Oncofusion

start / graph / gene-rbm15

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q6-interactome-rna-layer

RBM15 — m6A adaptor, AMKL fusion partner, and a named hit in the only KMT2A-biased RNA-protein CRISPR screen

also called OTT · OTT1 · SPEN family · RNA binding motif protein 15

RBM15 helps decide which messenger RNAs get chemically tagged, and it is itself broken into a fusion gene in a rare childhood leukaemia. A recent screen also named it among the proteins that KMT2A-rearranged myeloid leukaemia needs.

The same thing, for a scientist

RBM15 is an m6A adaptor that recruits the methyltransferase complex to specific transcripts, and is a leukaemia gene in its own right through the RBM15-MKL1 (OTT-MAL) fusion of infant acute megakaryoblastic leukaemia, which promotes leukaemia via m6A methylation and Wnt pathway activation [E1]. It was named alongside PTBP1 as an AML dependency with a bias toward the KMT2A-rearranged subtype in a domain-focused CRISPR screen of 341 RNA-associated proteins [E2]. DepMap gives it a Chronos mean of -0.267 with 18.9% of 1,208 lines depleted and index-arm differences of -0.036 B-lymphoid / -0.115 myeloid, the 10.3rd percentile of a 466-gene null [E3]. Its paralogue RBM15B carries a separate mechanism reading H3K79me2 to steer m6A to start codons, and the two are routinely conflated [E4].

Go deeper
A search count that title-checks almost entirely away

(MLL OR KMT2A) AND RBM15 returns 13 records, which looks like a real literature. Title-checked, it is not one. The records are paediatric acute megakaryoblastic leukaemia cytogenetics and outcome series in which RBM15-MKL1 and KMT2A rearrangements are both listed as recurrent abnormalities, two reviews of neonatal and infant leukaemia, a myeloid-leukaemia classification review, a NUP98-BPTF case report, and one preeclampsia paper about MLL1 in trophoblast. Exactly one is a functional MLL-rearranged study, and it is about RBM15B, not RBM15 5.

The retired alias OTT1 returns 0 against MLL or KMT2A, so alias expansion adds nothing here.

This is the mirror of the FTO case in the same beat: there, a spelled-out alias inflated 3 records to 8 with 6 artefacts. Here, co-occurrence in cytogenetic series inflates a real count of about 1 to 13. Both are why raw counts are not claims.

RBM15 as a fusion partner, and what that does not imply

RBM15-MKL1, historically OTT-MAL, defines a subtype of infant acute megakaryoblastic leukaemia and promotes leukaemia through m6A methylation and Wnt activation 1. That makes RBM15 a bona fide childhood-leukaemia oncogene, and it is tempting to read across.

It does not read across. A gene that is oncogenic when fused is not thereby a dependency when intact, and this graph holds the same distinction for KMT2A itself. What makes RBM15 relevant here is the independent CRISPR-screen result 2, not the fusion.

The fusion does supply one genuine cross-link, though. RBM15-MKL1 acts through m6A 1, which means two distinct infant leukaemias - t(4;11) via the RNA layer as this beat describes it, and RBM15-MKL1 AMKL - converge on m6A from different directions. That is a comparative-fusion-biology observation this graph is set up to hold.

Facts and figures
uniprotQ96T37
length AA977
gene symbolRBM15
uniprot resolutiongene_exact:RBM15 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as RBM15; length 977 aa checked
rna functionm6A adaptor recruiting the methyltransferase complex to specific transcripts; SPEN-family
fusion roleRBM15-MKL1 (OTT-MAL), infant acute megakaryoblastic leukaemia
subnuclear localizationNUCLEAR SPECKLES PLUS THE NUCLEAR ENVELOPE / NUCLEAR PORE - TWO NAMED COMPARTMENTS, BOTH EXPERIMENTALLY ATTRIBUTED, AND THE BEST-RESOLVED SUBNUCLEAR RECORD IN THIS TRANCHE. Speckles: 'Nucleus speckle' on two ECO:0000269 attributions, and the primary description is association with the SPLICING FACTOR COMPARTMENT [E_q91_rbm15_loc1][E_q91_rbm15_loc2]. Envelope and pore: RBM15 is a PERIPHERAL MEMBRANE PROTEIN of the nuclear membrane and colocalises at the NUCLEAR PORE with DBP5 and NXF1 [E_q91_rbm15_loc1][E_q91_rbm15_loc3]. This is directly comparable to two things this graph already holds - localization-nuclear-speckles-kmt2a-r, and the nuclear-pore comparators for NUP98-FG and for NUP214 (amended in this same tranche) - and RBM15 reaches both compartments in one protein. One caution recorded rather than hidden: the speckle attribution PubMed:19586903 is a paper about the PARALOGUE RBM15B/OTT3, in which RBM15 is the comparator arm; see the counter-note on E_q91_rbm15_loc2.
localization shown for fusionNO for KMT2A fusions, in any lineage; and NO for its OWN fusion either, which is the more surprising half. RBM15[tiab] AND MLL[tiab] = 5 and RBM15[tiab] AND KMT2A[tiab] = 7, none a localisation study [E_q91_rbm15_loc5] - and nothing retrieved reports where the RBM15::MKL1 (OTT-MAL) protein of infant AMKL sits, although RBM15's own compartments are well mapped and MKL1/MRTFA is a famously mechanosensitive nucleocytoplasmic shuttler. Under rule 86 the statement is: no fusion of any family has had this protein localised.
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": -0.2669,
  "pct_lines_depleted": 18.9,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": -0.0362,
  "kmt2a_r_myeloid_diff": -0.115,
  "null_percentile": 10.3
}
pubmed counts
{
  "RBM15 AND leukemia": 70,
  "(MLL OR KMT2A) AND RBM15": "13 raw, ~1 after title check and that one is RBM15B",
  "(MLL OR KMT2A) AND OTT1": 0
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.5114,
  "chronos_mean_other_leukaemia": -0.4251,
  "difference": -0.0863,
  "pan_essentiality_all_lines": -0.2669,
  "selectivity_rank": 1524,
  "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": "NUCLEAR, AND UNUSUALLY WELL RESOLVED WITHIN THE NUCLEUS. UniProt Q96T37 (fetched live 2026-09-21) carries FOUR location values, every one experimentally attributed: 'Nucleus speckle' (ECO:0000269 PubMed:19586903, 24100041), 'Nucleus, nucleoplasm' (ECO:0000269 PubMed:17001072, 19786495, 24100041), 'Nucleus envelope' (ECO:0000269 PubMed:19586903) and 'Nucleus membrane, peripheral membrane protein' (ECO:0000269 PubMed:19786495), with the note that RBM15 COLOCALIZES AT THE NUCLEAR PORE WITH DBP5 AND NXF1 [E_q91_rbm15_loc1]. No cytoplasmic value at all.",
  "THE_TWO_POOLS_ARE_ONE_MECHANISM": "The speckle pool and the envelope pool are the two ends of an mRNA export pathway, not two unrelated observations. RBM15 associates with the SPLICING FACTOR COMPARTMENT and with the NUCLEAR ENVELOPE and binds the mRNA export factors NXF1 and Aly/REF through its C-terminal region [E_q91_rbm15_loc2]; and RBM15 is TARGETED TO THE NUCLEAR ENVELOPE where it colocalises extensively with DBP5, binding human DBP5 specifically and facilitating its direct contact with mRNA in vivo [E_q91_rbm15_loc3]."
}
cell type
{
  "measured_in": "HeLa and human umbilical vein endothelial cells for the localisation and complex work [E_q91_rbm15_loc2][E_q91_rbm15_loc3][E_q91_rbm15_loc4]; not haematopoietic.",
  "IN_LEUKAEMIA_THE_CELL_TYPE_IS_SPECIFIC_AND_INFANT": "RBM15 is a leukaemia gene in its own right through the RBM15-MKL1 (OTT-MAL) fusion of INFANT ACUTE MEGAKARYOBLASTIC LEUKAEMIA (this entry's E1), i.e. a MEGAKARYOCYTIC lineage in the same age window as the index disease; and it was named as an AML dependency with a bias toward the KMT2A-rearranged subtype in a domain-focused CRISPR screen of 341 RNA-associated proteins (this entry's E2).",
  "never": "RBM15 protein has never been localised in any leukaemic cell, in either of the two diseases it is implicated in, and the RBM15::MKL1 fusion protein's compartment has never been reported in anything this beat retrieved."
}
cell state
{
  "not_a_redistribution_but_a_pathway_state": "No induced relocalisation of RBM15 is recorded in UniProt or retrieved here. What is state-dependent is the pathway it sits in: RBM15 facilitates DBP5's contact with mRNA at the pore, and DBP5 is the helicase that strips NXF1/Mex67 from mRNP on translocation to give export directionality [E_q91_rbm15_loc3]. RBM15's position is therefore coupled to mRNA export flux rather than to a signalling state.",
  "CELL_CYCLE_ADJACENT": "RBM15 is a subunit of the WTAP complex together with Hakai, Virilizer homolog, KIAA0853, BCLAF1, THRAP3 and general splicing regulators, and that complex is implicated in alternative splicing and the cell cycle [E_q91_rbm15_loc4]. That is the complex through which RBM15's m6A adaptor function runs."
}
localization bounded absence
{
  "date": "2026-09-22",
  "tool": "NCBI esearch db=pubmed",
  "run_by": "Q91",
  "queries": [
    {
      "query": "RBM15[tiab]",
      "querytranslation": "\"RBM15\"[Title/Abstract]",
      "count": 385,
      "role": "operand control on the current symbol"
    },
    {
      "query": "OTT1[tiab]",
      "querytranslation": "\"OTT1\"[Title/Abstract]",
      "count": 11,
      "role": "RETIRED-SYMBOL CONTROL - and it does NOT reproduce the RBM15B/OTT3 pattern; see the counter-note on E_q91_rbm15_loc5"
    },
    {
      "query": "RBM15[tiab] AND (\"nuclear speckle\"[tiab] OR \"nuclear speckles\"[tiab] OR nucleolus[tiab] OR subnuclear[tiab])",
      "querytranslation": "\"RBM15\"[Title/Abstract] AND (\"nuclear speckle\"[Title/Abstract] OR \"nuclear speckles\"[Title/Abstract] OR \"nucleolus\"[Title/Abstract] OR \"subnuclear\"[Title/Abstract])",
      "count": 2,
      "note": "only two records name a compartment in title or abstract, against four experimentally attributed UniProt values - the localisation evidence for this protein is in figures and methods, not in abstracts, which is rule 80 stated as a positive"
    },
    {
      "query": "RBM15[tiab] AND MLL[tiab]",
      "querytranslation": "\"RBM15\"[Title/Abstract] AND \"MLL\"[Title/Abstract]",
      "count": 5
    },
    {
      "query": "OTT1[tiab] AND MLL[tiab]",
      "querytranslation": "\"OTT1\"[Title/Abstract] AND \"MLL\"[Title/Abstract]",
      "count": 0
    },
    {
      "query": "RBM15[tiab] AND KMT2A[tiab]",
      "querytranslation": "\"RBM15\"[Title/Abstract] AND \"KMT2A\"[Title/Abstract]",
      "count": 7
    }
  ],
  "reading": "The absence bounded is a localisation in any leukaemic cell or any fusion. It is emphatically NOT an absence of localisation data for the protein. Titles and abstracts only (rule 80). Two symbols were run; OTT and the SPEN-family phrase were not.",
  "absence_class": "never-measured"
}
How this could bite us back
  • RBM15's headline leukaemia role is as a fusion partner, which is an oncogene argument and not a dependency argument, and the two are routinely elided.
  • RBM15 and RBM15B are conflated across the review literature; a programme citing the H3K79me2 mechanism while targeting RBM15 would be targeting the wrong paralogue.
  • A 977-residue SPEN-family adaptor with no catalytic pocket is not a tractable small-molecule target on current evidence.
Still open
  1. How large is RBM15's KMT2A-rearranged bias in the screen that named it, given that an independent DepMap scan puts it at the 10.3rd percentile against PTBP1's 0.0th?
  2. Do RBM15 and RBM15B buffer each other as m6A adaptors?
  3. Two infant leukaemias - t(4;11) and RBM15-MKL1 AMKL - both implicate m6A. Is that convergence or coincidence?
Every source behind this entry (10)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The RBM15-MKL1 fusion protein promotes leukaemia via m6A methylation and Wnt pathway activation.

    PMID 40435410DOI
    What this evidence cannot show

    Abstract-level. Acute megakaryoblastic leukaemia, not KMT2A-rearranged disease. Establishes RBM15's m6A adaptor function in a leukaemogenic context; says nothing about wild-type RBM15 as a dependency.

    Evidence that argues against this (1)
    • A gene that is oncogenic when fused is not thereby required when intact. This graph holds the same distinction for KMT2A, and reading a fusion partner's oncogenicity as a dependency argument would be the error. Cao Z et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A domain-focused CRISPR screen of 3,182 sgRNAs against 527 RNA-enzymatic and RNA-binding domains in 341 RNA-associated proteins identified the m6A reader RBM15, alongside PTBP1, as a dependency in AML with a bias toward the KMT2A-rearranged subtype.

    PMID 42637878DOI
    What this evidence cannot show

    Abstract read; full text not read. AML. RBM15 is named as a hit but only PTBP1's validation is described in the abstract, so RBM15's evidence within this paper is weaker than PTBP1's and the magnitude of its KMT2A-r bias is not stated.

    Evidence that argues against this (1)
    • This project's independent DepMap scan gives RBM15 an index-arm difference of only -0.036 in the four B-lymphoid lines and -0.115 in the seven myeloid ones, at the 10.3rd percentile - modest, and far from PTBP1's 0.0th percentile in the same scan. The two hits from the same screen behave very differently in an independent dataset. Oncofusion knowledge graph (internal analysis) 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives RBM15 a mean gene effect of -0.267 across 1,208 lines with 18.9% depleted below -0.5, and index-arm differences of -0.036 (four B-lymphoid) and -0.115 (seven myeloid) against 43 other leukaemia lines, at the 10.3rd percentile of a 466-gene null. Its paralogue RBM15B is -0.041 with 0.2% of lines depleted.

    What this evidence cannot show

    Eleven index lines. The difference is driven by the myeloid half and is small; -0.115 against a null SD of 0.075.

    Evidence that argues against this (1)
    • RBM15 and RBM15B may buffer each other as m6A adaptors, in which case both single-gene rows under-report the pair - the same problem this beat records for the IGF2BP, KHDRBS and YTHDF families. It is being invoked often enough in this beat that it should be treated as a pattern in the reasoning as well as in the biology. Oncofusion knowledge graph (internal analysis) 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q96T37 (RBM15_HUMAN) is human RBM15, 977 amino acids, and its paralogue RBM15B is Q8NDT2 at 890 amino acids; both resolved by gene_exact query with the primary gene name read back and the lengths checked.

    What this evidence cannot show

    Two single-hit queries with matching primary gene names. Recorded together because the two paralogues are routinely conflated in reviews and this beat cites a different primary paper for each.

    Evidence that argues against this (1)
    • Separate accessions and different lengths do not establish functional separation. The H3K79me2-reading mechanism is reported for RBM15B and the CRISPR-screen dependency for RBM15, and no published work shows both properties in either protein. Chen TQ et al. 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted, title-checked: `(MLL OR KMT2A) AND RBM15` returns 13 PubMed records of which the large majority are paediatric acute megakaryoblastic leukaemia cytogenetic and outcome series listing RBM15-MKL1 and KMT2A rearrangements side by side, plus reviews of neonatal leukaemia and myeloid leukaemia classification, a NUP98-BPTF case report and a preeclampsia study; exactly one is a functional MLL-rearranged mechanism paper and that paper is about RBM15B. The retired alias `OTT1` returns 0 against MLL or KMT2A.

    PubMed E-utilities esearch counts, alias-expanded and title-checked · NCBI PubMed (queries run 2026-09-20) 2026
    What this evidence cannot show

    Classification by title, not by full text. A cytogenetic series could in principle contain a subtype-specific RBM15 observation its title does not advertise.

    Evidence that argues against this (1)
    • Discounting co-occurrence records is a judgement. The AMKL series are genuinely about paediatric leukaemia genetics and the co-occurrence of these two fusions in the same nosology is itself the comparative observation this entry ends on. Mayday MY et al. 2025
  6. E_q91_rbm15_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q96T37 (RBM15_HUMAN, primary gene RBM15, 977 aa, read back live 2026-09-21) carries FOUR subcellular location values, all experimentally attributed and all nuclear: 'Nucleus speckle' (ECO:0000269 PubMed:19586903, PubMed:24100041), 'Nucleus, nucleoplasm' (ECO:0000269 PubMed:17001072, PubMed:19786495, PubMed:24100041), 'Nucleus envelope' (ECO:0000269 PubMed:19586903) and 'Nucleus membrane' with topology PERIPHERAL MEMBRANE PROTEIN (ECO:0000269 PubMed:19786495). The location note states that RBM15 COLOCALIZES AT THE NUCLEAR PORE WITH DBP5 AND NXF1 (ECO:0000269 PubMed:19786495). There is no cytoplasmic value and no by-similarity or rule-based call anywhere in the subcellular annotation.

    What this evidence cannot show

    Four values from three papers by overlapping groups (two share authors), so the apparent independence is less than the citation count suggests. A peripheral-membrane topology on a 977-residue SPEN-family RNA-binding protein means association, not insertion.

    Evidence that argues against this (1)
    • This record is the strongest in the tranche and the reason is worth stating, because it inverts what a coverage metric would say. RBM15 is not an important protein in this graph - it is a low-degree entry that arrived through an RNA-layer screen - and it has four experimentally attributed compartments, while SPI1, the master haematopoietic transcription factor, has one value resting on a PROSITE rule and a single disease-genetics paper, and SMC3 has three values that are all inter-species transfers. Depth of localisation annotation tracks which subfield adopted the protein - here the mRNA-export field, which images everything - and not the protein's importance. A graph that prioritises localisation work by degree, as this project's top-N-by-degree reporter encourages, will systematically chase the proteins whose records are hardest to complete. UniProt Consortium 2026
  7. E_q91_rbm15_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    RBM15's speckle and envelope pools were characterised alongside its paralogue: the human SPEN family proteins SHARP, RBM15/OTT1 and RBM15B/OTT3 share domain architecture but differ functionally, and OTT3, LIKE RBM15, has post-transcriptional regulatory activity where SHARP does not. OTT3 SHARES WITH RBM15 the association with the SPLICING FACTOR COMPARTMENT AND THE NUCLEAR ENVELOPE, as well as the binding to the mRNA export factors NXF1 and Aly/REF, and mutational analysis located the direct NXF1 interaction of both OTT3 and RBM15 in their C-terminal regions.

    The RNA-binding motif protein 15B (RBM15B/OTT3) acts as cofactor of the nuclear export receptor NXF1. · Uranishi H, Zolotukhin AS, Lindtner S, Warming S et al. · J Biol Chem 2009
    PMID 19586903DOI
    What this evidence cannot show

    The paper's subject is OTT3/RBM15B; RBM15 appears as the comparator against which OTT3 is characterised. 'Splicing factor compartment' is the term used and it is what UniProt renders as 'Nucleus speckle' - a controlled-vocabulary mapping the reader should be aware of. HeLa-type cells, not haematopoietic.

    Evidence that argues against this (1)
    • UniProt cites this paper as an ECO:0000269 for two of RBM15's four compartments, and it is a paper about the paralogue - a rule-91 configuration, and a mild one because RBM15 really is imaged here as the comparator arm rather than merely mentioned. Two things still follow. The compartment assignment for RBM15 in this paper is made in the service of establishing that OTT3 resembles it, which is the weakest experimental position a comparator can occupy; and this graph has a documented history with exactly this pair, since RBM15B's literature sits under OTT3 and this entry's own summary already warns that the two are 'routinely conflated'. The conflation is not only in the literature - it is inside RBM15's own UniProt record. Zolotukhin AS et al. 2009
  8. E_q91_rbm15_loc3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The nuclear-envelope pool has a dedicated source and a mechanism: the human NXF1-binding protein RBM15 binds specifically to human DBP5 and facilitates its direct contact with mRNA in vivo; RBM15 IS TARGETED TO THE NUCLEAR ENVELOPE, WHERE IT COLOCALIZES EXTENSIVELY with DBP5. The context is that NXF1 (Mex67 in yeast) assembles with mRNP in the nucleus and guides it through the nuclear pore complex, and the DEAD-box helicase DBP5 dissociates NXF1 from the mRNP upon translocation to generate directional passage.

    Nuclear export factor RBM15 facilitates the access of DBP5 to mRNA. · Zolotukhin AS, Uranishi H, Lindtner S, Bear J et al. · Nucleic Acids Res 2009
    PMID 19786495DOI
    What this evidence cannot show

    Read from the abstract; the cell system is not stated there. 'Targeted to the nuclear envelope' is not the same as 'is a nucleoporin' - UniProt's peripheral-membrane topology is the right reading. No leukaemia content.

    Evidence that argues against this (1)
    • The two compartments are usually recorded as a list and they are in tension as a steady state. A protein cannot be concentrated in speckles and concentrated at the envelope at the same instant unless it is in flux between them, which is what an mRNA export adaptor should be - and which means every static image of RBM15 is a snapshot of a distribution whose shape depends on export flux at that moment. For this graph the consequence is concrete: if RBM15's compartment tracks export activity, then comparing RBM15 localisation between a blast and a control cell measures the export state as much as it measures RBM15, and no published experiment controls for that. Lindtner S et al. 2006
  9. E_q91_rbm15_loc4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    RBM15 is a component of the WTAP complex: shotgun proteomics in human umbilical vein endothelial cells and HeLa cells identified WTAP-interacting proteins including Hakai, Virilizer homolog, KIAA0853, RBM15, the arginine/serine-rich domain-containing proteins BCLAF1 and THRAP3, and certain general splicing regulators, most with reported roles in post-transcriptional regulation; WTAP is described as required for cell-cycle progression through stabilisation of cyclin A2 mRNA.

    PMID 24100041DOI
    What this evidence cannot show

    A proteomics paper, and UniProt cites it as ECO:0000269 for both 'Nucleus speckle' and 'Nucleus, nucleoplasm' on RBM15. Whether it contains RBM15 imaging or only complex membership is not readable from the abstract, so this is a second possible rule-91 configuration on the same protein - though here the compartment it supports is independently established by two other papers.

    Evidence that argues against this (1)
    • This is the m6A writer complex, and it is the reason RBM15's speckle localisation is not a neutral descriptive fact for this graph. If the adaptor that targets m6A methylation sits in speckles, then m6A deposition has a subnuclear address, and this project holds an m6A axis for KMT2A-rearranged leukaemia with no spatial component at all. The objection to getting excited is that speckle residence is the default for almost every splicing-associated factor - this graph's own localization-nuclear-speckles-kmt2a-r entry exists partly to record how unspecific a speckle call is - so RBM15 being there discriminates between no hypotheses on its own. Uranishi H et al. 2009
  10. E_q91_rbm15_loc5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence with a retired-symbol control, 2026-09-22, querytranslation read back on every line. RBM15[tiab] = 385 and OTT1[tiab] = 11 as operand controls. Compartment terms in title or abstract: RBM15[tiab] against nuclear speckle(s), nucleolus and subnuclear = 2. Against the disease: RBM15[tiab] AND MLL[tiab] = 5, OTT1[tiab] AND MLL[tiab] = 0, RBM15[tiab] AND KMT2A[tiab] = 7 - none a localisation study.

    What this evidence cannot show

    Bounded by titles and abstracts (rule 80), and this gene is the clearest case in the tranche of why that bound matters: four experimentally attributed compartments in UniProt against two abstracts that name a compartment at all. Two symbols run; OTT and 'SPEN family' were not.

    Evidence that argues against this (1)
    • The retired-symbol control fails to reproduce the pattern that motivated it, and that is the second refutation of the brief's blanket expectation in this tranche (SGF29/CCDC101 is the first). The previous tranche found RBM15B's literature living under OTT3 and this beat was told to assume the same for its genes. For RBM15, OTT1 returns 11 records against the current symbol's 385 and ZERO against MLL - the old name carries essentially nothing. Across four genes tested here the retired symbol was decisive once (SKIC8: 2 versus 43 under SKI8), decisive once in the other direction (SEPTIN2: 0 versus 10 under SEPT2), and worthless twice (SGF29, RBM15). It is a hypothesis to test per gene, and the cost of testing it is one query. Eguchi M 2021
Connected to