Skip to content

Grouping Review: Tubulinopathies (2026-08-20)

Review of kb/groupings/Tubulinopathies.yaml — grouping structure and MONDO alignment, completeness of the curated member set, per-member content review, currency against the 2024–2026 literature, and the scientific knowledge gaps that literature opens.

Scope

Member entry Lines Compliance Evidence items Distinct refs Newest ref
TUBA1A-related_Tubulinopathy 645 78.9% 23 5 2017
TUBB2A_TUBB2B-related_Cortical_Malformation 1423 93.2% 71 8 2021
TUBB3-related_Tubulinopathy 1330 95.7% 64 8 2023
TUBB_TUBB5-related_Microcephaly 698 79.3% 22 4 2017
Total 4,096 — 180 21 —

Grouping: grouping_basis: [SHARED_GENE_FAMILY, SHARED_MECHANISM], one NECESSARY criteria block whose logic is a single CONFORMS_TO_MODULE leaf on microtubule_dependent_neuronal_migration_failure#Microtubule Apparatus Perturbation.

Method

  1. Structural + membership audit — just check-groupings, just validate-grouping, just compliance per member, section-coverage matrix across the four entries.
  2. MONDO alignment — EBI OLS4 API: descendants of MONDO:0100153 (tubulinopathy), hierarchical ancestors of each member's disease_term, and definitions of the whole CDCBM (complex cortical dysplasia with other brain malformations) series to recover each type's causal gene.
  3. Literature sweep — NCBI E-utilities over tubulinopath*, each member gene, and TUBG1/TUBA8, restricted to 2023–2026, with abstracts retrieved for every paper cited below. All PMIDs and titles in this report come from PubMed esummary/efetch responses, not from recall.

1. Grouping structure — sound

check-groupings reports structure: OK and all four listed members SATISFIED against the NECESSARY criterion. The design decisions are defensible and well argued in the file:

  • NECESSARY, not NECESSARY_AND_SUFFICIENT, is the right semantics. The grouping_rationale correctly observes that microtubule-dependent migration failure also arises from non-tubulin genes (LIS1/PAFAH1B1, DCX, and — from the module — KIF5C, KIF2A, DYNC1H1), so the mechanism cannot be sufficient. MONDO agrees: CDCBM2 is KIF5C and CDCBM3 is KIF2A, both non-tubulin.
  • Declining a HAS_GENE leaf is justified. With ~27 human tubulin genes, enumerating the isotypes in the logic would be brittle, and under audit-only NECESSARY semantics it would buy nothing the curated member list does not already convey.
  • Keeping the four as separate Disease entries, not one lump, is correct and is now better supported than when the grouping was written — see §5F on isotype-specific C-terminal tails and expression timing.

One structural observation, not a defect: the criterion tests the migration module's trigger node, so any tubulin disease that is not a cortical-malformation disorder fails it automatically. That makes criterion and intended scope agree — but the scope itself was stated only in display_name. Fixed in this review (§7).

2. MONDO alignment — the mapping is right, the upstream hierarchy is not

The skos:exactMatch to MONDO:0100153 is appropriate. The interesting finding is that MONDO's own tubulinopathy subtree is internally inconsistent, which is what made the grouping's member/MONDO alignment look worse than it is.

MONDO:0100153 has exactly five is-a descendants:

CURIE Label Gene
MONDO:0010912 fibrosis of extraocular muscles, congenital, 3A TUBB3
MONDO:0014337 complex cortical dysplasia with other brain malformations 5 TUBB2A
MONDO:0018763 tubulinopathy-associated dysgyria (tubulin, unspecified)
MONDO:0100144 Uner Tan Syndrome TUBB2B p.Arg390Gln
MONDO:0100154 TUBB3-related tubulinopathy TUBB3

But the CDCBM series is definitionally tubulin-caused for five of its types, and only one of those five is linked to tubulinopathy:

Term Type Gene (from MONDO definition) is-a descendant of MONDO:0100153?
MONDO:0013541 CDCBM1 TUBB3 no
MONDO:0014116 CDCBM2 KIF5C n/a (not a tubulin)
MONDO:0014170 CDCBM3 KIF2A n/a (not a tubulin)
MONDO:0014171 CDCBM4 TUBG1 no
MONDO:0014337 CDCBM5 TUBB2A yes
MONDO:0014341 CDCBM6 TUBB no
MONDO:0012399 CDCBM7 TUBB2B no

So CDCBM5 is classified as a tubulinopathy and CDCBM1/4/6/7 are not, despite all five carrying a tubulin gene in their own definition text. Similarly MONDO:0012703 (lissencephaly due to TUBA1A mutation) sits under lissencephaly type 3, not under tubulinopathy.

Conclusion: the "2/4 members are descendants" signal is an upstream MONDO gap, not a DisMech membership error. This is worth an upstream MONDO issue — adding is-a tubulinopathy to CDCBM1/4/6/7 and to MONDO:0012703 would make the subtree complete and would make the grouping's exactMatch verifiable by reasoning rather than by curator assertion.

Uncurated MONDO tubulin concepts (DisMech curation gaps)

CURIE Label Status in DisMech
MONDO:0014171 CDCBM4 (TUBG1) no entry at all — see §3
MONDO:0100144 Uner Tan Syndrome (TUBB2B R390Q) not covered by any entry or subtype
MONDO:0014337 CDCBM5 (TUBB2A) subsumed by the TUBB2A/2B entry, but its TUBB2A subtype carries no term: and the entry has no mappings: block
MONDO:0012399 CDCBM7 (TUBB2B) same
MONDO:0013541 CDCBM1 (TUBB3) subsumed by the TUBB3 entry (MONDO:0100154), no mapping recorded
MONDO:0010912 CFEOM3A (TUBB3) curated as a phenotype of the TUBB3 entry, not as a mapping

Per the CLAUDE.md MONDO-coverage rule, only disease_term and has_subtypes terms (and skos:exactMatch/narrowMatch mappings) retire a concept from the queue. None of the four members carries a mappings: block, so five of the six concepts above remain formally uncovered even where the biology is fully curated. Adding subtype term: values to the TUBB2A/2B subtypes and a mappings.mondo_mappings block to TUBB3 would close four of them cheaply.

3. Membership completeness

The one clear gap: TUBG1

TUBG1-related tubulinopathy (CDCBM4, MONDO:0014171) belongs in this grouping and has no DisMech entry. It passes both criteria on inspection — gamma-tubulin is a tubulin isotype gene, and the disease is a lissencephaly/microlissencephaly cortical malformation. The evidence for its in-scope status is already inside the KB: the TUBA1A member entry quotes PMID:24860126 as "The core phenotype of TUBA1A and TUBG1 tubulinopathies are lissencephalies and microlissencephalies", Lissencephaly_Spectrum_Disorders already carries TUBG1 as a lissencephaly gene with an hgnc: binding, and the TUBA1A entry's own notes explicitly says TUBA1A is "deliberately split from the beta-tubulin (TUBB2B/TUBB3/TUBB5) and gamma-tubulin (TUBG1) tubulinopathies".

It is also well-supported in the current literature, so it is a curatable entry today, not a stub:

PMID Year Contribution
PMID:40298439 2025 Two new cases expanding the phenotypic spectrum of TUBG1 missense variants
PMID:41070651 2025 Clinical phenotype and genetic analysis of a child with CDCBM4 and epilepsy
PMID:39215931 2025 tubg1 somatic-mutant zebrafish showing tubulinopathy-associated neurodevelopmental phenotypes — a ready animal_models entry
PMID:42177523 2026 Gene-specific long-term course, outcome and QoL including a TUBG1 lissencephaly patient
PMID:38919239 2024 Craniosynostosis with a novel TUBG1 variant — phenotype expansion
PMID:38912084 2024 TUBG1 lissencephaly and microcephaly case

Note PMID:37475831 ("A novel TUBG1 mutation with neurodevelopmental disorder…") is retracted and must not be cited.

Recommendation: file a claim issue for TUBG1-related tubulinopathy (CDCBM4, MONDO:0014171), curate the entry, and add it as a fifth member.

A contested candidate: TUBA8

TUBA8 is associated with autosomal-recessive polymicrogyria with optic nerve hypoplasia and appears in polymicrogyria gene panels (PMID:36211152, 2022). It is not in the KB at all. It should not be added to the grouping without resolving a genuine gene–disease validity question first: the Tuba8 mouse knockout has no brain phenotype (PMID:28388629), and PMID:28687668 is titled, pointedly, "Tubulin isotype specificity in neuronal migration: Tuba8 can't fill in for Tuba1a". Against that, PMID:32097653 (2020) reports Tuba8 driving cortical radial glia differentiation into apical intermediate progenitors via tubulin C-terminal modification, and PMID:41105144 (2026) reports TUBA8 promoting dendrite development. Recommendation: treat as a KNOWLEDGE_GAP-worthy candidate, not a member; a ClinGen-style validity check should precede any entry.

Deliberately out of scope (now documented in the grouping)

These are tubulin-family diseases that do not conform to the migration module, so the NECESSARY criterion excludes them correctly — but nothing in the file said so before this review:

  • TUBB4A — H-ABC hypomyelinating leukodystrophy / DYT4 dystonia. Still active research (PMID:42044700, 2026: H-ABC exhibits a cytoskeletal defect associated with microtubule stability; PMID:41547109/PMID:41685412, 2026: taiep rat model).
  • TUBB4B — Leber congenital amaurosis with early-onset deafness. Already curated in DisMech (kb/disorders/Leber_Congenital_Amaurosis_with_Early-Onset_Deafness.yaml) and deliberately modeled as a ciliopathy. New: PMID:41057290 (2025) extends it to cone-rod dystrophy with SNHL outside the canonical hotspot; PMID:41459724 (2026) identifies TUBB4B as the most abundant isotype governing ependymal ciliary polarity.
  • TUBA4A — ALS/FTD, and now myo-tubulinopathies, newly delineated in PMID:41678358 (Brain, 2026): 31 individuals from 19 families, with 17 families presenting myopathy without any CNS involvement, and — notably for a "dominant tubulinopathy" framing — three probands with recessive homozygous variants whose heterozygous carriers were asymptomatic.
  • TUBGCP2 / TUBGCP6 — gamma-tubulin complex genes rather than tubulin isotypes. PMID:42472988 (2026) reports a TUBGCP2-related tubulinopathy with cystic leukomalacia. If the grouping is ever re-scoped to "tubulin and tubulin-complex genes", this is the boundary that moves.

4. Per-member review

Section coverage across the four members

Section TUBA1A TUBB2A/2B TUBB3 TUBB/TUBB5
disease_term ✓ ✓ ✓ added in this review
inheritance — 1 1 1
epidemiology — 2 — —
prevalence — 1 — —
diagnosis — 2 2 —
progression — — 2 —
clinical_burden — — ✓ —
animal_models — — 1 —
has_subtypes — 2 — —
datasets — — 0 —
mappings — — — —
biological_scale on nodes — ✓ ✓ —
pathophysiology nodes 3 3 5 5
phenotypes 6 18 13 5
treatments 3 3 3 3

TUBA1A-related Tubulinopathy — the weakest entry for the most important disease

This is the review's headline finding. TUBA1A is the commonest and most severe tubulinopathy — the grouping's own differentiating_mechanisms says so, and the entry's own evidence quantifies it (~1% of classic lissencephaly and ~30% of lissencephaly with cerebellar hypoplasia, PMID:20466733). Yet it is the thinnest entry in the grouping: 645 lines, 23 evidence items, 5 distinct references, the newest from 2017, and 78.9% compliance.

Concretely missing: inheritance (the entry asserts "heterozygous, almost always de novo" in prose but never as a structured Inheritance block bound to HP:0000006), epidemiology/prevalence, diagnosis, progression, clinical_burden, animal_models (despite PMID:17218254 describing the founding ENU mouse, which the entry already cites as MODEL_ORGANISM evidence — it is a ready AnimalModel + modeled_mechanisms link), biological_scale tags on all three nodes, and mappings.

The 6 curated phenotypes are all neuroanatomical. The entry's notes is explicit and honest about why the clinical phenotypes (intellectual disability, motor delay, drug-resistant epilepsy, ataxia, ocular impairment) were left in prose: no quotable abstract snippet was found in the cited papers. That constraint has now lifted — PMID:42593952, PMID:42472988 and PMID:42177523 (§5I) all provide quantified, quotable clinical outcome statements.

The entry had no disease_term at all — a gap the grouping's own MONDO consistency note recorded. Fixed in this review: bound to MONDO:0014341 (CDCBM6, whose MONDO definition is precisely "cause of the disease is a mutation in the TUBB gene"); term validation and schema validation pass, and both the label cache and the diseaseterm enum-membership cache accepted the CURIE.

Remaining: 79.3% compliance, 22 evidence items, 4 references (newest 2017), 5 phenotypes, no biological_scale tags, and no animal_models despite the Breuss mouse (PMID:23246003, already cited) and the brand-new zebrafish (PMID:42241496, §5C). One discussion (gap_tubb_tubb5_natural_history_and_phenotype_breadth) has no evidence at all — flagged by just compliance as discussions[1] … evidence: MISSING. This is the entry's one outright schema-recommended omission and is now easy to fill from PMID:41152456 (Korean β-tubulinopathy cohort, which includes TUBB5) and PMID:42015805 (2026, intrafamilial variability with a novel missense TUBB variant).

TUBB2A/TUBB2B-related Cortical Malformation — structurally the best-built entry

93.2% compliance, 18 phenotypes, has_subtypes correctly used to hold the two genes as separate branches, epidemiology + prevalence present, and an explicit gap_tubb2ab_lumping_boundary discussion that argues the lump. Content is sound.

Two observations. First, the subtypes carry genes but no term:, so MONDO:0014337 (CDCBM5) and MONDO:0012399 (CDCBM7) are not retired. Second, the lumping argument's premise — "the published case evidence [for TUBB2A] is thinner than for TUBB2B" — is now weaker than when written: 2025–2026 produced a TUBB2A mutational hotspot paper (PMID:41080462), a TUBB2A epilepsy genotype–phenotype series (PMID:41872443), an adult TUBB2A ataxia phenotype (PMID:42050746), and a TUBB2A SUDEP report (PMID:42472988). The lump may still be right, but the gap_tubb2ab_lumping_boundary discussion should be re-argued against this evidence rather than left standing on the 2021 position.

95.7% compliance, 5 pathophysiology nodes including the module's Axon Guidance and Projection Wiring Defects branch (the only member that uses it), a real animal_models entry with two modeled_mechanisms links, progression, clinical_burden, diagnosis, three well-formed discussions, review_notes, and the only therapeutic_modality-tagged treatment in the grouping. This entry is the template the other three should be raised to.

Minor: the Cranial Motor Nerve Maldevelopment and Ocular Dysmotility node has no conforms_to, because the migration module has no node for cranial-motor-nerve maldevelopment. That is a reasonable local specialization, not a defect, but it is the one place where the module under-covers a member.

Cross-cutting

  • No member has datasets: (0% on all four).
  • No member has mappings:.
  • No treatment in any member uses target_mechanisms, so the module's drug-target pattern is entirely unexercised across the grouping. Historically correct — management was purely supportive — but §5B changes that.
  • Treatments are near-identical triplets (anti-seizure medication, supportive care, genetic counseling) with therapeutic_modality unset except on TUBB3's surgery entry. None of these has a mechanical treatment_term.term.id → modality mapping in CLAUDE.md's backfill table, so each needs a per-entry look rather than a blind rule.

5. Latest literature (2024–2026) and the knowledge gaps it opens

The whole grouping is 3–9 years behind the literature. The newest reference anywhere in the four entries is PMID:35915025 (2023, TUBB3 only); TUBA1A and TUBB/TUBB5 both stop at 2017. Nothing from 2024, 2025 or 2026 is cited anywhere. The 2025–2026 window has been unusually productive for this field, and several papers do not merely add citations — they open mechanism branches the pathographs do not model.

A. Ciliogenesis is a second, evidenced pathogenic arm — currently absent from the KB

PMID:41309602 (Nat Commun, Nov 2025) — Mutations in the β-tubulin TUBB impair ciliogenesis and are associated with ciliopathy-like phenotypes. A de novo heterozygous TUBB missense variant produces features of both ciliopathy and tubulinopathy. In patient-derived cells and gene-edited isogenic lines the variant impairs early cilium formation by altering microtubule dynamics and structure; knock-in mice show decreased ciliation in cerebellum and kidney. Two further conclusions matter for curation: the mechanism is explicitly not haploinsufficiency, and other patient TUBB mutations also affect cilium formation — so this is a subset mechanism, not a one-variant curiosity.

PMID:42091926 (npj Genom Med, May 2026) — Bridging the gap: an emerging link between tubulinopathies and ciliopathies — is the review that frames it.

Gap. TUBB_TUBB5-related_Microcephaly currently models a single chain (tubulin → spindle/cell-cycle → p53 apoptosis → microcephaly). A ciliogenesis branch is missing entirely, and DisMech has both a ciliopathy_dysfunction module and a Ciliopathies grouping for it to conform to. This is the single highest-value mechanistic addition identified by this review, and it also bears on the DisMech boundary between the Tubulinopathies and Ciliopathies groupings — and on the TUBB4B LCA entry, which DisMech already models as a ciliopathy.

B. The first mutation-independent therapeutic strategy

PMID:42589608 (Int J Mol Sci, Aug 2026) — Restoring the Balance: CRISPRa-Driven β-Tubulin Compensation as a Strategy for Tubulinopathy Treatment. CRISPR-Cas9 activation upregulates non-mutated β-tubulin isotypes; the authors demonstrate restoration of the microtubule network and of primary cilium formation. The design rationale is precisely the problem the grouping exists to describe: mutations are scattered across distinct tubulin genes, which defeats per-variant editing, so the strategy targets the shared downstream state instead.

Gap. No member entry has any disease-modifying treatment, and none uses target_mechanisms. This is the first candidate that would attach to a mechanism node (GENE_THERAPY/GENE_EDITING modality, target_mechanisms → the microtubule-apparatus node), and it should be curated with its preclinical status stated plainly.

C. A rescuable Notch arm in TUBB5

PMID:42241496 (Hum Mol Genet, Jun 2026) — tubb5 knockout in zebrafish causes neurodevelopmental defects via notch pathways. tubb5-null larvae show developmental delay, craniofacial malformation, uncoordinated movement, increased seizure susceptibility and impaired photomotor response. Transcriptomics show upregulated notch1a/her5 (which inhibit neural progenitor differentiation), with decreased neurogenin1/huc-positive and increased sox2-positive cells — i.e. progenitors stuck undifferentiated. DAPT (γ-secretase inhibition) rescued both developmental and locomotor deficits; carbamazepine and valproate ameliorated locomotor dysfunction and PTZ-induced seizure susceptibility.

Gap. This is (i) a ready animal_models entry for TUBB/TUBB5, which currently has none, with PERTURBS/RESCUES modeled_mechanisms links and concrete readouts; (ii) a distinct mechanistic arm from the entry's existing p53-apoptosis chain; and (iii) a second target_mechanisms treatment candidate. Curate it with its species caveat explicit — HUMAN_MODEL_MISMATCH is the right kind for the DAPT rescue arm.

D. A complete functional atlas of TUBA1A variants

PMID:42213754 (PNAS, Jun 2026) — Comprehensive mutagenesis defines the functional landscape of human α-tubulin. All 2,683 single-nucleotide coding variants of TUBA1A profiled by high-content live-cell imaging, resolving distinct mutation classes that disrupt folding, chaperone engagement, and protofilament geometry, with MD simulations showing how perturbations in GTP binding, dimer contacts and lateral interfaces propagate to filament architecture, and a predictive framework that generalizes across isotypes.

Gap. The TUBA1A entry currently models a single undifferentiated "Altered Alpha-Tubulin Function" node whose evidence is from 2007–2010. This paper supplies a variant-class structure the pathograph does not have, and it partially answers the entry's own open gap_tuba1a_human_organoid_translatability question — which should be re-scoped rather than left as originally posed.

E. Dominant-negative / gain-of-function is now a resolved question, and uncurated

No member entry carries a functional_impact_category on any GeneticContext, or discusses the mechanism class at all. The literature has settled it: tubulinopathy variants are missense-only (no nonsense, frameshift or whole-gene deletion), PMID:41309602 states outright that the TUBB mechanism "is not haploinsufficiency", and PMID:30517687 (2019) shows TUBA1A R402C/R402H patient alleles dominantly disrupt cortical migration in the developing mouse brain and impair dynein activity.

Gap. Per CLAUDE.md's decision tree, this belongs in GeneticContext.functional_impact_category (DOMINANT_NEGATIVE, and where supported GAIN_OF_FUNCTION), not in a descriptor modifier. Adding it makes an important, well-evidenced, currently-invisible claim machine-queryable across all four members. PMID:30517687 is also a missing mechanistic anchor for TUBA1A in its own right — the dynein link is the mechanistic bridge to the module's motor-protein biology.

F. The tubulin code — the evidence for the grouping's own central claim

The grouping's grouping_rationale asserts that members stay separate because isotypes have "distinct expression timing and binding partners". That claim is currently unevidenced anywhere in the KB, and 2026 supplied the evidence:

  • PMID:42179625 (2026) — C-terminal tails of TUBB2A/2B/2C/3/4A/5 differentially regulate cytoplasmic dynein motility; isotype-specific lateral protofilament interactions determine β-CTT proximity to the dynein microtubule-binding domain.
  • PMID:42275208 (2026) — spatial and temporal atlas of tubulin isotype gene expression during vertebrate embryonic development.
  • PMID:41847015 (2026) — the same for neural crest EMT.
  • PMID:41902449 (2026) — methods for visualizing specific tubulin isotypes and pathogenic variants in cellular microtubule arrays.
  • PMID:36943622 (2023) — role of α- and β-tubulin isotypes in early brain development.
  • PMID:41501376 (2026) — autoregulatory control of tubulin abundance, the mechanism that determines whether isotype compensation (§5B) can work.

Gap. The microtubule_dependent_neuronal_migration_failure module is the natural home for an isotype-specificity discussion; the grouping's rationale should cite it rather than assert it.

G. Phenotype expansion beyond what the entries model

PMID Year Expansion Affects
PMID:42310787 2026 Dandy-Walker malformation reclassified into the tubulinopathy spectrum (TUBB2B, TUBB3) TUBB2A/2B, TUBB3
PMID:41153399 2025 TUBB2B p.Ile202Thr causes syndromic CFEOM crosses the grouping's asserted TUBB3-owns-CFEOM boundary
PMID:42050746 2026 TUBB2A ataxia with preserved ambulation into adulthood TUBB2A/2B — far milder than curated
PMID:41872443 2026 TUBB2A-related epilepsy, novel variants + genotype–phenotype TUBB2A/2B
PMID:40729534 2025 TUBA1A infantile epileptic spasms syndrome + atypical absence seizures TUBA1A
PMID:42472988 2026 SUDEP in TUBB2A; movement disorders (dystonia, mirror movements) in 33% all
PMID:41048055, PMID:37713978 2026/2023 Congenital mirror movements (TUBB2B, TUBB3) TUBB2A/2B, TUBB3
PMID:42015805 2026 Marked intrafamilial phenotypic variability, novel TUBB variant TUBB/TUBB5

The PMID:41153399 finding deserves emphasis: the grouping's differentiating_mechanisms for TUBB3 says CFEOM3 is what distinguishes it from its siblings. A TUBB2B variant causing syndromic CFEOM does not invalidate that, but it does mean the differentiator is a tendency rather than a boundary, and the wording should soften.

H. Prenatal diagnosis is now a defined modality

PMID:41901019 (2026, narrative review of prenatal neurosonography in tubulinopathy) and PMID:41171976 (2026, prenatal diagnosis with a novel TUBA1A variant). TUBA1A and TUBB/TUBB5 have no diagnosis: block at all; TUBB2A/2B and TUBB3 have postnatal ones. Given that these are severe, often prenatally detectable malformations, a prenatal neurosonography diagnosis entry is a real content gap for all four.

I. Natural history and outcome data now exist for exactly this member set

PMID:42593952 (Am J Intellect Dev Disabil, Jul 2026) — Developmental Profiles Associated with TUBA1A, TUBB2A, TUBB2B, and TUBB3 Tubulinopathy Conditions. Caregiver Developmental Profile Scales-4 ratings for 32 individuals across precisely the grouping's gene set. Findings: challenges across all domains, motor skills most severely affected; TUBB3 conditions relatively milder; co-occurring vision impairment in 71%, and those individuals showed greater overall delays.

PMID:42177523 (Orphanet J Rare Dis, May 2026) — gene-specific long-term course, neurodevelopmental outcome and quality of life across LIS1/DCX/DYNC1H1/TUBA1A/TUBG1 lissencephaly, with PedsQL Family Impact Module scores (parental HRQL mean 61.23) and supportive-therapy effectiveness data.

PMID:41152456 (2026, Korean β-tubulinopathy cohort, n=12) and PMID:42472988 (2026, Turkish multicentre cohort, n=15, seven genes) and PMID:40179460 (2025, Japanese single-centre retrospective) supply the frequency denominators that phenotype frequency: qualifiers require (see docs/frequency-evidence-guidelines.md).

Gap. Only TUBB3 has progression and clinical_burden. PMID:42593952 directly substantiates the grouping-level claim that TUBB3 is the mildest member — which is currently nowhere in the KB — and would be strong evidence in the grouping's own differentiating_mechanisms.

6. Prioritized recommendations

# Action Target Effort
1 Curate TUBG1-related tubulinopathy (CDCBM4, MONDO:0014171) and add as a fifth member grouping new entry
2 Add the ciliogenesis branch to TUBB_TUBB5-related_Microcephaly (PMID:41309602, PMID:42091926), with conforms_to on ciliopathy_dysfunction TUBB/TUBB5 medium
3 Raise TUBA1A to parity — inheritance, diagnosis (incl. prenatal), progression, clinical_burden, animal_models (ENU mouse, already cited), biological_scale tags, and clinical phenotypes now that quotable cohort text exists TUBA1A large
4 Add animal_models for the tubb5 zebrafish with the Notch arm and DAPT rescue (PMID:42241496) TUBB/TUBB5 medium
5 Curate CRISPRa β-tubulin compensation (PMID:42589608) as an emerging target_mechanisms treatment — the grouping's first non-supportive therapy module or all members medium
6 Add functional_impact_category: DOMINANT_NEGATIVE to the GeneticContext blocks, with PMID:41309602 / PMID:30517687 all four small
7 Fill the empty evidence on gap_tubb_tubb5_natural_history_and_phenotype_breadth (PMID:41152456, PMID:42015805) TUBB/TUBB5 small
8 Add mappings.mondo_mappings and subtype term: values to retire CDCBM1/5/6/7 and CFEOM3A all four small
9 Refresh gap_tubb2ab_lumping_boundary against 2025–2026 TUBB2A evidence TUBB2A/2B small
10 Soften the TUBB3 differentiating_mechanisms CFEOM claim in light of PMID:41153399 grouping small
11 Add the isotype-specificity evidence (PMID:42179625, PMID:42275208) to the migration module, supporting the grouping's split rationale module small
12 File an upstream MONDO issue: link CDCBM1/4/6/7 and MONDO:0012703 under MONDO:0100153 upstream small

7. Changes applied

Update, same day. After this review was delivered, all twelve recommendations in §6 were applied. See §9 for what was done and for the two places where doing the work corrected the review itself.

The review pass itself was deliberately minimal. Three defects that the grouping file already named were fixed:

  1. kb/disorders/TUBB_TUBB5-related_Microcephaly.yaml — added the missing disease_term (MONDO:0014341, complex cortical dysplasia with other brain malformations 6). Term validation and schema validation pass; cache/mondo/terms.csv and cache/enums/diseaseterm_*.csv picked up the CURIE.
  2. kb/groupings/Tubulinopathies.yaml — rewrote the MONDO consistency.notes so it is accurate after fix 1 and records the upstream MONDO inconsistency (§2) as the actual cause of the 2/4-descendant signal.
  3. kb/groupings/Tubulinopathies.yaml — added a scope boundary paragraph to notes naming the deliberately-excluded tubulin diseases (TUBB4A, TUBB4B, TUBA4A, TUBGCP2/6) and a known membership gap paragraph naming TUBG1 and Uner Tan syndrome, so that an in-scope gap is no longer indistinguishable from an out-of-scope exclusion.

8. Verification

  • just check-groupings kb/groupings/Tubulinopathies.yaml — structure: OK, all four members SATISFIED.
  • just validate-grouping kb/groupings/Tubulinopathies.yaml — passed.
  • just validate kb/disorders/TUBB_TUBB5-related_Microcephaly.yaml — passed; Snippets checked: 22/22 verified against cached references.
  • just validate-terms kb/disorders/TUBB_TUBB5-related_Microcephaly.yaml — passed.
  • Every PMID and title in this report was retrieved live from NCBI E-utilities during the review. PMID:37475831 is flagged above as retracted. No PMID cited here has yet been added to a kb/ entry — doing so requires just fetch-reference plus exact-quote snippet verification per the evidence SOP.

9. Follow-up: all twelve recommendations applied (2026-08-20)

Every recommendation in §6 was carried out. Two of them could not be applied as written, and the reasons are findings in their own right.

The review was wrong about mechanism class (recommendation 6)

Recommendation 6 said to add functional_impact_category: DOMINANT_NEGATIVE to all four members. Reading the primary sources to write the annotation showed that is only right for two of them, and actively wrong for a third:

Member Applied value Why
TUBA1A DOMINANT_NEGATIVE PMID:30517687 resolves it directly — mutant α-tubulin "acts dominantly by populating microtubules with defective binding sites for dynein", and dynein impairment scales with mutant expression. A poisoning mechanism.
TUBB3 DOMINANT_NEGATIVE Folded mutant heterodimers still polymerize and a subset disrupts the kinesin interaction — a subunit that incorporates and degrades a motor-binding surface.
TUBB / TUBB5 GAIN_OF_FUNCTION PMID:41309602 tests haploinsufficiency directly and excludes it (TUBB-haploid cells ciliate normally), and concludes the variants "all act in a gain-of-function fashion". The review had read "not haploinsufficiency" as implying dominant-negative. It does not.
TUBB2A / TUBB2B UNKNOWN Impaired heterodimer formation is evidenced for both, but that finding does not discriminate a poisoning subunit from a reduced functional-heterodimer pool. UNKNOWN here means examined and unresolved, and is paired with a new gap_tubb2ab_functional_impact_class discussion.

The distinction is not bookkeeping: it decides whether the isotype-compensation strategy of §5B should be expected to work, since adding wild-type subunits dilutes a poisoning allele but does not remove it.

The ciliopathy module had to be widened first (recommendation 2)

TUBB_TUBB5 could not honestly conforms_to ciliopathy_dysfunction#Basal Body and Transition Zone Dysfunction as the node was written: it admitted only lesions in basal body, transition zone, BBSome or IFT component genes, and TUBB is none of those. Rather than force the conformance or drop it, the module node was widened to admit a third entry route — a dominant tubulin-subunit variant impairing cilium assembly directly — with PMID:41309602 and PMID:42091926 as evidence, plus two guardrails in the module notes: conformance needs evidenced ciliary impairment in that disorder (not merely a microtubule lesion), and a tubulin loss-of-function lesion is not evidence for this route, since the mechanism is dose-dependent gain-of-function.

Recommendation 12 landed as a DisMech issue, not an upstream one

This session has no push access to monarch-initiative/mondo, so the upstream request is filed as monarch-initiative/dismech#9113 with the full evidence table, for someone with upstream access to carry over.

What changed, by file

File Change
kb/disorders/TUBG1-related_Tubulinopathy.yaml New entry (R1). CDCBM4 / MONDO:0014171, 4 pathophysiology nodes, 8 phenotypes, 2 animal models, 2 knowledge gaps, 50/50 snippets verified.
kb/groupings/Tubulinopathies.yaml TUBG1 added as fifth member (R1); TUBB3 CFEOM differentiator softened with evidence (R10); MONDO consistency note and scope/gap notes updated.
kb/disorders/TUBA1A-related_Tubulinopathy.yaml Raised to parity (R3): inheritance, epidemiology, 3 diagnosis entries incl. prenatal neurosonography, 3 progression phases, clinical burden, 2 animal models, scale tags, 7 clinical phenotypes, DOMINANT_NEGATIVE context. 78.9% → ~88% compliance.
kb/disorders/TUBB_TUBB5-related_Microcephaly.yaml Ciliogenesis node (R2), Notch node + zebrafish model with DAPT rescue (R4), GAIN_OF_FUNCTION context (R6), gap evidence filled (R7), new translatability discussion.
kb/disorders/TUBB2A_TUBB2B-related_Cortical_Malformation.yaml Lumping gap re-argued against 2025–26 evidence (R9), UNKNOWN context + new mechanism-class gap (R6), Uner Tan boundary discussion, subtype terms + mappings (R8).
kb/disorders/TUBB3-related_Tubulinopathy.yaml DOMINANT_NEGATIVE context (R6), CDCBM1 + CFEOM3A mappings (R8), inherited-allele qualifier.
kb/modules/microtubule_dependent_neuronal_migration_failure.yaml CRISPRa β-tubulin compensation treatment with target_mechanisms (R5) — the grouping's first non-supportive therapy; isotype-specificity knowledge gap (R11).
kb/modules/ciliopathy_dysfunction.yaml Trigger node widened to admit tubulin-subunit lesions, with guardrails (R2).

Seven history records were written under history/. 30 new references were fetched with just fetch-reference; every snippet in every file above is quoted from a cached abstract and verified.

Two things deliberately not done

  • TUBA8 was not added as a member. §3 flagged its gene–disease validity as contested, and nothing in this pass resolved it. It stays a candidate.
  • TUBG1 was not added to Lissencephaly_and_Neuronal_Migration_Disorders, where it would also fit. Recommendation 1 scoped it to this grouping; widening to a second grouping is a separate call.

10. Coverage pass: every tubulin-family disease curated (2026-08-20)

The review above is scoped to the grouping — five members, all cortical-malformation diseases. A follow-on pass extended coverage to the whole tubulin gene family, on the principle that a disease should not go uncurated merely because it falls outside one grouping's boundary. Eight new entries:

Entry MONDO What it is
TUBB4A-related_Neurologic_Disorder MONDO:0800470 H-ABC hypomyelinating leukodystrophy / DYT4 dystonia — myelin, not migration
TUBA4A-related_Disorder MONDO:0014531 ALS22/FTD, hereditary spastic ataxia, and the 2026 myo-tubulinopathies (17/19 families with no CNS disease at all)
TUBB8-related_Oocyte_Maturation_Defect MONDO:0021573 Meiotic spindle assembly failure; a primate-specific isotype, so no mouse model is possible
TUBB1-related_Macrothrombocytopenia MONDO:0800047 Megakaryocyte marginal-band failure
TUBGCP4-related_Microcephaly_and_Chorioretinopathy MONDO:0014592 γ-TuRC component; microcephaly without cortical malformation
TUBGCP6-related_Microcephaly_and_Chorioretinopathy MONDO:0009624 As above, plus retinal dysfunction
TUBA8-related_Polymicrogyria_with_Optic_Nerve_Hypoplasia (none) Curated as association: Suspected — see below
Uner_Tan_Syndrome MONDO:0100144 Biallelic TUBB2B p.Arg390Gln; the recessive, cerebellar, basal-ganglia-sparing outlier

TUBA8 is the entry that changed shape while being written. The check-title-snippets gate refused a title-quoted snippet, which forced a read of the full PMID:28388629 abstract — where the authors report re-analysing the original human subjects by exome sequencing and finding a homozygous loss-of-function SNAP29 variant, "suggesting that SNAP29 deficiency, rather than TUBA8 deficiency, may underlie most or all of the neurodevelopmental anomalies." The linked-bystander scenario raised in §3 as a hypothetical had in fact already been found. The entry is curated with the SNAP29 finding as supports: REFUTE, and the gene–disease relationship as Suspected. This is a worked example of a validation gate catching a substantive error rather than a formatting one.

None of the eight joined the Tubulinopathies grouping: the NECESSARY criterion tests migration-module conformance, and none of them conform.

11. QA pass (2026-08-22)

An adversarial re-read of everything above. Cross-entry foreign-key checking over the 13 tubulin-family entries (every downstream.target, conforms_to anchor, target_mechanisms target, model link, discussion attaches_to, and subtype FK) came back clean, as did duplicate-disease_term checking across all 2,099 entries. The pass found five things worth fixing, three of them errors in the work above.

One more disease was uncurated

TUBGCP2-related_Lissencephaly_Spectrum_Disorder — no MONDO disease term exists (only the gene, HGNC:18599), which is why a MONDO-driven sweep missed it. The gene–disease relationship is nonetheless solid: AJHG 2019 delineation (PMID:31630790), independent replication with functional work in 2021 (PMID:33458610), a 2025 literature review (PMID:40017707), and inclusion in a 2026 multicentre cohort (PMID:42472988).

It is mechanistically the interesting one of the γ-TuRC genes. TUBGCP4 and TUBGCP6 cause microcephaly with chorioretinopathy and a structurally normal cortex; TUBGCP2 causes frank pachygyria and subcortical band heterotopia and conforms to all three nodes of the migration module. The literature calls it a tubulinopathy — PMID:40448381 does so in its title. It is still excluded from the grouping, on gene identity: GCP2 is a γ-tubulin complex protein, not a tubulin. That is a lumping decision, not a fact, and it is now recorded on both sides so it can be revisited. It also demonstrates why the grouping's criterion is NECESSARY (audit-only) rather than SUFFICIENT: a disorder can satisfy the mechanism half of the conjunction and still not be a member.

functional_impact_category is left UNKNOWN rather than defaulted to LOSS_OF_FUNCTION, with a knowledge gap explaining why: the allele spectrum contains a multi-exon deletion and a frameshift (unambiguously null) and a missense allele that leaves GCP2 protein levels normal while mislocalizing γ-tubulin, HAUS6 and NEDD1.

Three stale or contradictory statements introduced by the earlier passes

  1. The grouping described itself as alpha- and beta-tubulin only — in both description and grouping_rationale — while listing TUBG1 (gamma) as a member. Added in the same pass that added TUBG1. Fixed.
  2. The grouping's notes said Uner Tan syndrome "is still not covered by any member entry or subtype." It had been curated as its own entry two sections up in this very report. Fixed.
  3. The scope-boundary note listed TUBB4A / TUBB4B / TUBA4A / TUBGCP2 / TUBGCP6 as "out of scope" without saying that most of them now have entries, and omitted TUBGCP4, TUBB8, TUBB1 and TUBA8 entirely. Rewritten to name each entry and to separate the two distinct exclusion reasons (fails the mechanism criterion vs. excluded on gene identity).

Two scoping gaps

  1. Uner_Tan_Syndrome did not say that the eponym is broader than the entry. The MONDO binding is correct — MONDO:0100144 is defined as the TUBB2B R390Q entity — but the clinical literature also applies the name to VLDLR, CA8, WDR81 and ATP8A2 families, which MONDO keeps separate as CAMRQ 1–4. A reader could have taken the entry as covering all of them. Caveat added; the CAMRQ concepts are not tubulin disorders and remain uncurated.
  2. MONDO:1060115 (TUBB4B-related ciliopathy) was referenced nowhere in the KB, despite being the direct is-a parent of the curated LCAEOD entry — the last tubulin-family MONDO disease concept with no KB reference. Added as skos:broadMatch, not exactMatch: the umbrella's own definition says diagnoses under it can include primary ciliary dyskinesia, which this entry does not cover. As a broadMatch it does not retire the concept from the curation queue, which is the intended outcome.

Two additions from currency checking

The 2026 Turkish multicentre cohort (PMID:42472988) had not been read by any entry. Two findings were curated onto TUBB2A/TUBB2B-related Cortical Malformation: the first report of probable SUDEP in a TUBB2A patient — which makes the epilepsy here a mortality risk to counsel, not only a seizure burden — and movement disorders in 33.3% including dystonia and mirror movements. No frequency band was asserted for the latter, because the percentage spans all seven cohort genes rather than these two. Mirror movements are worth noting mechanistically: they implicate the module's axon-guidance branch, not the migration branch that explains the cortical malformation.

One deferred item closed

TUBG1 was added to Lissencephaly_and_Neuronal_Migration_Disorders (§9 left this open), along with TUBGCP2. All 17 members of that grouping audit as SATISFIED.

What changed, by file

File Change
kb/disorders/TUBGCP2-related_Lissencephaly_Spectrum_Disorder.yaml New entry. 4 pathophysiology nodes conforming across all three migration-module nodes, 8 phenotypes, 2 knowledge gaps, 27/27 snippets verified, 86.8% weighted compliance.
kb/groupings/Tubulinopathies.yaml Gamma-tubulin admitted in description and grouping_rationale; stale Uner Tan gap note corrected; scope boundary rewritten with entry names and the two exclusion classes.
kb/groupings/Lissencephaly_and_Neuronal_Migration_Disorders.yaml TUBG1 and TUBGCP2 added as members with differentiating mechanisms.
kb/disorders/TUBB2A_TUBB2B-related_Cortical_Malformation.yaml Probable-SUDEP evidence on the epilepsy phenotype; new Movement Disorder phenotype. Both from PMID:42472988.
kb/disorders/Uner_Tan_Syndrome.yaml Eponym scope caveat naming the four CAMRQ concepts.
kb/disorders/Leber_Congenital_Amaurosis_with_Early-Onset_Deafness.yaml MONDO:1060115 broadMatch mapping.

Four further history records written. Every new snippet is quoted from an abstract cached by just fetch-reference and verified.