CBX3 / HP1-gamma (Q13185, 183 aa) - the EUCHROMATIC HP1 paralogue, and the one CTCF actually prefers: the reader that carries BRCA1/BARD1 onto H3K9-methylated break chromatin, flat in DepMap, and a fusion partner in one KMT2A::AFDN AML patient

No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.

• CBX3 is flat in DepMap (-0.082, 1.1% of 1,258 lines below -0.5) and has two close paralogues, so an HP1-gamma-directed intervention would be both hard to make selective and likely to be buffered - and the paralogue that DOES carry a fitness cost, CBX1 at -0.349, is the one with no role in this cascade. A pan-HP1 reagent would therefore pay CBX1's toxicity to get CBX3's effect.
• Any experiment in this graph that uses the BARD1 PEELI mutant is pan-HP1 by construction and cannot support an HP1-gamma-specific conclusion. Reporting a PEELI result as an HP1-gamma result would be the same category error as reading a W6/32 signal as an HLA-A signal.
• This entry does not license reusing this graph's HP1-alpha findings for HP1-gamma or the reverse. `absence-no-h3k9-writer-or-reader-localised-at-the-kmt2a-locus` bounds CBX5 and remains correct for CBX5; it says nothing about CBX3, and CBX3 is the paralogue the DSB literature names.

1. ANSWERED IN CYCLE 63, AND THE ANSWER IS NO. All 13 released ENCODE BRCA1, CBX3 and CBX1 ChIP-seq tracks were read at seven KMT2A intervals: ZERO called IDR peaks in 13 of 13, from peak sets holding 280-26,789 peaks genome-wide, and the BCR core reads 0.566-1.235 fold over the genome-wide median of same-width tiles. The exon-11 element looked enriched in two BRCA1 tracks (2.05x, top 3.5% in HeLa-S3) and fell to 0.34-1.25x against 6,000 CTCF anchors of the same cell type, below the anchor median in 10 of 13. A conservative-IDR CTCF peak IS on the element in all six biosamples, so the anchor is there and the resection factors are not. See ana-q35-brca1-cbx3-and-cbx1-are-not-at-the-kmt2a-breakpoint-cluster-in-any-encode-cell. Resting state only - the damage state remains unmeasured in any cell type. Original text: AND THIS QUESTION WAS ALREADY PARTLY FALSE WHEN WRITTEN: gene-cbx5, written in cycle 60, records zero called peaks for CBX3 in K562 and GM12878 in its own open_questions. Two entries contradicted each other for two cycles and neither cited the other. Original text: Is CBX3 enriched at the KMT2A exon-11 CTCF element (GRCh38 chr11:118,488,159-118,489,206) in undamaged K562 or GM12878? Four ENCODE datasets exist, two in blood lines, and nobody has looked - this is a compute-only question.
2. Does HP1-gamma's euchromatic preference mean the cascade behaves differently at a transcribed locus such as KMT2A than at the pericentromeric heterochromatin where HP1-alpha and SUV39H1 normally work? The source paper's reporters are integrated transgenes and do not address it.
3. Is the CTCF preference for HP1-gamma over HP1-alpha absolute or quantitative? If quantitative, the paralogue used may be whichever is most abundant in the cell, and abundance in a fetal haematopoietic progenitor is unmeasured.
4. Is the CCDC32/CBX3 fusion recurrent? One patient in one 16-patient series is the entire evidence base, and a second case anywhere would change whether CBX3 belongs in this graph as a partner as well as a reader.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   CBX3 is UniProt Q13185, entry name CBX3_HUMAN, 183 amino acids, primary gene name CBX3 with NO gene synonyms, recommended protein name 'Chromobox protein homolog 3' and alternative names HECH, 'Heterochromatin protein 1 homolog gamma' and 'Modifier 2 protein'; the feature table annotates a Chromo domain at 30-88 and a chromoshadow-subtype Chromo domain at 121-179, and the record cross-references 8 PDB entries; its only coded subcellular location is Nucleus (one ECO:0000269, PubMed:28977666), with an uncoded note stating that it associates with euchromatin and is largely excluded from constitutive heterochromatin and may associate with microtubules and mitotic poles during mitosis.

   UniProtKB Q13185 (CBX3_HUMAN) - Chromobox protein homolog 3 / HP1 homolog gamma, Homo sapiens · UniProt Consortium · UniProt Knowledgebase 2026
   ▸What this evidence cannot show

   A database record fetched 2026-09-22. The euchromatin statement is UNCODED, so it is a curator's summary and not an ECO:0000269 experimental annotation - it is quoted because rule 91 says the uncoded comments carry the condition-scoped statements, not because it has the status of a measurement. The single coded citation for nuclear localisation is a thin evidence base compared with CBX5's eight.

   ▸Evidence that argues against this (1)

   • The euchromatin/heterochromatin split between CBX3 and CBX5 is a UniProt curator's summary of an older imaging literature, and HP1 paralogues redistribute on damage, on differentiation and through the cell cycle. A steady-state compartment assignment cannot be read as where the protein is at the moment of a double-strand break, which is the only moment this beat is about. Hwang SY et al. 2026
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   HP1-gamma is the HP1 paralogue in the CTCF-directed HR cascade, and the assignment was tested against the other two. In U2OS and HEK-293T cells, CTCF recruits HP1-gamma, KAP1, SUV39H1, SETDB1 and KDM5A to enable rapid BRCA1/BARD1 accumulation at DSBs. Among the three HP1 subtypes, CTCF interacts preferentially with HP1-gamma, modestly with HP1-alpha and weakly with HP1-beta by co-immunoprecipitation of endogenous proteins; repeating the comparison with ectopically expressed FLAG-tagged subtypes and HA-tagged CTCF, HP1-gamma interacted robustly despite showing the lowest ectopic expression of the three, excluding an antibody-sensitivity artefact. Endogenous HP1-gamma pulls down full-length CTCF and the ZF fragment (246-588) but not the N-terminus or C-terminus. HP1-gamma is rapidly and transiently recruited to DSBs in a CTCF-reliant manner; HP1-gamma depletion reduces the BARD1-CTCF interaction independently of etoposide; and concurrent ATM inhibition with HP1-gamma depletion does not aggravate impaired end resection or reduced HR efficiency beyond HP1-gamma depletion alone.

   CTCF primes repressive chromatin for rapid BRCA1/BARD1 recruitment to drive homologous recombination repair of DNA double-strand breaks. · Hwang SY, Choi D, Choi MJ, Jo HR, Jeong JH, Kang MA, Lee JS · Nucleic Acids Research 2026
   PMID 42475683DOI
   ▸What this evidence cannot show

   U2OS osteosarcoma reporter lines and HEK-293T only; no haematopoietic cell, no fusion, no KMT2A locus. The paralogue comparison is by co-immunoprecipitation abundance, which is a relative-affinity readout in over-expression and in whole-cell extract, not a quantitative binding measurement - 'preferentially, modestly, weakly' is the paper's own gradation and no Kd is reported. The HP1-gamma antibodies used are Millipore 05-690 and Abcam ab217999; no clone-level cross-reactivity data against HP1-alpha or -beta is given, which for a family of three proteins within 8 residues is the standing worry (the CLONE is the identifier, not the label). DOI taken from PubmedData/ArticleIdList only.

   ▸Evidence that argues against this (1)

   • The BARD1 side of the same cascade is explicitly NON-selective - the PEELI mutant 'largely lacks the ability to bind the CSD of all HP1 subtypes' - so the paralogue specificity rests entirely on the CTCF-HP1 co-IP gradient. If CTCF's preference is quantitative rather than absolute, HP1-alpha could substitute in a cell where it is the abundant paralogue, and the DSB literature separately names HP1-beta in other systems. Hwang SY et al. 2026
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   The three human HP1 paralogues are within 8 residues of each other and carry unevenly indexed synonyms: CBX3 Q13185 183 aa with no gene synonyms; CBX1 P83916 185 aa with the single gene synonym CBX and alternative protein names including 'Heterochromatin protein p25' and 'p25beta'; CBX5 P45973 191 aa with the gene synonym HP1A and the alternative protein name 'Antigen p25'. CBX5 carries coded Chromosome/centromere localisation from three citations and is described as a component of centromeric and pericentromeric heterochromatin; CBX1's note states only that it is unassociated with chromosomes during mitosis; CBX3's note states that it associates with euchromatin and is largely excluded from constitutive heterochromatin.

   UniProtKB Q13185, P83916 and P45973 - the three human HP1 paralogues, fetched in one gene_exact query set · UniProt Consortium · UniProt Knowledgebase 2026
   ▸What this evidence cannot show

   Database records fetched 2026-09-22; each gene_exact query returned exactly one reviewed human record with its primary gene name reading back correctly. The compartment statements for CBX3 and CBX1 are uncoded notes while CBX5's is coded, so the three are not documented to the same standard and the apparent contrast is partly a contrast in curation depth.

   ▸Evidence that argues against this (1)

   • The 8-residue spread across the family means verify-accessions.py cannot catch a CBX1/CBX3 swap (2 residues) and barely catches a CBX3/CBX5 one (8 residues). Length is not a defence in this family, so any downstream entry must carry the accession and rely on gene_primary - the same limitation the PRKAR1A row in the registry records for a pair at zero residues apart. Oncofusion Knowledge Graph identifier registry 2026
4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   CBX3 is not a CRISPR dependency and its paralogue CBX1 is: flattening every tissue block gives CBX3 1,258 screens, mean -0.0817, median -0.0749, minimum -0.6843, maximum 0.6011, with 1.11% of lines below -0.5 and none below -1.0, and a blood mean of -0.1088 over 147 screens. In the same call CBX5 gives mean -0.0830 over 1,258 screens (0.40% below -0.5) and CBX1 gives mean -0.3494 over 1,257 screens with 16.6% below -0.5 and a blood mean of -0.3187.

   Open Targets Platform GraphQL API, target.depMapEssentiality with all tissue blocks flattened, plus meta.dataVersion · Open Targets Platform; Broad Institute DepMap · Open Targets Platform 2026
   ▸What this evidence cannot show

   Rule 83: no DepMap release identifier in the payload; Open Targets data version year 26 month 06. The 1,257-versus-1,258 difference between CBX1 and its paralogues is one screen and is gene-specific variation, not a release difference. Rule 84: no median-split proportion reported. Live zero-screen control fired - POLR2A returned 0 blocks and 0 screens in the same call. A knockout screen in unstressed culture cannot test a damage-conditional function, and three close paralogues are the textbook case where single-gene knockout underestimates pathway importance.

   ▸Evidence that argues against this (1)

   • A flat gene-effect score for a member of a three-paralogue family is uninformative about requirement, because the redundancy is exactly what the screen cannot see. This figure must not be used to argue that HP1-gamma is dispensable for the cascade; the relevant experiment is a triple depletion or a separation-of-function allele, and neither is in DepMap. Open Targets Platform; Broad Institute DepMap 2026
5. E5weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

   CBX3 appears in KMT2A-rearranged AML as a secondary fusion partner, not as a locus-resident protein. In 16 chromosomal-rearrangement AML patients monitored across diagnosis, refractoriness and relapse, RNA-sequencing of a FISH-diagnosed KMT2A (11q23)/AFDN (6q27)-rearranged case identified a previously unknown CCDC32 (15q15.1)/CBX3 (7p15.2) gene fusion present at both diagnosis and relapse; the fusion significantly affected wild-type expression of both CCDC32 and CBX3 at relapse by quantitative PCR; the chimeric fusion persisted over diagnostic and relapse phases in a 21-year-old male with rapidly relapsed refractory AML; and over-expression of the CCDC32/CBX3 fusion in MV4-11 cells promoted cell-cycle progression, which the authors offer as functional validation of a role in treatment resistance.

   Exploring treatment-driven subclonal evolution of prognostic triple biomarkers: Dual gene fusions and chimeric RNA variants in novel subtypes of acute myeloid leukemia patients with KMT2A rearrangement. · Xu Y, Li SC, Xiao J, Liu Q, Cherukuri D, Liu Y, Mirshahidi S, Xu J, Chen X, et al. · Drug Resistance Updates 2025
   PMID 39823827DOI
   ▸What this evidence cannot show

   ONE patient, in a series of 16, in a 21-year-old male - not an infant and not KMT2A::AFF1 but KMT2A::AFDN. Abstract read; full text not fetched. The functional validation is over-expression of the fusion in MV4-11, a KMT2A::AFF1 line, so the patient's fusion context and the model's fusion context differ within the same paper. A cell-cycle-progression readout is not a DNA-repair readout and this paper makes no claim about HR, HP1 function, or the CTCF cascade. Nothing here places CBX3 protein at the KMT2A locus.

   ▸Evidence that argues against this (1)

   • A single recurrent-in-one-patient RNA fusion found by RNA-seq in a 16-patient series, without an independent cohort, is at high risk of being a private event or a read-through artefact; the authors report no second patient carrying CCDC32/CBX3. This entry cites it as the only CBX3-KMT2A-rearranged connection in the literature, which is a statement about the literature's thinness rather than about CBX3's importance. Xu Y et al. 2025
6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   ENCODE holds 4 released TF ChIP-seq experiments for target label CBX3, in GM12878, HCT116 and K562 - two of them blood-lineage cells - so CBX3 occupancy at the KMT2A locus is answerable from public data with compute and no wet work. In the same census CBX1 has 2 (HepG2, K562), CBX5 4 (GM12878, H1, HepG2, K562), BRCA1 7, KDM5A 3 (no blood line), and BARD1, EXO1, DNA2, SUV39H1 and RBBP8 have no ENCODE target record at all.

   ENCODE Portal experiment search, type=Experiment, assay_title=TF ChIP-seq, status=released, target.label=CBX3 · ENCODE Project Consortium · ENCODE Portal 2026
   ▸What this evidence cannot show

   Counted 2026-09-22 on the released TF ChIP-seq facet only; GEO individual depositions, ChIP-Atlas and ReMap were not swept. The count matches Q110's independent census of the same facet one day earlier (4 for CBX3, 2 for CBX1), which is a cross-check rather than an independent confirmation since it is the same endpoint. Crucially, all four experiments are in UNDAMAGED cultured cells, and the cascade this entry is about is damage-conditional and transient - so a re-analysis at the KMT2A locus tests only the resting state.

   ▸Evidence that argues against this (1)

   • Re-analysing these four datasets at the KMT2A locus could return a well-powered zero that is uninformative, for exactly the reason Q108's resting-state H3K9me3 null was: the model predicts the signal appears only after a break. An absence closed this way would need to be labelled as a resting-state absence, not as an absence. Hwang SY et al. 2026

• ← related toCCDC32::CBX3, a previously unreported second fusion found alongside KMT2A::AFDN in a relapsing AML - and the only record in PubMed that pairs CBX3 with KMT2A at all

  the first patient-level event in this graph connecting CBX3/HP1-gamma to a KMT2A rearrangement; wild-type CBX3 expression was altered at relapse

• ← recruitsCTCF - named in the id of an existing entry in this graph and having no entry of its own, and the protein whose binding sites break preferentially in KMT2A-rearranged patients

  CTCF prefers HP1-gamma over HP1-alpha and HP1-beta; ZF fragment 246-588 is the contact

• binds →BARD1 (Q99728, 777 aa) - BRCA1's obligate partner, and the member of the cascade that actually touches the chromatin: its PxVxL motif at 569-573 binds HP1, and CTCF binds BARD1 rather than BRCA1 in an undamaged cell

  HP1-gamma chromoshadow domain binds the BARD1 PxVxL motif at 569-573 (PEELI mutant abolishes)

• binds →H3K9me3 - measured in KMT2A::AFF1 cells, in a supplementary figure, and it rises GENERALLY where H3K27ac falls SPECIFICALLY in the same experiment

  chromo domain 30-88 reads H3K9me2/me3 at the break

• related to →CBX5 / HP1-alpha (P45973, 191 aa) - the H3K9me3 reader that recruits the writer, the writer's destroyer and the second writer; and the one member of that circuit with a published pro-leukaemic role, as part of an MLL leukaemia stem-cell programme that immortalises myeloid progenitors WITHOUT Hoxa or Meis

  paralogue within 8 residues; CBX3 is euchromatic, CBX5 pericentromeric - do not substitute one for the other

• co-occurs with →KMT2A::AFDN (MLL-AF6), t(6;11)(q27;q23)

  CCDC32/CBX3 secondary fusion in one KMT2A::AFDN AML patient at diagnosis and relapse

• ← co-occurs withTRIM28 / KAP1 / TIF1beta (Q13263) - the KRAB-ZFP corepressor scaffold in the pan-MLL-fusion core, with the weakest index-fusion p-value but a leukaemic role in another fusion family

  KAP1 is the scaffold of the CTCF-recruited repressive assembly that HP1-gamma joins

• ← recruitsThe CTCF licensing cascade: CTCF recruits HP1-gamma, KAP1, SUV39H1, SETDB1 and KDM5A to a break, that repressive chromatin licenses EARLY BRCA1/BARD1, and early BRCA1/BARD1 licenses EXO1/DNA2 long-range resection - measured entirely in U2OS and HEK-293T, with engineered breaks and one etoposide arm

  HP1-gamma is the reader limb

• ← related toNo BRCA1, BARD1, EXO1, DNA2, HP1-gamma or KDM5A has ever been localised at the KMT2A locus, in any cell type: 21 queries, every querytranslation read back, 0 on-topic records - and the absence splits THREE ways by class, which is what decides what it costs to close

  measured-not-analysed: 4 ENCODE datasets including K562 and GM12878, never read at this locus

• ← related toDepMap across the whole CTCF-BRCA1 licensing cascade in one call: CTCF is a near-total pan-essential at -1.83, BARD1 is deeper than BRCA1, the HP1 paralogue that matters is flat while the one that does not is four times deeper - and DNA2's row has 80 screens where everything else has 1,258

  CBX3 flat at -0.082 while the uninvolved paralogue CBX1 is four times deeper at -0.349

• ← testsRUN, and the absence closes: all 13 released ENCODE BRCA1, CBX3 and CBX1 ChIP-seq tracks read at the KMT2A locus with three nulls - zero called peaks in 13 of 13, the breakpoint cluster at 0.57-1.24x the genome, and the exon-11 CTCF element BELOW the median CTCF anchor of its own cell type in 10 of 13

  4 released ENCODE CBX3 experiments, 3 passing the dynamic-range gate; no occupancy above background

• ← related toCBX1 / HP1-beta (P83916, 185 aa) - the HP1 paralogue with the fitness cost and no role in this graph's cascade: DepMap mean -0.349 with 16.6% of 1,257 lines below -0.5, four times deeper than HP1-gamma or HP1-alpha, and zero papers place it anywhere near KMT2A

  HP1 paralogue; 185 vs 183 aa, and 4x the DepMap dependency

• ← testsThe HP1 paralogue asymmetry is a guide-cross-reactivity artefact until somebody shows it is not - and if it is real, HP1-beta is doing a job the other two cannot, which is the only HP1 question in this graph with a fitness phenotype attached

  the flat paralogue the asymmetry is measured against

• ← related toThe KMT2A exon-11 CTCF anchor carries a CBX3 peak and a BRCA1 peak in three of the very ENCODE experiments Q35 called empty - the peaks are in ENCODE's own pre-IDR files, present in both replicates, and deleted by the IDR threshold at global IDR 0.227 and 0.059

  CBX3/HP1-gamma peak at chr11:118,488,610-118,488,767 in K-562 (ENCSR000BRT), summit one base from the CTCF summit

• ← related toHP1-gamma holds a constitutive, sub-threshold foothold on the KMT2A exon-11 CTCF anchor in undamaged cells - the CTCF-primed model's missing resting state - and the reason nobody has seen it is that it sits below the reproducibility filter every public peak set applies

  HP1-gamma is the protein the foothold is claimed for

• ← testsRead all 76 archived CBX3 and BRCA1 tracks at the KMT2A exon-11 anchor at SIGNAL level against a CTCF-anchor-matched null, and score the anchor's rank among its own class in each - no wet work, no new data, and it decides whether the sub-IDR peak is a foothold or a threshold artefact

  45 archived CBX3 tracks at the exon-11 anchor against a class-matched CTCF-anchor null