Oncofusion

start / graph / gene-smc3

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q43-break-repair-machinery

SMC3 - the cohesin ATPase arm, the only subunit whose partial loss has been tested directly against a KMT2A fusion, and it was tolerated

also called SMC3 · SMC3L1 · CSPG6 · BAM · BMH · structural maintenance of chromosomes protein 3

SMC3 is one of the two long hinged arms of the cohesin ring. Cutting its dose in half in mouse blood progenitors that already carry a KMT2A fusion did essentially nothing - which quietly kills the idea that cohesin is a weak point you could attack in this leukaemia.

The same thing, for a scientist

SMC3 (Q9UQE7, 1217 aa) is one of the two SMC ATPase subunits of cohesin, acetylated on its head domain by ESCO1/2 to establish cohesion. Its place here rests on one positive association and one direct negative. **Positive:** 256 SMC3 binding sites were preferentially break-prone in KMT2A-rearranged AML patients' normal cells relative to non-fusion AML patients, one of the four features whose intersection defined the 34 most break-prone regions [E2]; and SMC3 sites falling within genes were enriched among highly expressed genes, consistent with the cohesin-plus-transcription model of TOP2-induced breakage [E2, E5]. **Negative, and it is the load-bearing one:** **Smc3 haploinsufficiency was well tolerated in MLL-AF9-expressing murine haematopoietic stem and progenitor cells**, with minimal effect on in-vitro self-renewal, alongside a tolerated complete Stag2 loss [E3]. The authors set out to test a synthetic-lethal hypothesis motivated by the rarity of cohesin mutations in MLL-rearranged AML and found the opposite of what they expected, proposing instead that cohesin loss and MLL fusions drive self-renewal through a **shared** HOXA-upregulation mechanism - so cohesin mutations are rare in MLL-r AML because they would add nothing, not because they would be lethal. In adult myeloid malignancy, SMC3 mutations group with RAD21 and SMC1A as de-novo-AML-like with better prognosis, against STAG2's MDS-derived secondary-AML pattern [E4].

Go deeper
Why a tolerated haploinsufficiency is the most useful fact in this sub-beat

A hypothesis this graph holds - hyp-kmt2a-aff1-shared-loop-anchor-hub - proposes cohesin-anchored loop geometry as the explanation for where KMT2A breaks, and names an acute RAD21 degron as the decisive experiment while flagging tolerability as the feasibility crux.

The Smc3 result speaks to both halves. On feasibility it is encouraging: halving an essential cohesin subunit in fusion-expressing progenitors was tolerated, so the degron readout is achievable. On the biology it is discouraging in a different way: if halving SMC3 does not perturb an MLL-AF9 cell's self-renewal, cohesin dose is not a lever on the established disease. Neither of those is a statement about rearrangement FREQUENCY, which is what the hypothesis actually predicts and which nobody has measured under cohesin depletion. Self-renewal and break frequency are different endpoints and the published experiment used the first.

A note on the acetylation axis, deliberately not claimed

SMC3's regulation by acetylation (ESCO1/ESCO2 writing, HDAC8 erasing) is one of the few places where cohesin meets a drug class this graph already holds an entire beat on. No KMT2A-rearranged evidence for that link was found in this beat and none is asserted. It is recorded as an open question rather than as a section with citations, because a plausible connection between two things the graph contains is exactly the kind of sentence that has previously entered this project as a fact.

Facts and figures
uniprotQ9UQE7
length AA1217
gene symbolSMC3
accession resolutiongene_exact:SMC3 AND organism_id:9606 AND reviewed:true returned exactly 1 row; gene_primary SMC3, length 1217, synonyms BAM BMH CSPG6 SMC3L1.
sister subunitgene-smc1a (Q14683, 1233 aa - only 16 residues apart, so length is a weak discriminator for this pair)
complexcohesin
break prone sites in kmt2a r patients256
clinical mutation classgroups with RAD21 and SMC1A as de-novo-AML-like, better prognosis (PMID 39033241)
acetylation axis not claimedSMC3 is acetylated by ESCO1/ESCO2 and deacetylated by HDAC8; no KMT2A-rearranged evidence for this axis was found in this beat and none is asserted
subnuclear localizationCHROMOSOME ARMS IN INTERPHASE, CENTROMERE FROM PROPHASE - a CHROMATIN address rather than a nuclear body, and it is stated here as such deliberately. SMC3 has no assignment to any membraneless compartment: SMC3[tiab] against nucleolus, nucleolar, nuclear speckle and nuclear matrix returns 9 records and none is a compartment assignment for the human protein [E_q91_smc3_loc3]. The defensible subnuclear statement is the cell-cycle-scoped chromatin distribution above, and in this graph it matters at a finer grain than 'chromatin': SMC3 binding sites are loop anchors, 256 of them were preferentially break-prone in the normal cells of KMT2A-rearranged AML patients relative to non-fusion AML patients, and SMC3 sites within genes were enriched among highly expressed genes (this entry's E2/E5). That is the closest thing this project has to a SUBNUCLEAR-scale localisation of SMC3 in this disease, and it is a locus-level map rather than an image.
localization shown for fusionNO for any image, in any fusion, in any lineage - SMC3 protein has never been localised in a fusion-bearing cell. What HAS been done in the index disease is locus-level and is the better fact: SMC3 binding-site positions were mapped in KMT2A-rearranged AML patients' normal cells and 256 of them are preferentially break-prone (this entry's E2). Bounded: SMC3[tiab] AND MLL[tiab] = 2 and SMC3[tiab] AND KMT2A[tiab] = 5; cohesin[tiab] AND MLL[tiab] = 7, run because rule 64 requires the complex form for a subunit absence, and the complex form adds five records the subunit form misses [E_q91_smc3_loc3].
direct test against a kmt2a fusion
{
  "perturbation": "Smc3 haploinsufficiency in MLL-AF9-expressing murine HSPCs",
  "result": "well tolerated, minimal effect on in-vitro self-renewal",
  "endpoint_tested": "self-renewal",
  "endpoint_NOT_tested": "KMT2A rearrangement frequency or within-BCR break position",
  "status": "preprint",
  "pmid": "38352423",
  "title": "Cohesin loss and MLL-AF9 are not synthetic lethal in murine hematopoietic stem and progenitor cells.",
  "authors": "Meyer A, Stelloh C, Zhu N, et al.",
  "venue": "Res Sq",
  "year": 2024,
  "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -1.3485,
  "chronos_mean_other_leukaemia": -1.2462,
  "difference": -0.1023,
  "pan_essentiality_all_lines": -1.169,
  "selectivity_rank": 1093,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
localization
{
  "CONSTITUTIVE": "NUCLEUS AND CHROMOSOME, WITH A CENTROMERE-RESTRICTED POOL FROM PROPHASE - and every one of those calls in UniProt is ECO:0000250, INFERRED BY SIMILARITY FROM MOUSE Q9CW03. UniProt Q9UQE7 (fetched live 2026-09-21) carries three location values, 'Nucleus', 'Chromosome' and 'Chromosome, centromere', and NOT ONE of them, nor the location note, carries an ECO:0000269 or any human experimental attribution at all [E_q91_smc3_loc1]. For a subunit of the ring that holds sister chromatids together this is a striking hole: the human protein's compartment is asserted entirely from the mouse orthologue.",
  "CELL_CYCLE_SCOPED": "The note, also by similarity: SMC3 associates with chromatin; BEFORE PROPHASE it is scattered ALONG CHROMOSOME ARMS; during prophase most cohesin dissociates from chromatin, probably through PLK phosphorylation, EXCEPT AT CENTROMERES where it remains; at anaphase RAD21 cleavage releases the complex from chromosomes [E_q91_smc3_loc1]. So 'where is SMC3' has four different answers within one cell cycle and no single resting-state answer is correct.",
  "direct_mammalian_imaging_that_does_exist": "Immunofluorescence of rodent testis sections and spermatocyte nuclear spreads shows intense chromatin association of SMC1 and SMC3 in meiotic prophase cells, weaker staining in round spermatids and absence in elongated spermatids, with both proteins localised IN A BEADED STRUCTURE ALONG THE AXIAL ELEMENTS of pachytene and diplotene synaptonemal complexes [E_q91_smc3_loc2]. That is a real image of SMC3 on chromatin - in meiosis, in rodent germ cells."
}
cell type
{
  "measured_in_for_localisation": "RODENT TESTIS - rat and mouse spermatocytes, round and elongated spermatids [E_q91_smc3_loc2]. The human compartment calls are transferred from mouse [E_q91_smc3_loc1].",
  "in_this_disease": "The locus-level SMC3 data are from patients: the 256 break-prone SMC3 binding sites were defined in the NORMAL cells of KMT2A-rearranged AML patients against non-fusion AML patients (this entry's E2). The functional test was murine MLL-AF9 haematopoietic stem and progenitor cells, where Smc3 haploinsufficiency was tolerated (this entry's E3).",
  "never": "SMC3 protein has never been imaged in a human haematopoietic cell of any lineage in anything this beat retrieved, and never in a KMT2A-rearranged blast."
}
cell state
{
  "CELL_CYCLE_IS_THE_STATE": "Interphase, prophase, metaphase and anaphase give four different SMC3 distributions [E_q91_smc3_loc1]. Any bulk measurement in an asynchronous blast population averages them, which matters here because this graph already holds cell-state-unsynchronisable-cell-cycle.",
  "MEIOSIS_IS_A_SEPARATE_STATE_WITH_ITS_OWN_STRUCTURE": "In meiotic prophase SMC1 and SMC3 sit in a beaded array along the axial/lateral elements of the synaptonemal complex - a structure that does not exist in a somatic cell - and the signal then declines through spermatid maturation [E_q91_smc3_loc2]. The best mammalian images of SMC3 are therefore of a structure irrelevant to a leukaemic blast.",
  "DOSE": "Smc3 haploinsufficiency was WELL TOLERATED in MLL-AF9-expressing murine HSPCs with minimal effect on in-vitro self-renewal (this entry's E3), so the state that matters therapeutically - partial cohesin loss in a fusion-bearing cell - has been tested and is uneventful."
}
localization bounded absence
{
  "date": "2026-09-22",
  "tool": "NCBI esearch db=pubmed",
  "run_by": "Q91",
  "queries": [
    {
      "query": "SMC3[tiab] AND (nucleolus[tiab] OR nucleolar[tiab] OR \"nuclear speckle\"[tiab] OR \"nuclear matrix\"[tiab])",
      "querytranslation": "\"SMC3\"[Title/Abstract] AND (\"nucleolus\"[Title/Abstract] OR \"nucleolar\"[Title/Abstract] OR \"nuclear speckle\"[Title/Abstract] OR \"nuclear matrix\"[Title/Abstract])",
      "count": 9,
      "on_topic_as_a_compartment_assignment": 0
    },
    {
      "query": "SMC3[tiab] AND cohesin[tiab] AND prophase[tiab]",
      "querytranslation": "\"SMC3\"[Title/Abstract] AND \"cohesin\"[Title/Abstract] AND \"prophase\"[Title/Abstract]",
      "count": 50,
      "role": "operand control - the cell-cycle redistribution literature is large and real"
    },
    {
      "query": "SMC3[tiab] AND MLL[tiab]",
      "querytranslation": "\"SMC3\"[Title/Abstract] AND \"MLL\"[Title/Abstract]",
      "count": 2
    },
    {
      "query": "SMC3[tiab] AND KMT2A[tiab]",
      "querytranslation": "\"SMC3\"[Title/Abstract] AND \"KMT2A\"[Title/Abstract]",
      "count": 5
    },
    {
      "query": "cohesin[tiab] AND MLL[tiab]",
      "querytranslation": "\"cohesin\"[Title/Abstract] AND \"MLL\"[Title/Abstract]",
      "count": 7,
      "note": "RULE 64 complex-form control. The complex name returns more than the subunit symbol, as rule 64 predicts, and this entry's own two load-bearing sources are in both sets."
    }
  ],
  "reading": "The absence being bounded is an IMAGE of SMC3 in a human haematopoietic or fusion-bearing cell, not the existence of SMC3 data in this disease - there is locus-level data and this entry already carries it. Titles and abstracts only (rule 80). Retired symbols CSPG6, SMC3L1, BAM and BMH were not run.",
  "absence_class": "never-measured"
}
How this could bite us back
  • Cohesin is essential for chromosome segregation and its partial loss is independently leukaemogenic in mouse. SMC3 is not proposed as a target and the tolerability findings here are cited to argue that a research degron is feasible, not that depletion is safe.
  • The one direct test of cohesin loss against a KMT2A fusion was negative, is a preprint, is murine, and used a self-renewal endpoint. Treating it as settling the question in either direction would be over-reading a single unrefereed negative.
  • If cohesin loss and KMT2A fusion converge on HOXA upregulation, then architectural ideas in this disease are a second route to a target the menin inhibitors already address, and would have to beat them rather than complement them.
Still open
  1. Does cohesin depletion change KMT2A rearrangement FREQUENCY? The only direct test against an MLL fusion used self-renewal as its endpoint, which is not what this graph's loop-anchor hypothesis predicts.
  2. SMC3 is deacetylated by HDAC8, and this graph holds an entire acetylation-cycle beat with several HDAC inhibitors in KMT2A-rearranged disease. Does any HDAC inhibitor used in this disease change SMC3 acetylation, cohesin function or loop architecture? No evidence either way was found, and the question has an obvious experiment behind it.
  3. The non-synthetic-lethality preprint argues cohesin mutations are rare in MLL-r AML because both lesions act through HOXA. Is that testable by asking whether cohesin-mutant non-KMT2A AML responds to menin inhibition?
Every source behind this entry (8)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    SMC3 is UniProt Q9UQE7, structural maintenance of chromosomes protein 3, 1217 residues, primary gene name SMC3 with synonyms BAM, BMH, CSPG6 and SMC3L1.

    What this evidence cannot show

    A naming fact. gene_exact:SMC3 AND organism_id:9606 AND reviewed:true returned exactly one row with gene_primary SMC3. The synonym CSPG6 (chondroitin sulfate proteoglycan 6) is a misassignment from the protein's original characterisation as a basement-membrane component and is a live text-mining hazard.

    Evidence that argues against this (1)
    • SMC1A is the sister ATPase arm at 1233 aa against SMC3's 1217 - a 16-residue difference, which is the smallest separation between any paralogue pair in this beat and the least protected by a length check. For this pair the accession is the only reliable discriminator, as it is for the histones. UniProt Consortium 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    256 SMC3 binding sites were preferentially break-prone in normal cells from KMT2A-rearranged AML remission patients relative to non-fusion AML patients; SMC3 sites falling within gene regions were enriched among highly expressed genes; and SMC3 was one of the four features whose intersection defined 59 shared regions, 34 of which had a KMT2A-to-non-fusion break ratio at least equal to that of the KMT2A region itself.

    PMID 34405474DOI
    What this evidence cannot show

    n=3 per arm, adult, treated remission samples, comparator is non-fusion AML rather than healthy donors. SMC3 site definitions come from reference ChIP datasets intersected with the patient break map, not from ChIP in these patients. Full text read via the PMC HTML route after fullTextXML returned HTTP 500.

    Evidence that argues against this (1)
    • SMC3's 256 sites are the smallest of the four feature sets and the sets are collinear by biology, so the SMC3-specific contribution cannot be separated from RAD21's or CTCF's in this design. Nothing in the study perturbs SMC3. Canela A et al. 2019
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.preprint

    Haploinsufficiency of Smc3, and complete loss of Stag2, were well tolerated in MLL-AF9-expressing murine haematopoietic stem and progenitor cells with minimal effect on in-vitro self-renewal; cohesin loss and MLL-AF9 are therefore not synthetic lethal, and the authors propose that the rarity of cohesin mutations in MLL-rearranged AML reflects a common HOXA-upregulation mechanism of action rather than a lethal interaction.

    PMID 38352423DOI
    What this evidence cannot show

    A PREPRINT, not peer-reviewed at the time of reading, read from its abstract. Murine, MLL-AF9, in-vitro self-renewal as the readout. A negative from a short in-vitro assay does not exclude an effect on rearrangement frequency, on in-vivo leukaemogenesis or on a different subunit. The perturbation is haploinsufficiency, not deletion.

    Evidence that argues against this (1)
    • Tolerance of Smc3 haploinsufficiency for SELF-RENEWAL says nothing about the endpoint this graph's loop-anchor hypothesis actually predicts, which is etoposide-induced KMT2A rearrangement frequency and its position within the BCR. Cohesin binding is a stated precondition for TOP2 trapping, so a break-frequency effect could exist with no self-renewal effect at all. Canela A et al. 2019
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 790 cohesin-mutant myeloid-malignancy patients, mutations in RAD21, SMC1A and SMC3 share features with de novo AML, better prognosis and co-occurrence with de novo AML-type lesions, in contrast to STAG2 mutations which are acquired at MDS stage and associate with secondary AML and adverse prognosis.

    Subunit-specific analysis of cohesin-mutant myeloid malignancies reveals distinct ontogeny and outcomes. · Jann JC, Hergott CB, Winkler M, Liu Y, Braun B, Charles A, et al. · Leukemia 2024
    PMID 39033241DOI
    What this evidence cannot show

    Retrospective two-centre adult cohort, 390 with outcome data. KMT2A rearrangement is not a stratifying variable and no KMT2A-specific claim is made by the paper or by this entry from it.

    Evidence that argues against this (1)
    • Grouping three subunits together on clinical features is the opposite move from the one this beat makes in writing one entry per subunit. The paper's own framing is 'suggestive of distinct ontogenies' - a co-mutation-pattern observation, not a demonstration that SMC3 mutations do something SMC1A mutations do not. Mullenders J et al. 2015
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Cohesin binding together with transcriptional activity alone predict the kinetics of TOP2-induced DSBs, and TOP2 chromatin localisation and etoposide trapping require pre-existing cohesin binding.

    Topoisomerase II-Induced Chromosome Breakage and Translocation Is Determined by Chromosome Architecture and Transcriptional Activity. · Canela A, Maman Y, Huang SN, Wutz G, Tang W, Zagnoli-Vieira G, et al. · Mol Cell 2019
    PMID 31202577DOI
    What this evidence cannot show

    Genome-wide modelling in cultured cells at the level of the complex; no subunit is isolated, and the KMT2A BCR is not examined specifically.

    Evidence that argues against this (1)
    • The complex-level requirement does not transfer to a subunit. Depleting SMC3 removes the whole functional ring, so a break-frequency phenotype under SMC3 loss would be a cohesin phenotype and not an SMC3 phenotype - the same naming-versus-measurement problem this beat documents for the TOP2 isoforms. Canela A et al. 2017
  6. E_q91_smc3_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q9UQE7 (SMC3_HUMAN, primary gene SMC3, 1,217 aa, read back live 2026-09-21) carries THREE subcellular location values - 'Nucleus', 'Chromosome' and 'Chromosome, centromere' - and ALL THREE, together with the location note, are attributed solely to ECO:0000250 with source Q9CW03, the MOUSE orthologue. There is no ECO:0000269, no ECO:0000305 and no human experimental attribution of any kind on the localisation of this protein. The note, also by similarity, is cell-cycle-scoped: SMC3 associates with chromatin and before prophase is scattered ALONG CHROMOSOME ARMS; during prophase most cohesin complexes dissociate from chromatin, probably because of phosphorylation by PLK, EXCEPT AT CENTROMERES where cohesin remains; at anaphase RAD21 cleavage releases the complex from chromosomes; and the Ser-1083-phosphorylated form is preferentially associated with unsynapsed chromosomal regions.

    What this evidence cannot show

    A faithful transfer from a well-studied orthologue, and this entry does not dispute the biology - cohesin's behaviour is conserved and the calls are almost certainly right. What is recorded is the EVIDENCE GRADE: this graph's own audits distinguish ECO:0000269 from ECO:0000305, and this is a third and weaker tier again, an inter-species transfer with no human experiment behind it.

    Evidence that argues against this (1)
    • The interesting objection is that this is not a gap in the literature but a gap in the DATABASE, and the two demand opposite responses. Human cohesin has been imaged extensively; the absence here is of a curated attribution, not of an experiment - which makes it absence_class measured-not-curated rather than never-measured for the somatic compartment, and a beat bounding SMC3's localisation by UniProt alone would reach the wrong conclusion about what remains to be done. Contrast the meiotic record, where a direct mammalian image exists and is cited below, and which UniProt also does not attribute. Eijpe M et al. 2000
  7. E_q91_smc3_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Direct mammalian immunofluorescence of SMC3: in rodent testis sections SMC1 and SMC3 show intense chromatin association in MEIOTIC PROPHASE cells, weaker staining in round spermatids and absence from elongated spermatids; in spermatocyte nuclear spreads both proteins localise IN A BEADED STRUCTURE ALONG THE AXIAL ELEMENTS of the synaptonemal complexes of pachytene and diplotene chromosomes, and both are present in rat spermatocytes and enriched in synaptonemal complex preparations.

    PMID 10652260DOI
    What this evidence cannot show

    Rodent, germ line, meiotic. The synaptonemal complex does not exist in a somatic cell, so this image cannot be transferred to a blast; it is cited as the best DIRECT mammalian localisation of SMC3 that exists rather than as evidence about leukaemia. Read from the abstract.

    Evidence that argues against this (1)
    • Reading this as 'SMC3 is a chromosome-axis protein' over-generalises from the one state in which the axis is visible. In a somatic interphase nucleus there is no axial element to bead along, and the UniProt note describes SMC3 as SCATTERED along chromosome arms before prophase - a distribution with no substructure at all. The meiotic image is the sharpest picture available and it is sharp because the substrate is unusual, which is a reason to treat it as a special case rather than the canonical view. This graph's cohesin reasoning is about interphase loop anchors, and there is no image of that state. UniProt Consortium 2026
  8. E_q91_smc3_loc3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence, 2026-09-22, querytranslation read back on every line. SMC3 has no membraneless-compartment assignment: SMC3[tiab] against nucleolus, nucleolar, nuclear speckle and nuclear matrix = 9 records, none a compartment assignment. SMC3 has never been imaged in a KMT2A-rearranged cell: SMC3[tiab] AND MLL[tiab] = 2, SMC3[tiab] AND KMT2A[tiab] = 5, cohesin[tiab] AND MLL[tiab] = 7. Operand control through the identical shape: SMC3[tiab] AND cohesin[tiab] AND prophase[tiab] = 50.

    What this evidence cannot show

    Bounded by titles and abstracts (rule 80); a cohesin immunofluorescence panel inside one of the seven cohesin/MLL papers would not have been seen. The retired and alias symbols CSPG6, SMC3L1, BAM and BMH were NOT run, and CSPG6 in particular indexes an older chondroitin-sulphate-proteoglycan literature under which early human SMC3 work may sit.

    Evidence that argues against this (1)
    • The rule-64 control changes what this absence means rather than confirming it. cohesin[tiab] AND MLL[tiab] returns 7 against the subunit symbol's 2 - the complex name reaches more than three times as much - so a beat that had bounded SMC3 in this disease by its own symbol would have had access to less than a third of the relevant literature and would have reported a stronger absence than is true. The absence of an IMAGE survives both forms; the impression of an empty field does not. Meyer A et al. 2024
Connected to