No KMT2A-rearranged haematopoietic cell line in CCLE expresses IRX1 - including all four KMT2A::AFF1 lines MV4-11, NALM-19, RS4;11 and SEM at 0.0000-0.0053 RPKM
also called q149 CCLE IRX1 zero · no IRX1-positive KMT2A-r model
None of the eleven laboratory cell lines carrying a KMT2A fusion has the IRX1 gene switched on - not even the four t(4;11) lines that are the standard models for this disease. So there is no cell line in which the IRX1 half of this biology can be studied.
The same thing, for a scientist
**Measured, not asserted.** In `ccle_broad_2019` (RNA-seq RPKM), taking the KMT2A fusion partner from the study's own structural-variant calls and restricting to lines passing a PTPRC (CD45) gate, **IRX1 is at or below the per-sample silent-gene floor in 11 of 11 KMT2A-rearranged haematopoietic lines** [E1]: | partner | line | IRX1 RPKM | HOXA9 | IRX3 | IRX5 | PROM1 | |---|---|---:|---:|---:|---:|---:| | **AFF1** | MV4-11 | **0.0000** | 1.51 | **52.23** | 3.09 | 0.00 | | **AFF1** | NALM-19 | **0.0000** | **0.00** | 0.12 | 0.81 | **160.66** | | **AFF1** | RS4;11 | **0.0053** | 0.30 | 1.32 | 0.79 | 4.65 | | **AFF1** | SEM | **0.0000** | 4.47 | 0.72 | 0.37 | **105.93** | | MLLT3 | MOLM-13, MONO-MAC-1, MONO-MAC-6, NOMO-1, THP-1 | all OFF, max **0.0086** | ON in 5 of 5 | - | - | - | | MLLT1 | KOPN-8 | 0.0106 (OFF) | ON | - | - | - | | AFDN | OCI-AML2 | 0.0000 (OFF) | ON | - | - | - | **The null:** 54 of 290 non-KMT2A-rearranged haematopoietic lines in the same study are IRX1-ON (18.6%). So the zero is not the platform's inability to report IRX1. **Rule of three on the KMT2A-r lines: the true rate is below 27.3% with 95% confidence** (3/11) - a weak bound at this n, and the useful statement is the per-line naming rather than the rate. **Two consequences.** 1. **Cell lines cannot serve as the KMT2A::AFF1 IRX1-positive reference arm.** The q149 pre-registration listed the missing AFF1 arm as a NEITHER condition; CCLE does not supply it, and this is why. 2. **It extends an existing corpus entry from one line to four and from chromatin to RNA.** `finding-sem-the-standard-t411-model-has-no-irx1-signal` recorded that SEM has no IRX1 occupancy signal. **All four AFF1 lines are IRX1-negative at the transcript level**, which is the stronger statement and explains the chromatin one. **And IRX3 is not off.** MV4-11 carries **52.2 RPKM of IRX3** while IRX1 is exactly zero, so the IRX cluster neighbourhood is transcribable in an AFF1 line; the silence is specific to IRX1.
What was excluded, by name, and why
Five samples carrying a KMT2A structural variant were excluded before any value was read, and they were named in the pre-registration: C32 (skin, partner PKIG), EFE184 (endometrium, UBE4A), HCC38 (breast, ATP5MG), TEN (endometrium, CD3G), ST486 (Burkitt lymphoma, RBM47). None is a known KMT2A fusion partner and four are non-haematopoietic, so these are probable passenger or mis-called breakpoints rather than fusions.
850 of 1,156 CCLE samples failed the PTPRC (CD45) gate and were excluded - correctly, since they are solid-tumour lines. That is a large exclusion and it is stated rather than absorbed: the 290-line null is the haematopoietic null, not the CCLE-wide one.
Nothing else was removed. The KMT2A-r arm went from 16 SV-carrying samples to 11 by these two named routes.
The bound, and what it is a bound on
This is measured-not-analysed: the data were public and deposited; nobody had asked this question of them.
The absence is bounded by a per-line RNA-seq value in one release of one cell-line panel, not by a literature search. Specifically:
- It does NOT say IRX1 is unexpressed in every KMT2A-r line that exists - only in the 11 in this panel. Lines absent from CCLE (for instance primary-derived KMT2A-r cultures, or PDX-derived lines) are NEVER-CHECKED, not verified-absent.
- It does NOT say IRX1 cannot be induced in these lines. A gene at zero in standard culture can be inducible.
- The units are RPKM from
ccle_broad_2019; DepMap's own current expression release was UNREACHABLE (seeconcept-q149-a-resource-status-has-three-values-and-an-http-200-from-depmap-is-a-bot-wall), so this is a 2019-era measurement and has not been refreshed against a current release.
| absence class | measured-not-analysed |
|---|---|
| resource | ccle_broad_2019, ccle_broad_2019_rna_seq_mrna (RPKM), via cBioPortal REST API 2026-09-24 |
| partner source | ccle_broad_2019_structural_variants, KMT2A entrez 4297 |
| n kmt2a r lines scored | 11 |
| n irx1 on | 0 |
| rule of three upper PCT | 27.3 |
| null on rate non kmt2a haem lines | 0.1862 |
| excluded by PTPRC gate | 850 |
| script | analyses/q149-is-irx1-expressed-in-non-aff1-kmt2a-r/stats149.py |
lines and values
{
"AFF1": {
"MV411": 0,
"NALM19": 0,
"RS411": 0.0053,
"SEM": 0
},
"MLLT3": {
"MOLM13": 0,
"MONOMAC1": 0,
"MONOMAC6": 0,
"NOMO1": 0,
"THP1": 0.0086
},
"MLLT1": {
"KOPN8": 0.0106
},
"AFDN": {
"OCIAML2": 0
}
}excluded by name nonhaem or nonpartner
[ "C32_SKIN", "EFE184_ENDOMETRIUM", "HCC38_BREAST", "TEN_ENDOMETRIUM", "ST486_HAEMATOPOIETIC_AND_LYMPHOID_TISSUE" ]
measurements — the numbers behind the chart
[
{
"quantity": "count_IRX1_ON",
"subject": "KMT2A-rearranged haematopoietic CCLE lines, all partners",
"value": 0,
"units": "lines",
"n": 11,
"universe": "ccle_broad_2019 RNA-seq RPKM, PTPRC gate passed, known KMT2A partner from the study's own SV calls",
"null": 0.1862,
"note": "rule of three: true rate below 27.3% at 95%; the null is 54 of 290 non-KMT2A-r haematopoietic lines"
},
{
"quantity": "IRX1_rpkm",
"subject": "KMT2A::AFF1 lines MV4-11 / NALM-19 / RS4;11 / SEM",
"value": 0.0053,
"units": "RPKM (the MAXIMUM of the four)",
"n": 4,
"universe": "ccle_broad_2019 rna_seq_mrna",
"null": null,
"note": "the other three are exactly 0.0000. n=4 - NAMED, NOT COMPUTED"
},
{
"quantity": "IRX3_rpkm",
"subject": "MV4-11 (KMT2A::AFF1)",
"value": 52.232,
"units": "RPKM",
"n": 1,
"universe": "ccle_broad_2019 rna_seq_mrna; same sample's IRX1 is 0.0000",
"null": null,
"note": "the IRX cluster neighbourhood is transcribable in an AFF1 line, so IRX1's zero is gene-specific not regional"
}
]- Every preclinical result about IRX1 in this disease that comes from a cell line is a result in a model where IRX1 is off. A compound screened for activity against the iALL-IRX1 subgroup in SEM or RS4;11 has not been screened against that subgroup at all.
- NALM-19 and SEM are PROM1-high with IRX1 at zero. If PROM1 is used as a surrogate for IRX1 status - and this corpus has a PROM1 entry precisely because the occupancy beat validated a chromatin proxy against PROM1 - these two lines show the surrogate can be positive while IRX1 is absent.
- Is there ANY cultured model - line, PDX or primary culture - that is KMT2A-rearranged and IRX1-positive? This corpus has recorded the absence for SEM alone and now for all eleven CCLE lines, and the field's stated priority of establishing one has no recorded attempt.
- Can IRX1 be induced in SEM or RS4;11? A gene at 0.0000 RPKM in standard culture is not necessarily a closed locus, and the four AFF1 lines would be the cheapest place to test whether the IRX1 address is inducible at all.
- Does NALM-19 - HOXA9 at exactly 0.00 and PROM1 at 160.7 RPKM - represent the HOXA-low arm of the dichotomy without the IRX1 arm? If so it is the only HOXA-negative AFF1 line and nothing in this corpus mentions it.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In the Cancer Cell Line Encyclopedia RNA-seq data, IRX1 is at or below a per-sample ten-gene silent-tissue floor in all 11 haematopoietic lines carrying a KMT2A structural variant with a known fusion partner, including MV4-11 (0.0000 RPKM), NALM-19 (0.0000), RS4;11 (0.0053) and SEM (0.0000), while 54 of 290 non-KMT2A-rearranged haematopoietic lines are above that floor.
Next-generation characterization of the Cancer Cell Line Encyclopedia · Ghandi M, Huang FW, Jane-Valbuena J, et al. · Nature 569:503-508 2019What this evidence cannot show
My analysis of their deposited data served through the cBioPortal API, not their published result. RPKM, 2019-era; the current DepMap expression release could not be reached and this has not been refreshed against it. The KMT2A partner comes from the study's own structural-variant calls, which is a strength (not recall) but means a line whose fusion was not called by their pipeline is absent from the arm entirely. n=11 gives a rule-of-three bound of only 27.3%, so the value here is in the named per-line zeros and not in a rate. Five KMT2A-SV samples were excluded as non-haematopoietic or non-partner and are named.
Evidence that argues against this (1)
- MV4-11 carries 52.2 RPKM of IRX3 with IRX1 at exactly 0.0000, so the locus neighbourhood is transcriptionally competent and IRX1's zero is specific. That is the argument AGAINST reading the zero as a technical floor - and it also means a reader cannot infer 'the IRX cluster is closed in AFF1 lines' from this entry. Ghandi M et al. 2019
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In primary KMT2A::AFF1 infant ALL the same threshold rule calls IRX1 ON in 21 of 48 patients, so the uniform cell-line zero is a property of the models rather than of the fusion.
Gene expression profiles predictive of outcome and age in infant acute lymphoblastic leukemia: a Children's Oncology Group study · Kang H, Wilson CS, Harvey RC, et al. · Blood 119:1872-1881 2012What this evidence cannot show
Different platform (Affymetrix array versus RNA-seq), different units, different threshold magnitudes - the comparison is of CALLS made by the same rule, not of values. The four cell lines are not drawn from the 48 patients and the comparison is across studies.
Evidence that argues against this (1)
- It is equally possible that the array calls 21 of 48 patients ON because a bulk patient sample contains non-blast cells that express IRX1, while a pure cell line cannot. In that reading the cell lines are right and the patient samples are contaminated - which would be a serious problem for the whole iALL-IRX1 subgroup. No purity measurement was available for either dataset. Nagel S et al. 2022
extends that finding from SEM chromatin to all four AFF1 lines at the transcript level
- related to →IRX1 — Iroquois homeobox 1, the gene that names the HOXA-negative half of infant t(4;11) ALL
IRX1 is 0.0000-0.0106 RPKM in every KMT2A-r haematopoietic CCLE line
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
MV4-11, NALM-19, RS4;11 and SEM are all IRX1-negative
NALM-19 and SEM are PROM1-high at 160.7 and 105.9 RPKM with IRX1 at zero - the surrogate dissociates
four KMT2A::AFF1 cell lines carrying the index fusion with IRX1 at zero
CCLE via cBioPortal was the route used because DepMap was unreachable
IRX1 0.0000 RPKM in CCLE RNA-seq, PROM1 105.9
IRX1 0.0053 RPKM in CCLE RNA-seq - the highest of the four AFF1 lines and still below floor
- related to →MV4-11 (CVCL_0064) — KMT2A::AFF1, yes, but a ten-year-old's acute monocytic leukaemia with FLT3-ITD
IRX1 0.0000 RPKM while IRX3 is 52.2 - the cluster neighbourhood is transcribable
- related to →KOPN-8 (CVCL_1866): the field's default KMT2A::MLLT1 line, and its fusion protein has no YEATS domain
IRX1 0.0106 RPKM, below its own silent-gene floor
IRX1 0.0086 RPKM, the highest of the five MLLT3 lines and still below floor