EPHA7 is the one core member with a published KMT2A::AFF1 promoter ChIP - and this beat's own data put it second on separation while denying it to the AF9 samples
also called q152 EPHA7 convergence
Of the seventeen genes, one - EPHA7 - was already shown in 2007 to have the MLL-AF4 fusion protein sitting on its promoter, and switching that fusion off in a t(4;11) cell line turns EPHA7 down. Our data agree for the AFF1 patients. They disagree with the same paper about MLL-AF9, where we see nothing.
The same thing, for a scientist
Three independent systems converge on EPHA7 and one disagrees with the others on partner scope. (1) PMID 17726105 shows by ChIP that tagged ALL1/AF4 and ALL1/AF9 fusion proteins occupy the EphA7 promoter, that EphA7 is transcriptionally up-regulated in K562 cells expressing either fusion, and that siRNA knockdown of ALL1/AF4 in SEMK2 t(4;11) cells down-regulates EphA7. (2) This project's q117 re-analysis of murine MLL-AF9 (GSE260742) puts Epha7 in the top twenty fusion-occupied 2 kb bins and records a VTP50469/DMSO share ratio of 0.178, one of the largest locus-specific losses under a menin inhibitor. (3) This beat finds EPHA7 occupied in 3 of 8 human KMT2A::AFF1 samples of GSE252378 (p247, p279, p454) at 2.02-2.23x and in 0 of 19 non-AFF1 samples, with the highest non-AFF1 reading 1.09x - the second-widest separation margin (1.84) in the whole 17. **The disagreement:** the 2007 paper reports AF9 as well as AF4, and the three primary KMT2A-AF9 samples here are all negative.
What the 2007 paper actually shows
Read in the abstract, not the title: screening K562 cells producing recombinant ALL1/AF4 or ALL1/AF9 revealed transcriptional up-regulation of EphA7; siRNA suppression of ALL1/AF4 in SEMK2 cells carrying t(4;11) down-regulated EphA7; and ChIP analysis demonstrated the occupancy of tagged ALL1 fusion proteins on the EphA7 promoter, identifying EphA7 as a direct target. Up-regulation is accompanied by ERK phosphorylation, and an ERK-phosphorylation blocker caused apoptosis specific to t(4;11) cells 1.
This is one of the very few published direct-promoter-occupancy claims for a KMT2A::AFF1 target gene outside the HOXA/MEIS1 axis, and it was not cited anywhere in this graph before this beat.
The partner disagreement, stated rather than smoothed
The 2007 work reports EphA7 up-regulation for ALL1/AF9 as well as ALL1/AF4, in K562. q117 confirms it for murine MLL-AF9 at occupancy level. This beat finds no EPHA7 promoter occupancy in the three primary human KMT2A-AF9 samples (A93949, A95941, TB040576) nor in any of the other 16 non-AFF1 samples, with the highest non-AFF1 value 1.09x in a KMT2A-ENL sample 2.
Three readings are open and this beat cannot separate them: the promoter-scale 2.0x rule may be underpowered in those particular libraries; a K562 or murine system may not predict a primary human AF9 blast; or EPHA7 may genuinely be AFF1-preferential in primary disease. The honest statement is that the published partner scope and this deposit disagree, and the count at issue is 0 of 3 AF9 samples, which is a number too small to settle it.
Why this is reported alongside a negative headline
The headline result of this beat is that the 17-promoter core is not more recurrence than chance gives. That is a statement about the set. EPHA7 is the counter-example that shows why the set-level negative must not be read gene-by-gene: one member has independent, published, direct-occupancy support, and it happens to be the member with the second-cleanest separation in the data. A negative about a set is not a negative about each member.
| status | run |
|---|---|
| gene | EPHA7 |
| disagreement | the 2007 paper and q117 both implicate AF9; the three primary human KMT2A-AF9 samples in GSE252378 are negative at promoter scale |
systems converging
[ "K562 + recombinant ALL1/AF4 and ALL1/AF9 ChIP (PMID 17726105)", "SEMK2 t(4;11) siRNA knockdown (PMID 17726105)", "murine MLL-AF9 CUT&RUN, this project q117 (GSE260742)", "primary human KMT2A::AFF1 AutoCUT&RUN, this beat (GSE252378)" ]
measurements
[
{
"quantity": "n_samples_occupied",
"subject": "EPHA7 2 kb promoter, KMT2A::AFF1 arm",
"value": 3,
"units": "samples of 8",
"n": 27,
"universe": "27 GSE252378 KMT2A-fusion samples, two-antibody 2.0x plus per-library floor rule",
"null": 2.73,
"note": "null is q151's expectation of 2.73 of 27 samples for an arbitrary promoter; EPHA7 is 3 of 27 overall, i.e. at the chance rate in total but concentrated in one arm"
},
{
"quantity": "separation_margin",
"subject": "EPHA7, weakest AFF1 occupied call over strongest non-AFF1 value",
"value": 1.84,
"units": "ratio",
"n": 27,
"universe": "min-of-two-KMT2A-N-replicates within-sample shares, 8 AFF1 vs 19 non-AFF1",
"null": 1,
"note": "2.02x weakest occupied over 1.09x highest non-AFF1; second widest of the 17 core promoters after WT1 (2.17)"
}
]- The 2007 paper's therapeutic proposal is an ERK-phosphorylation blocker, and MEK/ERK inhibition in leukaemia carries the full RAS-pathway resistance repertoire: RAS and FLT3 co-mutations are already recorded in this graph as menin-inhibitor resistance routes and would apply here too.
- EPHA7 has a secreted splice variant and is silenced by promoter hypermethylation in other tumour types, so a receptor-directed agent could meet a target that is absent, soluble, or epigenetically off in a subset of patients.
- If EPHA7 occupancy is genuinely AFF1-restricted, an EPHA7-directed strategy excludes the AF9 and AF10 patients who make up much of the KMT2A-rearranged AML population.
- Is EPHA7 occupancy present in the primary KMT2A-AF9 blasts at gene-body scale, where q147 found IRX1 signal the 2 kb window missed? That single re-scoring would resolve the partner-scope disagreement at no cost.
- Does EPHA7 promoter occupancy collapse under revumenib in the treated KMT2A-ENL samples of this deposit, matching q117's murine 0.178 ratio?
- The 2007 ERK result implies a pharmacological handle. Has anyone measured EPHA7 or ERK phosphorylation in a modern KMT2A::AFF1 xenograft, or does the line stop in 2007?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
ChIP analysis demonstrates occupancy of tagged ALL1 (MLL) fusion proteins on the EphA7 promoter; EphA7 is transcriptionally up-regulated in K562 cells producing recombinant ALL1/AF4 or ALL1/AF9; siRNA-mediated suppression of ALL1/AF4 in SEMK2 t(4;11) cells down-regulates EphA7; and up-regulation is accompanied by ERK phosphorylation, with an ERK-phosphorylation blocker producing apoptosis specific to t(4;11) cells.
ALL1 fusion proteins induce deregulation of EphA7 and ERK phosphorylation in human acute leukemias · Nakanishi H, Nakamura T, Canaani E, Croce CM · Proc Natl Acad Sci U S A 104:14442-7 2007What this evidence cannot show
Abstract read in full; the figures were not opened. The ChIP is on TAGGED fusion protein in a cell line, not on endogenous fusion in a patient blast, and the K562 arm is an ectopic-expression system in a CML background. The knockdown arm (SEMK2) is the only endogenous t(4;11) experiment and it measures transcript, not occupancy.
Evidence that argues against this (1)
- EPHA7 is silenced by promoter hypermethylation in colorectal cancer and is expressed at substantial levels in most lung cancers, so its promoter state is strongly tissue-dependent and a leukaemia-specific occupancy call cannot be generalised. Tsuboi M et al. 2010
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
In GSE252378 the EPHA7 2 kb promoter (hg38 chr6:93,418,559-93,420,559, minus-strand refGene TSS 93,419,559, NM_001288629) is occupied in 3 of 8 KMT2A::AFF1 samples (p247, p279, p454) at min-of-replicate within-sample shares of 2.02-2.23x and in 0 of 19 non-AFF1 samples including all three KMT2A-AF9 samples, with the highest non-AFF1 value 1.09x in p186 (KMT2A-ENL) and no non-AFF1 sample above 1.5x.
Q152 per-core-gene occupancy and separation margins over GSE252378 · Iqbal Lab oncofusion knowledge graph, beat Q152 · this project 2026What this evidence cannot show
3 of 8 samples is 3 of 3 distinct patients (p247, p279, p454), but 3 positives cannot distinguish a partner-specific target from a patient-specific one. Confidence is moderate, not strong, because the headline analysis of this same beat shows that 3-of-8 recurrence at this pass rate is within the chance range.
Evidence that argues against this (1)
- The three positive samples are p247, p279 and p454, and p279 also carries IRX1 and 13 of the 17 core promoters, so it is the single most permissive AFF1 library in the arm. EPHA7's separation may be partly a property of which samples happen to be deep rather than of the gene. NCBI Gene Expression Omnibus 2024
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
This project's q117 re-analysis of murine MLL-AF9 fusion occupancy (GSE260742) places Epha7 among the top twenty fusion-occupied 2 kb bins and records a VTP50469/DMSO share ratio of 0.178, among the largest locus-specific losses under a menin inhibitor.
ana-q117-the-published-named-locus-pattern-reproduces-for-meis1-and-jmjd1c-and-breaks-for-mef2c-and-myb (this graph) · Iqbal Lab oncofusion knowledge graph, beat Q117 · this project 2026What this evidence cannot show
Murine MLL-AF9, recovered from TDF files with no spike-in normalisation; q117's own entries record the blacklist and unfiltered-deposit caveats that apply to every locus in that scan.
Evidence that argues against this (1)
- q117's own entries record that GSE260742 was deposited without spike-in scale factors and that 152 of its 158 samples are TDF-only, and that one 2 kb bin carried 18% of a BCOR track because half the deposit is unfiltered. A top-twenty position recovered from a TDF under those conditions is suggestive, not established. Iqbal Lab oncofusion knowledge graph et al. 2026
the gene entry for the one core member with published fusion-promoter occupancy
q117 independently puts Epha7 in the top twenty murine MLL-AF9 fusion-occupied bins at a 0.178 VTP50469/DMSO ratio
a set-level negative is not a gene-level negative: this member has independent published support
- related to →KMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)
the published partner scope includes ALL1/AF9, which this deposit's three primary AF9 samples do not reproduce
EPHA7 is present in the authors' table but in only one sample, p279