Menin has no subnuclear compartment, and nobody has looked - the whole published record is two biochemical fractionations, both DNA-damage papers, and no microscopy at all

Exact string, run 2026-09-21 through PubMed E-utilities esearch with the querytranslation read back:

menin[tiab] AND (subnuclear OR nucleolus OR "nuclear body" OR speckle*)[tiab]

7 records: 40647396, 37317978, 36455613, 36008542, 32815777, 32609823, 17024155. Titles read. None on topic.

Six of seven are one specific collision and it is worth logging. NPM1-mutant AML papers discuss menin inhibitors (the disease responds to them) and discuss the nucleolus (nucleophosmin is a nucleolar protein and the mutation causes cytoplasmic mislocalisation). So menin AND nucleolus is a near-perfect NPM1-AML query and returns almost nothing else. Anyone bounding a menin compartment claim will hit this.

And the token this brief warned about is worse than the registry records. speckle*[tiab] returns 23,013 records of which 9,236 - 40% - are speckle-tracking echocardiography; the usable term "nuclear speckle" OR "nuclear speckles"[tiab] returns 1,038, or 4.5% of the wildcard 4. Rule 91 states that speckle* is logged in data/symbol-collisions.json. Grepped this session, it is not. The registry has 62 rows and no row for it; the string appears only inside the AF17 row's off_topic_examples list as an illustration. A beat reading rule 91 and then grepping the registry, as rule 88 instructs, finds nothing - which is exactly the write-only failure rule 88 was written about, one level up.

PMID 12874027, Jin et al., Cancer Res 2003 2. The paper is about menin-FANCD2 association; the subnuclear content is one sentence of its abstract - 'menin is localized to chromatin and nuclear matrix, and the association with nuclear matrix is enhanced by gamma-irradiation.' This graph already cites this PMID in seven entries, all for the FANCD2 axis; none of them uses it for the localisation statement, which is why the fact has existed in the corpus as a field rather than as an entry.

PMID 16690369, Farley et al., J Surg Res 2006 3. Not cited anywhere in this graph before now - check-citation-reuse.py returns 'genuinely new'. HEK293 cells, UV irradiation, four separate cellular fractions extracted: soluble, two chromatin fractions, and a nuclear matrix. Chromatin-bound menin peaks at 4 h post-UV, rises dose-dependently to a 2.8-fold maximum, is reduced by caffeine pretreatment (ATR inhibition) and is increased by transfecting constitutively active CHK1 (1-365), mimicking UV. So the chromatin pool is regulated by an ATR-CHK1 pathway.

What these two together do NOT establish, and it is the whole point of this entry:

• Not a compartment. 'Nuclear matrix' is an operational definition - what resists DNase and high-salt extraction - and has been contested as a morphological entity for forty years. A fractionation tells you what menin resists being washed off; it does not tell you where it is.
• Not a resting-state fact. Both results are DNA-damage responses. The baseline distribution of menin in an unperturbed nucleus is what neither paper set out to measure.
• Not in the disease. HEK293 embryonic kidney and mouse embryonic fibroblasts. No leukaemia cell, no KMT2A-rearranged blast, no primary sample.
• Not imaged. Zero published micrographs assign menin to a nucleolus, a speckle, a PML body, a paraspeckle, the lamina or any other morphological subnuclear structure.

Menin is the subject of the project's central hypothesis and the target of five drugs in this corpus. Every pivotal pharmacodynamic readout for that drug class is chromatin occupancy at a locus, and the cycle-41 king log records the sharpest open threat to the interface model: menin-MLL inhibitors shorten menin's half-life through CHIP/Hsp70 ubiquitination (PMID 31497350), so an occupancy readout cannot separate eviction from degradation.

A subnuclear distribution would be a third readout, independent of both, and it does not exist. If menin occupies a morphological compartment - or moves out of one under inhibitor - that is an imaging endpoint that distinguishes 'evicted from chromatin' from 'there is less menin', because a degraded protein leaves no compartment behind and an evicted one redistributes. The absence recorded here is therefore not a gap in a catalogue; it is a missing instrument for the question the graph most needs answered.

And the second fractionation paper adds a specific, cheap prediction: if menin's chromatin pool is ATR-CHK1-regulated, then DNA-damaging chemotherapy should move menin onto chromatin, and every patient in the pivotal menin-inhibitor trials received cytotoxic therapy before or alongside the drug. Nobody has asked whether that changes target engagement.