Oncofusion

start / graph / gene-cxxc1

Proteinmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q5-interactome5

CXXC1 (CFP1) - a CpG-reading COMPASS subunit that binds MLL1 and recruits it to HoxA7, and whose CXXC domain cannot substitute for KMT2A's in a fusion

also called CXXC1 · CFP1 · CGBP · PCCX1 · PHF18 · CpG binding protein · CXXC finger protein 1

CFP1 recognises unmethylated CpG islands - the same kind of DNA address the KMT2A fusion uses - and it physically travels with MLL1 to switch on HOX genes. But when researchers swapped CFP1's CpG-reading domain into the leukaemia fusion in place of KMT2A's own, the fusion stopped causing leukaemia, even though it still bound DNA and still reached the right gene.

The same thing, for a scientist

CXXC1/CFP1 (Q9P0U4, 656 aa) carries a **PHD-type zinc finger at 28-76** and a **CXXC-type zinc finger at 160-209** [E4]. It co-purifies with MLL1, MLL2 and hSet1, co-localises with them in vivo, binds the **HoxA7 promoter**, and its antisense knockdown **diminishes MLL1 recruitment and lowers H3K4 trimethylation at HoxA7**, reducing expression [E1]. That places a CpG-reading protein inside the MLL1 targeting machinery, which is why round 4 flagged the CpG-reader parallel to the fusion's retained CXXC domain as interesting. **The decisive experiment says the parallel does not transfer.** When the MLL CXXC domain was replaced, in the context of the leukemogenic MLL-AF9 fusion, by the CXXC domain of CGBP/CFP1, colony-forming ability and leukaemogenicity were **abrogated** - while the DNMT1 CXXC domain supported robust colony formation and leukaemogenesis comparable to MLL-AF9 itself [E2]. The isolated CGBP CXXC domain does bind unmethylated DNA, and all swapped fusions still targeted the Hoxa9 locus; what distinguished the functional from the non-functional swaps was **which CpG residues were protected from methylation**, with protection preventing H3K9me3 and enabling Hoxa9 expression - a property shared by MLL and DNMT1 CXXC domains and not by CGBP's [E2]. Separately, CXXC1 has a genuine **subnuclear** result: ES cells lacking Cfp1 show **subnuclear mislocalisation of SETD1A and of H3K4me3 into regions of heterochromatin** [E3]. On DepMap CXXC1 ranks 414 of 17,728 genome-wide (-0.1606), but the lineage split **flags it as a myeloid-identity artefact** (myeloid z -0.97, B-lineage z -0.28, OCI-AML3 z -2.29) [E5][E6].

What it looks like
CXXC-type zinc finger protein 1656 aa
PHD (plant homeodomain) zinc fingerPHD (plant homeodomain) zinc finger · residues 28–76CXXC (CpG-binding zinc finger)CXXC (CpG-binding zinc finger) · residues 160–209Coiled coilCoiled coil · residues 422–4740200400600
loading AF-Q9P0U4-F1…
AlphaFold AF-Q9P0U4-F1 ↗
colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)
I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
Resolving the round 4 deferral: written, with the attractive hypothesis closed off

Round 4 deferred CXXC1 with the note that 'rank 414 (-0.161) is a real signal and the CpG-reader parallel to the fusion's retained CXXC domain is interesting; deferred for want of a primary paper tying it to a KMT2A fusion'. Both halves of that note turn out to need revision, in opposite directions.

The primary papers exist. Two of the eight records in the CXXC1 query tie it directly to MLL: a 2008 co-purification and HoxA7 recruitment study 1, and a 2013 domain-swap study performed in the context of the MLL-AF9 fusion 2. The count was 8 in round 4 and is 8 now; the papers were always in it.

The interesting parallel is closed rather than opened by them. The 2013 swap is as clean a negative as this literature produces: put CFP1's CXXC domain where KMT2A's was, and the fusion still binds unmethylated DNA, still reaches Hoxa9, and no longer transforms 2. The graph's existing target-kmt2a-cxxc-dna-interface and mechanism-fusion-dna-targeting-cxxc-athooks entries reason about the retained CXXC domain as a DNA-targeting module. This result says DNA binding and locus targeting are not sufficient - the discriminating property is protection of specific CpGs from methylation, which then prevents H3K9me3 at Hoxa9.

And the dependency signal is a lineage artefact. Rank 414 looked like the strongest number round 4 had on this gene. Split by lineage it is myeloid z -0.97 against B-lineage z -0.28, and the run's own classifier labels CXXC1 a MYELOID-IDENTITY ARTEFACT - one of only two genes in the 23-gene panel so labelled, alongside MED1 6. The genome-wide rank should not be quoted without that split.

What the swap experiment tells the CXXC-domain target programme

This graph carries a CXXC-domain target dossier. The swap result constrains it usefully and is worth stating as a list, because each line changes what a screen would have to measure.

  1. A CXXC domain that binds unmethylated DNA is not thereby functional in the fusion. CGBP's does bind, and the fusion carrying it does not transform 2.
  2. Locus targeting is not the discriminator either. All swapped fusions, including the non-functional ones, still targeted the Hoxa9 locus in primary bone marrow progenitors 2.
  3. The discriminator is protection of specific CpGs from methylation, which prevents H3K9me3 of the locus and enables Hoxa9 expression 2.
  4. So a CXXC-domain inhibitor screened purely on DNA binding could produce compounds that do nothing to the oncogenic function - and, symmetrically, a compound that blocks DNA binding may not be needed if the relevant activity is methylation protection.

The assay implication is concrete: the right functional readout for a CXXC-directed agent is CpG methylation and H3K9me3 at the target locus, not an electrophoretic mobility shift.

The caveats are equally concrete. The work is MLL-AF9 in mouse primary bone marrow progenitors, not KMT2A::AFF1 in human infant B-ALL; domain swaps can fail for folding or spacing reasons that have nothing to do with the mechanism proposed; and 'protected from methylation' is an inference from methylation profiling at one locus.

Cell type, cell state, subcellular and subnuclear localisation, epigenetics

Subnuclear - the strongest axis, and rare in this beat. Murine ES cells lacking Cfp1 are viable but show increased global H3K4 methylation, decreased SETD1A levels, and subnuclear mislocalisation of both SETD1A and H3K4me3 into regions of heterochromatin 3. Structure-function work in the same study shows that either the N-terminal fragment (aa 1-367) or the C-terminal fragment (aa 361-656) restores SETD1A levels, but full-length Cfp1 is required to restrict SETD1A and H3K4me3 to euchromatin - so protein level and correct subnuclear targeting are separable functions requiring different parts of the protein 3. UniProt independently annotates CXXC1 to nuclear speckles 4.

This is the clearest statement in this beat of a chromatin factor whose job is where a mark goes rather than how much of it there is - and it is the same distinction SETD1B's H3K4me3-breadth phenotype makes from a different direction.

Cell type: human cell lines for the MLL1 co-purification 1; murine primary bone marrow progenitors and transplantation for the swap 2; murine ES cells for the localisation 3. No leukaemia patient material and no KMT2A::AFF1 system. Cell state: not established in leukaemia; in ES cells, Cfp1 loss is compatible with viability but not with normal differentiation.

Epigenetic readouts, three of them and all different: H3K4me3 at HoxA7, reduced by CXXC1 knockdown with reduced MLL1 recruitment 1; H3K9me3 at Hoxa9, which appears when CpGs are not protected from methylation 2; and global H3K4 methylation level plus its euchromatin/heterochromatin distribution 3.

Facts and figures
uniprotQ9P0U4
gene symbolCXXC1
protein nameCXXC-type zinc finger protein 1
length AA656
alphafold IDAF-Q9P0U4-F1
accession resolutionUniProt REST query gene_exact:CXXC1 AND organism_id:9606 AND reviewed:true, run 2026-09-19; one reviewed hit; primary gene name read back as CXXC1 and length 656 confirmed from the same fetched record.
alias hazardUniProt lists PHF18 as a synonym, but PHF18 is also a separate human gene; it was excluded from all queries to avoid inflation.
complexSET1A/SET1B COMPASS (CXXC1/CFP1 is a SET1-specific subunit, absent from MLL1-4 complexes in most preparations)
domains noteAll three boundaries copied from the fetched UniProt feature table. The 1-367 and 361-656 fragments used in the euchromatin-restriction study are experimental constructs, not domains, and are deliberately not given lanes.
functionCpG-island-reading COMPASS subunit; recruits and correctly positions SETD1A, and is required in full-length form to restrict SETD1A and H3K4me3 to euchromatin
classificationan MLL1-complex-associated CpG reader whose CXXC domain is NOT functionally interchangeable with the KMT2A CXXC domain in a fusion
interaction evidence classaffinity purification and reciprocal IP; antisense knockdown with promoter ChIP; domain-swap transformation and transplantation assays; ES-cell knockout with imaging
shown for fusionMLL-AF9, as a NEGATIVE - the CGBP CXXC domain in place of the MLL CXXC domain abrogates leukaemogenicity
cell type evidencehuman cell lines (MLL1 co-purification); murine primary bone marrow progenitors and transplantation (swap); murine ES cells (localisation). No leukaemia patient material, no KMT2A::AFF1 system.
cell state evidencenot established in leukaemia; in ES cells Cfp1 loss is compatible with viability
epigenetic readoutH3K4me3 at HoxA7 (reduced by knockdown, with reduced MLL1 recruitment); H3K9me3 at Hoxa9 (appears when CpGs are not protected from methylation); global H3K4 methylation level and its euchromatin/heterochromatin distribution
domains
PHD (plant homeodomain) z…28.0PHD (plant homeodomain) zinc finger: 28CXXC (CpG-binding zinc fi…160.0CXXC (CpG-binding zinc finger): 160Coiled coil422.0Coiled coil: 422
enumeration record counts
(MLL OR KMT2A) AND (CXXC1…8.000(MLL OR KMT2A) AND (CXXC1 OR CFP1 OR CGBP OR PCCX1): 8(CXXC1 OR CFP1) AND SETD1…9.000(CXXC1 OR CFP1) AND SETD1A AND (CpG OR H3K4): 9(CXXC1 OR CFP1) AND (leuk…10.0(CXXC1 OR CFP1) AND (leukemia OR leukaemia): 10
PDB IDS
[
  "3QMB",
  "3QMC",
  "3QMD",
  "3QMG",
  "3QMH",
  "3QMI"
]
aliases retired or alternate
[
  "CFP1",
  "CGBP",
  "PCCX1"
]
complex partners
[
  "gene-setd1a",
  "gene-setd1b",
  "gene-kmt2a"
]
domains — the numbers behind the chart
[
  {
    "name": "PHD (plant homeodomain) zinc finger",
    "start": 28,
    "end": 76
  },
  {
    "name": "CXXC (CpG-binding zinc finger)",
    "start": 160,
    "end": 209
  },
  {
    "name": "Coiled coil",
    "start": 422,
    "end": 474
  }
]
localization
{
  "subcellular": "nucleus (UniProt)",
  "subnuclear": "nuclear speck (UniProt); and functionally, Cfp1 loss mislocalises SETD1A and H3K4me3 into heterochromatin in ES cells - full-length Cfp1 is required to restrict them to euchromatin"
}
essentiality
{
  "depmap_scan": "analyses/depmap-kmt2a-selective-scan-full",
  "pan_essentiality_all_lines": -0.3833,
  "n_all_lines": 1208,
  "mean_kmt2a_rearranged": -0.4878,
  "mean_other_leukemia": -0.3272,
  "kmt2a_selectivity_difference": -0.1606,
  "rank": 414,
  "n_scanned": 17728,
  "percentile": 2.335,
  "lineage_split": {
    "analysis": "analyses/depmap-interactome5-lineage-split",
    "z_kmt2a_myeloid": -0.97,
    "z_kmt2a_b_lineage": -0.28,
    "z_split_gap": -0.69,
    "z_OCI_AML3": -2.29,
    "z_U937": 0.85,
    "verdict": "MYELOID-IDENTITY ARTEFACT - one of only two genes so classified in the 23-gene panel"
  },
  "interpretation": "rank 414 is the best genome-wide number in this beat and it does not survive the lineage split; quote both halves or neither"
}
enumeration record counts — the numbers behind the chart
[
  {
    "query": "(MLL OR KMT2A) AND (CXXC1 OR CFP1 OR CGBP OR PCCX1)",
    "records": 8,
    "records_after_title_check": 2
  },
  {
    "query": "(CXXC1 OR CFP1) AND (leukemia OR leukaemia)",
    "records": 10
  },
  {
    "query": "(CXXC1 OR CFP1) AND SETD1A AND (CpG OR H3K4)",
    "records": 9
  },
  {
    "date": "2026-09-19"
  }
]
round4 deferral resolved
{
  "round4_reason": "rank 414 (-0.161) is a real signal and the CpG-reader parallel to the fusion's retained CXXC domain is interesting; deferred for want of a primary paper tying it to a KMT2A fusion",
  "resolution": "written. Two primary MLL papers were already inside the same 8-record query. The dependency signal is reclassified as a myeloid-identity artefact by the lineage split, and the CpG-reader parallel is CLOSED rather than opened: the CGBP CXXC domain cannot substitute for the MLL CXXC domain in a transforming fusion."
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.4878,
  "chronos_mean_other_leukaemia": -0.3272,
  "difference": -0.1606,
  "pan_essentiality_all_lines": -0.3833,
  "selectivity_rank": 414,
  "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."
}
How this could bite us back
  • A CXXC-domain-directed agent screened on DNA binding could be functionally inert: the CGBP CXXC domain binds unmethylated DNA perfectly well and still fails to support transformation. The functional readout has to be CpG methylation and H3K9me3 at the target locus.
  • Conversely, blocking DNA binding may not be necessary. If the oncogenic property is protection of specific CpGs from methylation, an agent that permits binding but not protection would be sufficient - and no screen in this programme is set up to find one.
  • CXXC1's genome-wide rank of 414 is the kind of number that starts a programme. It is a myeloid-identity artefact on the lineage split, and this project's index disease is infant B-ALL.
  • Full-length CXXC1 is required to keep H3K4me3 in euchromatin; partial inhibition or truncation could redistribute an activating mark into heterochromatin rather than reducing it, which is an unpredictable rather than a graded perturbation.
Still open
  1. Does CXXC1 associate with KMT2A::AFF1, or only with wild-type MLL1 complexes? The 2008 co-purification predates fusion-specific proteomics and used cells with no fusion.
  2. Is the discriminating property really CpG methylation protection, or does the CGBP swap fail for a structural reason? Expression-matched chimeras with methylation profiling at multiple loci would separate these, and the published work tested one locus.
  3. Does CXXC1 loss mislocalise SETD1A in a leukaemic blast as it does in ES cells? If it does, CXXC1 becomes a lever on WHERE H3K4me3 goes, which is the same axis SETD1B's breadth phenotype sits on.
  4. Is the myeloid-only dependency real biology or a lineage artefact? Four B-lineage lines cannot distinguish 'no requirement in KMT2A-rearranged B-ALL' from 'underpowered', and the graph should not treat the artefact label as a finding.
Every source behind this entry (7)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Human CpG binding protein (CGBP/CXXC1) co-purifies with three H3K4-specific methyltransferases, MLL1, MLL2 and hSet1; independent immunoprecipitation of each complex recovers CGBP; CGBP co-localises with them in vivo and binds the promoter of the MLL target gene HoxA7; and antisense knockdown of CGBP diminished recruitment of MLL1 and downregulated H3K4 trimethylation at the HoxA7 promoter, affecting its expression.

    Human CpG binding protein interacts with MLL1, MLL2 and hSet1 and regulates Hox gene expression · Ansari KI, Mishra BP, Mandal SS · Biochimica et Biophysica Acta 1779(1):66-73 2008
    PMID 18082152DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, every identifier copied from that fetched record. **Antisense knockdown in human cell lines with no KMT2A fusion**; 2008 methodology, so the co-localisation is microscopy-level rather than genomic, and the promoter binding is at a single locus. A knockdown that reduces both MLL1 recruitment and H3K4me3 cannot separate cause from consequence at one promoter.

    Evidence that argues against this (1)
    • The same protein's CXXC domain cannot substitute for the MLL CXXC domain in a transforming fusion, which shows that being present in the MLL1 complex and being able to read CpGs does not make CXXC1's DNA-binding module interchangeable with KMT2A's. Presence in a complex is weaker evidence than it looks. Risner LE et al. 2013
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Replacing the MLL CXXC domain, in the context of the leukemogenic MLL-AF9 fusion, with the CXXC domain of CGBP/CFP1 or of MBD1, or with the MBD1 methyl-CpG-binding domain, abrogated colony-forming ability and leukaemogenicity, whereas MLL(DNMT1 CXXC)-AF9 showed robust in vitro colony formation and in vivo leukaemogenesis similar to MLL-AF9. Isolated MLL, DNMT1 and CGBP CXXC domains each bound unmethylated DNA, with differing affinities, while the isolated MBD1 CXXC and MBD1 MBD domains did not. All substituted domains still allowed targeting of the MLL fusions to the Hoxa9 locus in primary bone marrow progenitor cells; the property that distinguished functional from non-functional swaps was protection of specific CpG residues in the Hoxa9 locus from methylation, which prevented H3K9 trimethylation of the locus and enabled Hoxa9 expression - a property shared by MLL and DNMT1 CXXC domains but not by CGBP's.

    Functional specificity of CpG DNA-binding CXXC domains in mixed lineage leukemia · Risner LE, Kuntimaddi A, Lokken AA, Achille NJ, Birch NW, Schoenfelt K, Bushweller JH, Zeleznik-Le NJ · Journal of Biological Chemistry 288(41):29901-10 2013
    PMID 23990460DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, identifiers copied from that fetched record. **MLL-AF9 in murine primary bone marrow progenitors, not KMT2A::AFF1 and not human infant B-ALL.** A domain swap can fail for folding, spacing or expression reasons unrelated to the mechanism proposed, and the abstract does not report expression-matching of the chimeras. 'Protected from methylation' is an inference from methylation profiling at one locus. This is nonetheless the most decisive negative result this beat found.

    Evidence that argues against this (1)
    • CXXC1 is a genuine component of MLL1-containing complexes and its knockdown reduces MLL1 recruitment and H3K4me3 at a HOX promoter, so it is not irrelevant to MLL1 biology; the swap result constrains what its CXXC domain can substitute for, not whether the protein matters. Ansari KI et al. 2008
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Murine embryonic stem cells lacking Cfp1 are viable but show increased global H3K4 methylation, decreased levels of SETD1A, and subnuclear mislocalisation of both SETD1A and H3K4 trimethylation into regions of heterochromatin. Expression of either the N-terminal fragment (aa 1-367) or the C-terminal fragment (aa 361-656) of Cfp1 is sufficient to restore appropriate SETD1A levels, and either DNA-binding activity or association with the Setd1 complex suffices for that; but full-length Cfp1 is required to restrict SETD1A and H3K4 trimethylation to euchromatin.

    CXXC finger protein 1 restricts the Setd1A histone H3K4 methyltransferase complex to euchromatin · Tate CM, Lee JH, Skalnik DG · FEBS Journal 277(1):210-23 2010
    PMID 19951360DOI
    What this evidence cannot show

    Abstract fetched from PubMed and read on 2026-09-19; full text not read, every identifier copied from that fetched record. **Murine embryonic stem cells - not haematopoietic, not leukaemic, and a cell type with unusual chromatin.** 'Subnuclear mislocalisation into regions of heterochromatin' is a microscopy-level statement in the abstract; the imaging method and quantification were not read. The residue ranges 1-367 and 361-656 are the paper's own fragment boundaries, not domain boundaries.

    Evidence that argues against this (1)
    • An ES-cell phenotype transfers poorly to leukaemic blasts, and the graph should not assume CXXC1 performs the same euchromatin-restriction role in a KMT2A-rearranged cell. It is recorded because it is the only genuine subnuclear result available for any protein in this beat's COMPASS block. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q9P0U4 (CXXC1_HUMAN, CXXC-type zinc finger protein 1) is 656 aa with synonyms CFP1, CGBP, PCCX1 and PHF18; it is annotated to nucleus and nuclear speck; it has six PDB cross-references (3QMB, 3QMC, 3QMD, 3QMG, 3QMH, 3QMI); and it carries a PHD-type zinc finger at 28-76, a CXXC-type zinc finger at 160-209 and a coiled coil at 422-474.

    UniProtKB Q9P0U4 (CXXC1_HUMAN), CXXC-type zinc finger protein 1 · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
    What this evidence cannot show

    Record fetched 2026-09-19 via UniProt REST with gene_exact:CXXC1 AND organism_id:9606 AND reviewed:true; **primary gene name read back as CXXC1 and length 656 confirmed from the same fetched record**, per the standing accession rule. One reviewed hit. The listed synonym PHD18 - given as PHF18 in the record - is a collision hazard with the separate PHF18 gene and was not used in any query.

    Evidence that argues against this (1)
    • The six PDB entries are of the isolated CXXC domain, so there is no structure of CXXC1 in a COMPASS context and nothing structural constrains how it positions SETD1A. The euchromatin-restriction function requires the full-length protein, precisely the form for which no structure exists. Tate CM et al. 2010
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    In the genome-wide KMT2A-rearranged selectivity scan of 17,728 genes, CXXC1 has a pan-essentiality of -0.3833 across 1,208 lines and a KMT2A-rearranged difference of -0.1606 (KMT2A-rearranged mean -0.4878 against other-leukaemia mean -0.3272), ranking 414 (2.335th percentile).

    Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
    What this evidence cannot show

    Row read out of analyses/depmap-kmt2a-selective-scan-full/results.json on 2026-09-19 in a scripted pass. Eleven KMT2A-rearranged against 43 other leukaemia lines, AML-heavy, none infant-derived. **This number must not be quoted without the lineage split in E6, which reclassifies it.** Rank 414 is the strongest genome-wide number for any gene written in this beat and it does not survive being split.

    Evidence that argues against this (1)
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    On the lineage-split run for this beat, CXXC1 is one of only two genes in a 23-gene panel classified as a MYELOID-IDENTITY ARTEFACT (the other is MED1): KMT2A-rearranged myeloid z -0.97, B-lineage z -0.28, split gap -0.69, OCI-AML3 z -2.29, U-937 z +0.85, on a pan-essentiality of -0.384.

    Interactome round 5 panel, lineage-split DepMap dependency (FACT, NSL, MSL, COMPASS remainder, Mediator) · Broad Institute DepMap / Open Targets Platform; analyses/depmap-interactome5-lineage-split 2026
    What this evidence cannot show

    Rows read from analyses/depmap-interactome5-lineage-split/results.json on 2026-09-19. Six myeloid and four B-lineage index lines, so a single line moves a mean; the split is by lineage LABEL rather than expression signature. The OCI-AML3 value of z -2.29 comes from one cell line and is reported for reference, not as a dependency claim.

    Evidence that argues against this (1)
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    As of 2026-09-19, PubMed returns 8 records for '(MLL OR KMT2A) AND (CXXC1 OR CFP1 OR CGBP OR PCCX1)' and 10 for '(CXXC1 OR CFP1) AND (leukemia OR leukaemia)'. Title-checking the 8: two tie CXXC1 directly to MLL (the 2008 MLL1/MLL2/hSet1 co-purification and the 2013 CXXC domain-swap study); the remainder are a CXXC structural-determinants survey, a ZF-CxxC review, a fungal COMPASS paper, an H2B-ubiquitylation study of KMT2 complexes, an FBXL19 in silico characterisation and an Mll2 embryonic-stem-cell bivalency study.

    What this evidence cannot show

    Counts from the esearch `count` field, queried 2026-09-19, with **every record in the 8-record query title-checked individually**. Symbol set: CXXC1, CFP1, CGBP, PCCX1. Deliberately not searched: PHF18, which UniProt lists as a synonym but which is also a separate gene, and which would inflate. Round 4 recorded the same count of 8 for this query, so the difference between that beat's conclusion and this one is a reading difference, not a retrieval difference.

    Evidence that argues against this (1)
    • Round 4 deferred this gene 'for want of a primary paper tying it to a KMT2A fusion' on a query returning the same 8 records that contain such a paper. Counting is not reading, and an identical count supported opposite conclusions in two cycles - which argues that enumeration ledgers should record the title-checked relevant set, not just the integer. Risner LE et al. 2013
Connected to