Oncofusion

start / graph / gene-meaf6

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q126-the-partners-named-but-never-written

MEAF6 - a 191-residue subunit shared between the NuA4/TIP60 and MOZ/MORF acetyltransferase complexes, whose only two UniProt localisations both cite papers that are not localisation studies

also called MYST/Esa1 associated factor 6 · EAF6 · hEAF6 · Esa1-associated factor 6 homolog · CENP-28 · NY-SAR-91 · sarcoma antigen NY-SAR-91 · C1orf149 (PREVIOUS OFFICIAL SYMBOL, not an alias)

MEAF6 is a very small protein that sits inside more than one of the machines that acetylate histones and switch genes on. In a rare uterine tumour and a rare soft-tissue tumour it gets fused to PHF1, a component of the opposite, gene-silencing machinery.

The same thing, for a scientist

**MEAF6 is UniProt Q9HAF1 (EAF6_HUMAN, 191 aa), HGNC:25674 `MYST/Esa1 associated factor 6` at 1p34.3, primary Ensembl ENSG00000163875; `gene_exact:MEAF6 AND organism_id:9606 AND reviewed:true` returned exactly one row, and its PREVIOUS official symbol is `C1orf149`** [E1][E2]. At 191 residues it is the smallest gene entry in this beat. **It is a shared subunit, which is the whole reason it matters here.** EAF6 is part of the tetrameric MOZ/MORF acetyltransferase complexes with ING5 and BRPF1/2/3, where BRPF bridges the acetyltransferase to ING5 and EAF6 and the three form a trimeric core conserved from *Drosophila* to human [E3]; and the ESS-fusion literature names its conserved region the **'histone acetyltransferase subunit NuA4 domain'** [E4]. A protein that belongs to two acetyltransferase families at once is exactly the kind of item a complex-defined fusion family is built out of. **The fusion.** In an endometrial stromal sarcoma with t(1;6)(p34;p21) as the sole karyotypic anomaly, **MEAF6 exon 5 is fused in frame to PHF1 exon 2**, producing a predicted 750-residue protein containing MEAF6's NuA4 domain and PHF1's Tudor, PHD zinc-finger and MTF2 domains [E4]. MEAF6-PHF1 was independently found in an **ossifying fibromyxoid tumour** by RNA sequencing, in a series where PHF1 rearrangement occurs in 80% of cases [E5] - so this is not an ESS-specific fusion, which is a refinement to how `concept-low-grade-ess-is-a-nua4-to-polycomb-fusion-family` reads and is recorded in `gene-phf1`. **Dependency.** Over 1,258 DepMap screens MEAF6's mean Chronos gene effect is **-0.3565 with 24.3% of lines below -0.5 and 0.3% below -1.0** - very close to EPC1 (-0.336, 22.7%) and to MEN1 (-0.3616, 28.9%) on comparable calls [E6].

Go deeper
Both of MEAF6's UniProt localisations point at papers that are not about localisation - rule 91, twice on one 191-residue protein

UniProt Q9HAF1's location block reads: 'Nucleus, nucleolus {ECO:0000269|PubMed:18794358}. Chromosome, centromere, kinetochore {ECO:0000269|PubMed:20813266}.' Both carry the experimental evidence code. This beat fetched both papers and neither is a localisation study of MEAF6 1.

annotationcited paperwhat the paper actually is
Nucleus, nucleolusPMID 18794358Molecular architecture of quartet MOZ/MORF histone acetyltransferase complexes (Ullah et al., MCB 2008) - a complex-reconstitution and interaction-mapping paper 3
Chromosome, centromere, kinetochorePMID 20813266The protein composition of mitotic chromosomes determined using multiclassifier combinatorial proteomics (Ohta et al., Cell 2010) - a proteomics plus machine-learning classifier paper that explicitly predicts 'up to 97 new centromere-associated proteins remain to be discovered in our data set'

The second one is the sharper case, and it explains the alias. MEAF6's alias CENP-28 is a centromere-protein name, and the 2010 paper is where the CENP numbering of predicted chromosomal proteins comes from. So the alias, the kinetochore annotation and the evidence code all trace to one classifier output, not to an imaging or biochemical demonstration that MEAF6 is at kinetochores. The paper did GFP-validate 30 of 34 predicted chromosomal proteins, but this beat did not establish that MEAF6 was among the validated ones and therefore does not claim it was.

Consequence for this graph, stated plainly: gene-meaf6 records no experimentally established subnuclear localisation. That is a worse position than it looks from the database, and it is the third instance in this one tranche - NUP98's coded citations are TPR papers, NSD1's chromosome call is ECO:0000305 curator inference, and PHF1's localisations are entirely uncoded. Four proteins, four different ways the localisation annotation is not what it appears to be.

What a shared subunit means for the complex-defined-family idea

Q123's concept entry argues that a fusion family can be defined by the complexes its halves belong to rather than by a shared gene. MEAF6 is the member that tests how sharp that definition can be, because MEAF6 is not in one complex. It is in the MOZ/MORF (KAT6A/KAT6B) tetramers with ING5 and BRPF 3, and it is assigned a NuA4 domain and treated as a NuA4-family subunit in the fusion literature 4; concept-low-grade-ess-is-a-nua4-to-polycomb-fusion-family records it as a 'NuA4/HBO1-family subunit'.

So the 5' half of MEAF6-PHF1 does not donate 'the NuA4 address'. It donates a module that several acetyltransferase complexes use, and which of them the chimera actually joins is, as far as this beat found, unmeasured. That is not a defeat for the family rule - it is where the rule becomes testable. If MEAF6-PHF1 docks into MOZ/MORF rather than NuA4, the family is 'MYST-acetyltransferase-to-Polycomb' and not 'NuA4-to-Polycomb', and the two make different predictions about which acetyltransferase inhibitor would matter.

And it bears directly on the retrieval problem the concept entry names. Q123 wrote that the ESS family is 'structurally invisible to the way this graph searches' because no gene-symbol query reaches it. MEAF6 shows a second layer of the same problem: EAF6[tiab] AND (HBO1 OR MOZ OR MORF OR NuA4)[tiab] returns 9 records while MEAF6[tiab] with the same operands returns 6, and the two sets are not nested - the complex literature uses EAF6 and the fusion literature uses MEAF6. A beat searching one symbol sees one literature. C1orf149 is a third string, and it is the previous official symbol, not an alias, so rule 69 applies.

Facts and figures
uniprotQ9HAF1
length AA191
gene primaryMEAF6
hgncHGNC:25674
hgnc nameMYST/Esa1 associated factor 6
ensemblENSG00000163875
cytoband1p34.3
accession resolutiongene_exact:MEAF6 AND organism_id:9606 AND reviewed:true returned exactly one row, Q9HAF1, primary gene name MEAF6; HGNC:25674 independently carries Q9HAF1 and ENSG00000163875. Both fetched live 2026-09-23. Rule 62 does not fire.
coverage check verdict pasted not paraphrased`python scripts/is-symbol-covered.py MEAF6`, run 2026-09-23 immediately before this entry was written, one symbol per invocation: MEAF6 no match under 9 candidate names
complex membershipSHARED between acetyltransferase families, which is the point of the entry. Documented as a subunit of the tetrameric MOZ/MORF (KAT6A/KAT6B) complexes with ING5 and BRPF1/2/3, where its association is bridged by BRPF and assisted by ING5 (PMID 18794358); and carrying a 'histone acetyltransferase subunit NuA4 domain' and treated as a NuA4-family subunit in the fusion literature (PMID 22761769). WHICH complex the MEAF6-PHF1 chimera joins is, as far as this beat found, unmeasured.
subcellular localizationnucleus
subnuclear localizationNO EXPERIMENTALLY ESTABLISHED SUBNUCLEAR LOCALISATION was found in this beat. UniProt annotates nucleolus and centromere/kinetochore with ECO:0000269 codes, but the two cited papers are a MOZ/MORF complex-architecture paper and a mitotic-chromosome proteomics/machine-learning paper respectively - neither is a MEAF6 localisation study, and the CENP-28 alias traces to the same classifier output. Recorded as UNESTABLISHED rather than as nucleolar.
epigenetic readouthistone acetylation, indirectly: MEAF6 is a non-catalytic subunit of MYST-family histone acetyltransferase complexes (MOZ/MORF, and NuA4/HBO1-family by the fusion literature's assignment). It has no catalytic activity of its own and this beat found no measurement of a mark attributable to MEAF6 specifically.
fusion junctionMEAF6 exon 5 to PHF1 exon 2, in frame; predicted 750-residue protein retaining MEAF6's NuA4 domain and the ENTIRE PHF1 coding region (PMID 22761769)
retrieval trapFOUR strings for one gene and the literatures are split by which one they use. MEAF6 (current), C1orf149 (PREVIOUS OFFICIAL SYMBOL, rule 69), EAF6/hEAF6 (the complex literature's string), CENP-28 (from a mitotic-chromosome classifier), NY-SAR-91 (cancer-antigen literature). Measured this beat: EAF6[tiab] with MYST-complex operands returns 9 records, MEAF6[tiab] with the same operands returns 6, and the sets are not nested. No row exists in data/symbol-collisions.json for any of them.
relationship to existing entriesconcept-low-grade-ess-is-a-nua4-to-polycomb-fusion-family names MEAF6 in its membership table as a 'NuA4/HBO1-family subunit'. gene-phf1 (this tranche) holds its fusion partner and the OFMT extension of the family. gene-epc1 (Q123) holds the other small NuA4 5' donor. This entry holds MEAF6 itself: accession, the four-string retrieval trap, the shared MOZ/MORF-and-NuA4 membership that makes 'the NuA4 address' underdetermined for this fusion, the exon-level junction, and the double rule-91 defect in its localisation annotation.
alias symbol
[
  "NY-SAR-91",
  "FLJ11730",
  "Eaf6",
  "CENP-28"
]
prev symbol
[
  "C1orf149"
]
fusions donating this gene as the 5 prime partner
[
  "MEAF6-PHF1, t(1;6)(p34;p21), endometrial stromal sarcoma (PMID 22761769)",
  "MEAF6-PHF1, ossifying fibromyxoid tumour (PMID 24285434)"
]
essentiality
{
  "depmap_source": "Open Targets GraphQL target.depMapEssentiality, fetched 2026-09-23",
  "depmap_release": null,
  "open_targets_dataVersion": "year 26, month 06, iteration null",
  "n_screens": 1258,
  "chronos_mean": -0.3565,
  "frac_below_minus_0_5": 0.2432,
  "frac_below_minus_1_0": 0.0032,
  "chronos_median": -0.3505
}
How this could bite us back
  • Nothing here supports an intervention and the acetyltransferase framing should not be allowed to imply one. If a reader reasons from 'MEAF6 is a NuA4 subunit' toward a KAT5 or KAT6A inhibitor, they are targeting a catalytic subunit far more essential than MEAF6 itself, on the basis of a complex assignment this entry has just shown to be underdetermined for the fusion.
  • MEAF6's own dependency profile (24.3% of lines below -0.5) sits in the same range as MEN1's, which in this project is the profile of a real but non-pan-essential target. That similarity is a coincidence of a whole-panel mean and must not be read as an argument that MEAF6 is comparably tractable: there is no structure, no ligand, no probe, and the protein is 191 residues of mostly bridged protein-protein interface.
  • The kinetochore annotation, if it is ever substantiated, would be a mitotic liability of exactly the kind gene-nup98 records for RAE1-NUP98 - and it currently sits in the database as a fact with an experimental evidence code and no experiment behind it that this beat could find. An agent designing against MEAF6 on the strength of 'it is a chromatin protein' would not see it either way.
Still open
  1. Which acetyltransferase complex does MEAF6-PHF1 actually join - NuA4/TIP60, HBO1, or MOZ/MORF? MEAF6 is documented in MOZ/MORF tetramers and is assigned a NuA4 domain by the fusion literature. The family rule in concept-low-grade-ess-is-a-nua4-to-polycomb-fusion-family says 'N-terminal NuA4 component'; for this member that is an assignment by domain name, not by measurement. One FLAG-IP of the chimera would settle it and would either sharpen the rule or rename it 'MYST-to-Polycomb'.
  2. Does MEAF6's bridged, ING5-assisted, explicitly WEAK association with BRPF survive in a fusion that keeps only exons 1-5? The complex paper reports that the BRPF1-EAF6 contact is weak and needs ING5. A 5' fragment of an already weakly-binding bridged subunit may not dock into anything, in which case the 5' half of MEAF6-PHF1 is not delivering a complex address at all and the family rule fails for this member.
  3. Is MEAF6 at kinetochores? The CENP-28 alias, the UniProt annotation and the evidence code all trace to one 2010 machine-learning classifier over mitotic-chromosome proteomics. That paper GFP-validated 30 of 34 predicted chromosomal proteins; this beat did not establish whether MEAF6 was among them. Reading that paper's validation list is an afternoon and would either give MEAF6 a real subnuclear localisation or remove one from the database's face value.
  4. How many MEAF6-PHF1 cases exist? Two published reports in this beat's reading - one ESS and one OFMT index case - and the OFMT abstract does not state how many the cohort screen found. A fusion supporting a family rule on an n of about two should be labelled that way whenever the rule is invoked.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    MEAF6 is UniProt Q9HAF1 (EAF6_HUMAN), a 191-residue 'Chromatin modification-related protein MEAF6' also named MYST/Esa1-associated factor 6, hEAF6, CENP-28 and Sarcoma antigen NY-SAR-91, with a subcellular-location block reading 'Nucleus, nucleolus {ECO:0000269|PubMed:18794358}. Chromosome, centromere, kinetochore {ECO:0000269|PubMed:20813266}.' Both cited papers were fetched in this beat and neither is a localisation study of MEAF6: PMID 18794358 is a MOZ/MORF complex reconstitution and interaction-mapping paper, and PMID 20813266 is a mitotic-chromosome proteomics paper whose centromere assignments are machine-learning predictions.

    What this evidence cannot show

    Fetched live 2026-09-23 via gene_exact:MEAF6 AND organism_id:9606 AND reviewed:true, exactly ONE row. The accession, length and gene names are established. THE LOCALISATION PART OF THIS CLAIM IS A STATEMENT ABOUT WHAT UNIPROT SAYS AND WHAT THE CITED PAPERS ARE, not a claim that MEAF6 is or is not in the nucleolus or at kinetochores. Both cited titles and abstracts were fetched by efetch this session; neither full text was read, so it remains possible that a localisation experiment sits inside one of them that the abstract does not mention - which is exactly rule 80's shape and is the honest limit of this check. What can be said is that neither paper is ABOUT MEAF6's localisation, so UniProt's pointer should not be cited as though it were.

    Evidence that argues against this (1)
    • Only the abstracts of the two cited papers were read, so the conclusion that neither contains a MEAF6 localisation experiment is bounded by abstracts - which is rule 80's exact failure mode, and this project has already paid for it once with an H3K79me2 ChIP that existed only in Figure 6A. A nucleolar or kinetochore image could sit in a supporting panel of either paper. The claim as stated is that neither paper is ABOUT MEAF6 localisation, which is solid; the stronger reading, that no such experiment exists, is not established.
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    HGNC:25674 assigns the symbol MEAF6 the name 'MYST/Esa1 associated factor 6' at 1p34.3, with primary Ensembl gene id ENSG00000163875, UniProt Q9HAF1, alias symbols NY-SAR-91, FLJ11730, Eaf6 and CENP-28, and PREVIOUS SYMBOL C1orf149.

    HGNC record for MEAF6 (HGNC:25674) · HUGO Gene Nomenclature Committee · HGNC REST 2026
    What this evidence cannot show

    Fetched live 2026-09-23; exactly one document returned. RULE 69: C1orf149 is a PREVIOUS OFFICIAL SYMBOL, not an alias, and is where the earliest literature is indexed. Measured in this beat: the complex literature and the fusion literature use DIFFERENT strings - EAF6[tiab] with MYST-complex operands returns 9 records, MEAF6[tiab] with the same operands returns 6, and the sets are not nested. Any bounded absence about MEAF6 must union EAF6, MEAF6 and C1orf149. None of the four strings has a row in data/symbol-collisions.json as of 2026-09-23.

    Evidence that argues against this (1)
    • The EAF6-versus-MEAF6 comparison here is a COUNT comparison, and this project's own rules say to compare ID SETS rather than counts when two naming variants are in play. 9 records against 6 records tells you the totals differ; it does not establish that the sets are non-nested, which is what the entry asserts. That assertion is the weakest sentence in this entry and it was not verified by intersecting the two id lists.
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MOZ and MORF form tetrameric complexes with ING5, EAF6 and the bromodomain-PHD finger protein BRPF1, -2 or -3; BRPF proteins bridge the association of MOZ and MORF with ING5 and EAF6, with an N-terminal BRPF1 region binding the acetyltransferases and the enhancer-of-polycomb (EPc) homology domain in the middle binding ING5 and EAF6; the BRPF1-EAF6 association is weak but ING5 increases the affinity; and these three form a trimeric core conserved from Drosophila to humans.

    Molecular architecture of quartet MOZ/MORF histone acetyltransferase complexes · Ullah M, Pelletier N, Xiao L, Zhao SP, et al. · Mol Cell Biol 2008
    PMID 18794358DOI
    What this evidence cannot show

    Abstract read this session; full text not read, author list is the four in the fetched record plus 'et al.'. Reconstituted complexes and deletion mapping, not endogenous stoichiometry. Note the detail that the BRPF1-EAF6 association is explicitly described as WEAK and ING5-dependent - so MEAF6's membership in these complexes is bridged and affinity-assisted rather than direct, which matters for any claim that a MEAF6 fusion 'docks into' an acetyltransferase complex. This is also the paper UniProt cites for MEAF6's nucleolar localisation, and its abstract describes no localisation experiment at all.

    Evidence that argues against this (1)
    • This paper establishes MEAF6's membership in MOZ/MORF complexes by reconstitution and deletion mapping, not by endogenous stoichiometry, and it explicitly reports the BRPF1-EAF6 association as WEAK and ING5-dependent. MEAF6's other claimed membership - NuA4 - rests on a domain name in the fusion literature and not on this paper at all. So the 'shared subunit' framing this entry builds on combines one reconstituted, bridged, weak association with one annotation-level assignment, and those are different evidence classes being added together.
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In an endometrial stromal sarcoma carrying t(1;6)(p34;p21) as the sole karyotypic anomaly, MEAF6 from 1p34 is recombined with PHF1 from 6p21; exon 5 of MEAF6 is fused in frame to exon 2 of PHF1, so the entire PHF1 coding region becomes the 3' terminal part of a predicted 750-residue fusion protein containing the histone acetyltransferase subunit NuA4 domain of MEAF6 and the Tudor, PHD zinc-finger and MTF2 domains of PHF1. The authors propose that rearrangement of genes involved in acetylation (EPC1, MEAF6) and methylation (PHF1) is a common theme in ESS pathogenesis.

    Novel fusion of MYST/Esa1-associated factor 6 and PHF1 in endometrial stromal sarcoma · Panagopoulos I, Micci F, Thorsen J, Gorunova L, et al. · PLoS One 2012
    PMID 22761769DOI
    What this evidence cannot show

    Abstract read this session; full text not read, author list is the four in the fetched record plus 'et al.'. A SINGLE TUMOUR. The 750-residue protein is PREDICTED from the sequenced junction (5'-RACE, RT-PCR, sequencing); no protein was detected or characterised, and no functional experiment was performed. The authors' own framing is explicitly uncertain: 'Although the specific functions of the MEAF6 and PHF1 proteins and why they are targeted by a neoplasia-specific gene fusion are not directly apparent'. NOTE that this paper's acetylation/methylation framing PREDATES the 2019 NuA4 assignment of JAZF1 by seven years and reaches a similar conclusion independently, which is a point in favour of the family rule that concept-low-grade-ess-is-a-nua4-to-polycomb-fusion-family does not currently make.

    Evidence that argues against this (1)
    • One tumour, and the fusion protein was predicted from a cDNA junction rather than detected. The paper performs no functional experiment and the authors state plainly that why these two genes are targeted is 'not directly apparent'. Everything this entry says about what MEAF6 donates to the chimera is therefore reasoning from a predicted sequence.
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    MEAF6-PHF1 was identified by RNA sequencing in an ossifying fibromyxoid tumour, in a cohort of 39 OFMT in which PHF1 rearrangement is the most common abnormality at about 80% of cases, with fusion to EP400 in 44%; the paper's title states that this molecular characterisation shows genetic overlap with endometrial stromal sarcoma.

    PMID 24285434DOI
    What this evidence cannot show

    Abstract read this session; full text not read, author list is the four in the fetched record plus 'et al.'. MEAF6-PHF1 was found in ONE index case (OFMT3) by RNA-seq and then screened across the cohort; the abstract does not state how many additional MEAF6-PHF1 cases the screen found, so no frequency for MEAF6-PHF1 in OFMT is stated here. Referral-centre cohort of a rare tumour; the partner proportions are not population frequencies.

    Evidence that argues against this (1)
    • MEAF6-PHF1 was found in ONE index case by RNA sequencing, and the abstract does not report how many further cases the cohort screen identified - so no frequency for MEAF6-PHF1 in OFMT can be stated, and this entry does not state one. The two published MEAF6-PHF1 reports between them may describe as few as two tumours, which should be attached to any invocation of MEAF6 as a member of the family.
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    MEAF6's DepMap profile over 1,258 screens is a mean Chronos gene effect of -0.3565, median -0.3505, with 24.3% of lines below -0.5 and 0.32% below -1.0 - close to EPC1 (-0.336, 22.7% below -0.5, recorded in gene-epc1) and to MEN1 (-0.3616, 28.9%) on the same endpoint.

    What this evidence cannot show

    Fetched live 2026-09-23; persisted in analyses/q126-the-partners-named-but-never-written/depmap-essentiality.json (rule 90). RULE 83: meta.dataVersion {year 26, month 06, iteration null} is the OPEN TARGETS platform release, not a DepMap release. Leukaemia strata for MEAF6 are in the JSON. The comparison to EPC1 is drawn from gene-epc1's stored figures, which came from a DIFFERENT call in cycle 83; MEN1 was re-fetched in the same call as MEAF6 here and is therefore the sounder comparator. No ESS or OFMT line is known to this beat to be in the panel. Per rule 84, no median-split or 'anti-selective half' claim is made.

    Evidence that argues against this (1)
    • The EPC1 comparison crosses two different API calls made in two different cycles, which this project's own staleness rules warn against: absolute Chronos values move between releases, so a 0.02 difference between MEAF6 and EPC1 is not interpretable. Only the MEN1 comparator was re-fetched in the same call as MEAF6 and is therefore the only sound one in this claim.
Connected to