KIF24 curation notes

Deep research status

just deep-research-falcon human KIF24 was run (first with the unavailable perplexity-lite fallback, then --timeout 2400). See the end of this file for the outcome. The review was written from cached primary literature.

Kinesin-13-like protein that binds CP110/CEP97 and localizes to mother centrioles PMID:21620453
Loss causes CP110 loss from mother centrioles and aberrant ciliation in cycling cells PMID:21620453
Binds and depolymerizes microtubules; remodels centriolar MTs specifically PMID:21620453
NEK2 phosphorylation stimulates KIF24 and prevents cilium outgrowth in proliferating cells PMID:26290419
Intramolecular N/C-terminal interaction (autoinhibition) relieved by NEK2 PMID:26290419
Recruits MPHOSPH9, which recruits CP110-CEP97 PMID:30375385
UniProt notes KIF24 does not disassemble fully formed axonemes; so I did not propose a NEW cilium disassembly annotation (raised as a question instead).

Motor activity (IBA/IDA/IEA): ACCEPT with caveat (kinesin-13 depolymerase; GO lacks a depolymerase MF). Microtubule-based movement IEA: MARK_AS_OVER_ANNOTATED (not a translocating motor).
Microtubule depolymerization, microtubule binding, centrosome/centriole, negative regulation of cilium assembly: ACCEPT.
Cilium assembly IMP: MODIFY -> negative regulation of cilium assembly (direction of phenotype).
Identical protein binding: MARK_AS_OVER_ANNOTATED (intramolecular fold-back, not homo-oligomer). Protein binding x10: REMOVE.

Module: stage 1 cap component, "MT-depolymerizing kinesin at mother centriole (ciliogenesis suppressor)"; the cap participant notes KIF24 contributes MT-depolymerizing activity and lists negative regulation of cilium assembly.
HPA v25: no data ("not in HPA"); no HPA-sourced GOA rows. The HPA cilium atlas therefore provides no evidence for or against.
Assessment: core_functions agree with the module (microtubule binding/depolymerization and negative regulation of cilium assembly at the mother centriole). Minor point: KIF24 is better described as the upstream recruiter of the cap (via MPHOSPH9) plus a depolymerase, rather than a structural "cap component" in the same sense as CP110/CEP97.

Falcon succeeded on retry (--timeout 2400): KIF24-deep-research-falcon.md. Consistent with the review. Two refinements taken on board: (1) the 2011 KIF24 motor fragment depolymerized microtubules similarly with ATP and AMP-PNP, so ATP-hydrolysis dependence is not demonstrated for KIF24 (the description and motor-activity review were worded accordingly); (2) biallelic KIF24 missense variants cause a skeletal ciliopathy/acromesomelic dysplasia (Reilly et al. 2022, JBMR, DOI 10.1002/jbmr.4639), with patient cells paradoxically showing fewer cilia and centriole amplification - noted in the description, not used for annotations. The deep research also supports NEK2-KIF24 mainly maintaining the disassembled state rather than disassembling full axonemes, consistent with not proposing a cilium disassembly annotation.