTN002389358 IBD supports the IBA, and the literal mouse P11798 donor has an IMP record from PMID:17610578. Alpha-dependent synaptic signaling and receptor-targeting work independently corroborate the neuronal plasticity role. Mouse experiments are not relabeled as human neuronal assays, and the short donor list is not treated as weak evidence.

Summary: Human-alpha-specific positive calcium-transport regulation remains unresolved in the donor evidence.

Reason: The mouse P11798 source joins lead to PMID:23051746, which studies cardiac CaMKII inhibition, mitochondrial preparations and active T287D constructs. The inspected Methods do not establish an alpha-specific endogenous cardiac reagent for this claim. Moreover, the reported direct MCU-current mechanism was challenged in a subsequent primary communication (DOI:10.1038/nature13626); the published reply was not accessed. This does not refute every mitochondrial or injury phenotype, but the precise human CAMK2A positive calcium-transport regulation transfer cannot be finalized from the available evidence.

Summary: Human-alpha-specific mitochondrial permeability regulation remains unresolved in the donor evidence.

Reason: The mouse P11798 source joins lead to PMID:23051746, which studies cardiac CaMKII inhibition, mitochondrial preparations and active T287D constructs. The inspected Methods do not establish an alpha-specific endogenous cardiac reagent for this claim. Moreover, the reported direct MCU-current mechanism was challenged in a subsequent primary communication (DOI:10.1038/nature13626); the published reply was not accessed. This does not refute every mitochondrial or injury phenotype, but the precise human CAMK2A mitochondrial permeability regulation transfer cannot be finalized from the available evidence.

Summary: Presynaptic alpha kinase regulates activity-dependent transmitter release in mouse hippocampus.

Reason: PMID:12629219 uses conditional presynaptic alpha-CaMKII deletion and distinguishes preserved basal transmission from altered activity-dependent release. PMID:17660813 further separates the alpha protein's presynaptic role from its kinase activation and synapsin phosphorylation. Retain this contextual regulatory function without equating all transmitter secretion with catalytic activity or assigning a postsynaptic scaffold on this evidence.

Summary: Self-association is a supported assembly mechanism of the kinase.

Reason: Human alpha kinase domains self-associate in solution, and engineered alpha holoenzyme structures plus WT/E183V coassembly establish target self-interaction. Isolated-domain dimerization and full holoenzyme organization are distinguished; an engineered crystal conformation is not assumed universal. Retain the source dimer/identical-binding claims as assembly mechanisms within the catalytic core.

Summary: Retain the curated nuclear detection in the human sperm proteome.

Reason: PMID:21630459 analyzes highly purified human sperm nuclei by mass spectrometry. Its complete abstract and target-specific HDA curation support a specialized nuclear detection, although the individual CAMK2A peptide table was not inspected. This does not establish general neuronal nuclear function or finer nucleoplasmic localization.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-432164 cached summary was read. The event describes calcium entry rather than catalysis by CAMK2A. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-4332356 cached summary was read. The event reports CAMK2/MAP3K7 coassociation. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-4332358 cached summary was read. The event distinguishes CaM dissociation and inhibitory-site autophosphorylation. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-4332359 cached summary was read. The event describes calcium/CaM activation of homo- or heteromeric CaMKII. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-4332363 cached summary was read. The event describes intersubunit regulatory autophosphorylation. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-4332388 cached summary was read. The event notes that direct purified CAMK2 phosphorylation of MAP3K7 was not demonstrated. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-5672965 cached summary was read. RAS nucleotide exchange is performed by the GEF, not by the kinase. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: The target-specific nucleoplasmic location is not independently resolved.

Reason: The HSF1 Ser230 phosphorylation event is nucleoplasmic, but the cached Reactome summary focuses on HSF1 and CaMKII-dependent transactivation rather than imaging endogenous CAMK2A in that compartment. The human sperm nucleus HDA assertion is broader and a different cell context. Preserve the source while leaving the finer target-localization claim open.

Summary: Retain the source cytosolic location of the intracellular kinase.

Reason: The complete Reactome R-HSA-909552 cached summary was read. STAT1 phosphorylation does not by itself relocate all upstream kinase molecules to the nucleus. Its cytosolic assignment is compatible with independently established intracellular CAMK2A pools. The event's downstream chemistry is not transferred to CAMK2A.

Summary: The precise endocytic-vesicle membrane assignment remains unresolved.

Reason: The Reactome event concerns AMPA-receptor trafficking, and CAMK2-dependent phosphorylation participates in that pathway. The available event summary does not independently localize the kinase itself to the endocytic-vesicle membrane; the exact ARBA rule premises were not inspected. This is a source-resolution gap, not evidence that the kinase cannot associate with that membrane.

Summary: The kinase directly carries out protein phosphorylation in the BCL10 signaling context.

Reason: The complete PMID:17052756 abstract reports CaMKII-dependent BCL10 Ser138 phosphorylation and kinase-inhibitor/siRNA experiments. This is catalytic participation, not merely necessity. The original alpha-specific construct details remain uninspected; retain the target curation with that limit and keep the immune signaling context outside the primary neuronal core.

Summary: Use the established calcium/calmodulin-dependent kinase activity for this target.

Reason: The broad kinase assignment is compatible with CAMK2A. Human alpha structural/biochemical and E183V functional data establish the more specific calcium/calmodulin-regulated serine/threonine kinase. This target-level refinement does not imply that every original assay independently tested calcium dependence or that autonomous phosphorylated CAMK2A requires ongoing calcium binding.

Proposed replacements: calcium/calmodulin-dependent protein kinase activity

Summary: The obsolete positive-direction term and the source's opposing regulatory steps need reconciliation.

Reason: The machine GO:0051092 identifier/label are preserved. The abstract discusses positive CARMA1-dependent signaling but reports that preventing BCL10 Ser138 phosphorylation prolongs NF-kappaB activation, consistent with an attenuating step. Without the exact original experiment supporting this positive annotation, a sign-specific replacement would conflate distinct substrates and steps. Do not silently replace the obsolete source term with a guessed current identifier.

Summary: Retain the curated CaMKII-associated synaptic-protein interaction.

Reason: The exact partner-specific IPI assertion from PMID:28130356 is preserved. Its available abstract/Discussion supports altered CaMKII-associated synaptic complexes, but does not reproduce this individual partner assay. Defer to target-specific curation without claiming a direct interface, transferring the partner function, or rejecting the edge because the title emphasizes a variant.

Summary: The literal partner GRIN2B supports ionotropic glutamate receptor binding.

Reason: The Q13224 IPI assertion is retained from the CAMK2A variant study. Its individual assay was not recovered in the partial cache, but independent direct alpha/NR2B binding experiments in PMID:19453375 and the exact canonical UniProt interaction corroborate this partner. Refine to the defined receptor-binding class without implying kinase channel activity.

Proposed replacements: ionotropic glutamate receptor binding

Summary: Retain the curated CaMKII-associated synaptic-protein interaction.

Reason: The exact partner-specific IPI assertion from PMID:28130356 is preserved. Its available abstract/Discussion supports altered CaMKII-associated synaptic complexes, but does not reproduce this individual partner assay. Defer to target-specific curation without claiming a direct interface, transferring the partner function, or rejecting the edge because the title emphasizes a variant.

Summary: Human CAMK2A coassociates with SHANK3 in the tested expression system.

Reason: The source Discussion reports disruption of SHANK3 coimmunoprecipitation by E183V despite retained WT/mutant holoassembly. Retain the literal Q9BYB0 IPI association with its expression-system/access limits. Coimmunoprecipitation supports coassociation; it does not by itself make CAMK2A a general scaffold or prove a direct binary interface.

Summary: Refine the supported HSP90AB1 interaction to Hsp90 protein binding.

Reason: The source records UniProtKB:P08238 (HSP90AB1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The study compares affinity-purified signaling networks in HCT116 and HKE3 cells with quantitative mass spectrometry. The exact pairwise target entry was not inspected; retain the curated association without transferring colorectal-cancer or KRAS mechanisms to CAMK2A. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: Hsp90 protein binding

Summary: Refine the supported CAMK2D interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13557 (CAMK2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The study compares affinity-purified signaling networks in HCT116 and HKE3 cells with quantitative mass spectrometry. The exact pairwise target entry was not inspected; retain the curated association without transferring colorectal-cancer or KRAS mechanisms to CAMK2A. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with CDC37 in its tested interaction context.

Reason: The source records UniProtKB:Q16543 (CDC37), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The study compares affinity-purified signaling networks in HCT116 and HKE3 cells with quantitative mass spectrometry. The exact pairwise target entry was not inspected; retain the curated association without transferring colorectal-cancer or KRAS mechanisms to CAMK2A.

Summary: Retain the curated association with RBM47 in its tested interaction context.

Reason: The source records UniProtKB:A0AV96 (RBM47), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP23-1 in its tested interaction context.

Reason: The source records UniProtKB:A1A580 (KRTAP23-1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with FAM168B in its tested interaction context.

Reason: The source records UniProtKB:A1KXE4-2 (FAM168B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with RBFOX2 in its tested interaction context.

Reason: The source records UniProtKB:O43251-10 (RBFOX2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with HYAL3 in its tested interaction context.

Reason: The source records UniProtKB:O43820 (HYAL3), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRT75 in its tested interaction context.

Reason: The source records UniProtKB:O95678 (KRT75), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with VARS1 in its tested interaction context.

Reason: The source records UniProtKB:P26640 (VARS1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with SOX5 in its tested interaction context.

Reason: The source records UniProtKB:P35711-4 (SOX5), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with SUOX in its tested interaction context.

Reason: The source records UniProtKB:P51687 (SUOX), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TIAL1 in its tested interaction context.

Reason: The source records UniProtKB:Q01085-2 (TIAL1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRT76 in its tested interaction context.

Reason: The source records UniProtKB:Q01546 (KRT76), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with ARID5A in its tested interaction context.

Reason: The source records UniProtKB:Q03989 (ARID5A), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Refine the supported CAMK2B interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13554-3 (CAMK2B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported CAMK2G interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13555-5 (CAMK2G), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported CAMK2D interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13557-8 (CAMK2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with LASP1 in its tested interaction context.

Reason: The source records UniProtKB:Q14847-2 (LASP1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with DAZAP2 in its tested interaction context.

Reason: The source records UniProtKB:Q15038 (DAZAP2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with CDC37 in its tested interaction context.

Reason: The source records UniProtKB:Q16543 (CDC37), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP6-1 in its tested interaction context.

Reason: The source records UniProtKB:Q3LI64 (KRTAP6-1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP6-2 in its tested interaction context.

Reason: The source records UniProtKB:Q3LI66 (KRTAP6-2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP6-3 in its tested interaction context.

Reason: The source records UniProtKB:Q3LI67 (KRTAP6-3), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP19-5 in its tested interaction context.

Reason: The source records UniProtKB:Q3LI72 (KRTAP19-5), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP15-1 in its tested interaction context.

Reason: The source records UniProtKB:Q3LI76 (KRTAP15-1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP19-7 in its tested interaction context.

Reason: The source records UniProtKB:Q3SYF9 (KRTAP19-7), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with C1orf94 in its tested interaction context.

Reason: The source records UniProtKB:Q6P1W5 (C1orf94), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with RBPMS2 in its tested interaction context.

Reason: The source records UniProtKB:Q6ZRY4 (RBPMS2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP19-3 in its tested interaction context.

Reason: The source records UniProtKB:Q7Z4W3 (KRTAP19-3), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TFAP2D in its tested interaction context.

Reason: The source records UniProtKB:Q7Z6R9 (TFAP2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with RALYL in its tested interaction context.

Reason: The source records UniProtKB:Q86SE5-3 (RALYL), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with KRTAP8-1 in its tested interaction context.

Reason: The source records UniProtKB:Q8IUC2 (KRTAP8-1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with GLB1L2 in its tested interaction context.

Reason: The source records UniProtKB:Q8IW92 (GLB1L2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TTC5 in its tested interaction context.

Reason: The source records UniProtKB:Q8N0Z6 (TTC5), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with FAM168A in its tested interaction context.

Reason: The source records UniProtKB:Q92567-2 (FAM168A), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with RBPMS in its tested interaction context.

Reason: The source records UniProtKB:Q93062-3 (RBPMS), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TSR2 in its tested interaction context.

Reason: The source records UniProtKB:Q969E8 (TSR2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with NTAQ1 in its tested interaction context.

Reason: The source records UniProtKB:Q96HA8 (NTAQ1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with SPMIP9 in its tested interaction context.

Reason: The source records UniProtKB:Q96LM6 (SPMIP9), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with LENG8 in its tested interaction context.

Reason: The source records UniProtKB:Q96PV6 (LENG8), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with RHOXF2 in its tested interaction context.

Reason: The source records UniProtKB:Q9BQY4 (RHOXF2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TAB2 in its tested interaction context.

Reason: The source records UniProtKB:Q9NYJ8 (TAB2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with MRPL11 in its tested interaction context.

Reason: The source records UniProtKB:Q9Y3B7 (MRPL11), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Retain the curated association with TCAF1 in its tested interaction context.

Reason: The source records UniProtKB:Q9Y4C2-2 (TCAF1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The HuRI framework tests binary interactions using multiple yeast two-hybrid versions. The exact target-pair table was not independently inspected. A curated positive edge is not rejected for lacking a native-cell assay, but no endogenous physiological role or binding-domain mechanism is inferred.

Summary: Refine the supported CAMK2D interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13557 (CAMK2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The kinase interaction network uses affinity purification/mass spectrometry and controlled network filtering. The individual target-pair record was not inspected. Co-purification supports association without proving a direct interface or transferring the partner kinase activity. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with TOM1 in its tested interaction context.

Reason: The source records UniProtKB:O60784-2 (TOM1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with PNP in its tested interaction context.

Reason: The source records UniProtKB:P00491 (PNP), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with S100B in its tested interaction context.

Reason: The source records UniProtKB:P04271 (S100B), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with GAPDH in its tested interaction context.

Reason: The source records UniProtKB:P04406 (GAPDH), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with APP in its tested interaction context.

Reason: The source records UniProtKB:P05067 (APP), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Refine the supported PRKCB interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:P05771-2 (PRKCB), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Refine the supported HSP90AA1 interaction to Hsp90 protein binding.

Reason: The source records UniProtKB:P07900 (HSP90AA1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: Hsp90 protein binding

Summary: Retain the curated association with VIM in its tested interaction context.

Reason: The source records UniProtKB:P08670 (VIM), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with UNG in its tested interaction context.

Reason: The source records UniProtKB:P13051-2 (UNG), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with ATP1A3 in its tested interaction context.

Reason: The source records UniProtKB:P13637 (ATP1A3), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with PECAM1 in its tested interaction context.

Reason: The source records UniProtKB:P16284 (PECAM1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with RAB5A in its tested interaction context.

Reason: The source records UniProtKB:P20339 (RAB5A), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with BDNF in its tested interaction context.

Reason: The source records UniProtKB:P23560-2 (BDNF), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with PIK3R1 in its tested interaction context.

Reason: The source records UniProtKB:P27986-2 (PIK3R1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with SHC1 in its tested interaction context.

Reason: The source records UniProtKB:P29353-7 (SHC1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Refine the supported AKT1 interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:P31749 (AKT1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with CBS in its tested interaction context.

Reason: The source records UniProtKB:P35520 (CBS), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with NBL1 in its tested interaction context.

Reason: The source records UniProtKB:P41271-2 (NBL1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with UBE2K in its tested interaction context.

Reason: The source records UniProtKB:P61086 (UBE2K), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with ACTA1 in its tested interaction context.

Reason: The source records UniProtKB:P68133 (ACTA1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with SMAD3 in its tested interaction context.

Reason: The source records UniProtKB:P84022 (SMAD3), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with MAPK8IP2 in its tested interaction context.

Reason: The source records UniProtKB:Q13387-4 (MAPK8IP2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with APBB2 in its tested interaction context.

Reason: The source records UniProtKB:Q92870-2 (APBB2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with UTP14A in its tested interaction context.

Reason: The source records UniProtKB:Q9BVJ6 (UTP14A), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with VPS29 in its tested interaction context.

Reason: The source records UniProtKB:Q9UBQ0-2 (VPS29), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Retain the curated association with STUB1 in its tested interaction context.

Reason: The source records UniProtKB:Q9UNE7 (STUB1), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. The complete abstract describes a neurodegenerative-disease-focused yeast two-hybrid interaction map. The exact target-pair table was not inspected. Retain the curated interaction without inferring a disease mechanism or aggregating the preserved isoform-specific partners.

Summary: Refine the supported CAMK2D interaction to protein serine/threonine kinase binding.

Reason: The source records UniProtKB:Q13557 (CAMK2D), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. BioPlex compares affinity-purification interaction networks across 293T and HCT116 cellular contexts. Available framework text and the curated target edge support association; the individual target table was not inspected, and direct physical contact or disease participation is not inferred. The partner identity supports this defined binding class; the source accession and any isoform suffix remain unchanged.

Proposed replacements: protein serine/threonine kinase binding

Summary: Retain the curated association with TAB2 in its tested interaction context.

Reason: The source records UniProtKB:Q9NYJ8 (TAB2), and the same literal accession is independently present in the canonical UniProt IntAct interaction list. BioPlex compares affinity-purification interaction networks across 293T and HCT116 cellular contexts. Available framework text and the curated target edge support association; the individual target table was not inspected, and direct physical contact or disease participation is not inferred.

Summary: Calmodulin binding is the activation mechanism of the kinase rather than an additional core function.

Reason: Human alpha-domain biochemistry and activation studies support CaM recognition. The PMID:20668654 CaM co-crystal specifically uses delta, so it is not recast as an alpha structure. Preserve the individual P0DP23/P0DP24/P0DP25 source assertions: their canonical calmodulin proteins share the relevant sequence, but this does not identify which CALM gene supplied an assay reagent. The cached CaM-binding IBD at PTN002389359 includes target evidence legitimately.

Summary: Calmodulin binding is the activation mechanism of the kinase rather than an additional core function.

Reason: Human alpha-domain biochemistry and activation studies support CaM recognition. The PMID:20668654 CaM co-crystal specifically uses delta, so it is not recast as an alpha structure. Preserve the individual P0DP23/P0DP24/P0DP25 source assertions: their canonical calmodulin proteins share the relevant sequence, but this does not identify which CALM gene supplied an assay reagent. The cached CaM-binding IBD at PTN002389359 includes target evidence legitimately.

Summary: Retain the interferon/ferroptosis context with its literal process relation.

Reason: PMID:40281343 reports interferon-gamma/CaMKII-dependent PSAT1 Ser337 phosphorylation upstream of GPX4 stabilization and resistance to ferroptosis. Cached GO-CAM 680ad14200007201 explicitly assigns human Q9UQM7 catalytic activity to negative regulation of ferroptosis. The complete abstract was read but detailed isoform-specific Methods were not available. Preserve both distinct acts_upstream_of and involved_in source records without collapsing them; neither transfers PSAT1 catalysis, PHD3 hydroxylation or GPX4 peroxidase activity to CAMK2A.