Oncofusion

start / graph / ana-q118-the-mll1-n-antibody-masks-the-occupancy-collapse-four-fold-in-rs4-11

Analysisstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q118-menin-at-the-human-index-fusion

At the same bins in the same cells, menin's share ratio is 0.20-0.26 and the MLL1 N-terminus antibody's is 0.80-0.91 - so a shared-epitope MLL1-N ChIP under-reports the menin inhibitor's occupancy effect by 3.5 to 4.0-fold, and the number is now measured rather than asserted

also called shared-epitope masking · criterion D · MLL1n vs MENIN at the same bins

The antibody most papers use to ask whether the fusion protein is still on the DNA also detects the normal, non-cancerous version of the same protein. Measured side by side at the same places in the same cells, that antibody shows a four times smaller drug effect than a menin antibody does, so it hides most of what the drug is doing.

The same thing, for a scientist

**Measured at genome scale in human RS4;11 (KMT2A::AFF1), 2 kb bins, universe 1,389,514, top 0.1% = 1,390 bins** [E1]: | held set | MENIN share ratio | MLL1n share ratio | masking D | |---|---:|---:|---:| | MLL1n's own top 0.1% (75.2% CpG island) | 0.2560 | **0.9069** | **3.54x** | | MENIN's own top 0.1% (50.4% CpG island) | 0.1999 | **0.7996** | **4.00x** | Same cells, same bins, same drug, same library-free within-sample statistic. The only difference is the antibody, and it changes the answer four-fold. The same contrast in enrichment and in the ceiling-free retention statistic, at each track's own top 0.1% of the 1,389,514 universe bins: **MLL1-N 14.20x -> 12.18x over its own genome median, with 43.67% of its held bins still in its own top 0.1%; MENIN 14.91x -> 1.96x, with 10.58%.** The two tracks start from the same enrichment and end four to six-fold apart. **Partitioning the two held sets makes it sharper** [E1]: | partition | n | CpG island | MLL1n ratio | MENIN ratio | |---|---:|---:|---:|---:| | MLL1n **AND** MENIN (the closest available fusion proxy) | 692 | 69.5% | **0.8305** | 0.2005 | | MLL1n only | 698 | 80.8% | **1.0422** | 0.4014 | | MENIN only | 698 | 31.4% | 0.6948 | 0.1991 | **At the sites where menin and an MLL1 N-terminus co-occupy - the best proxy for fusion-occupied chromatin available without a tag - menin falls 5-fold and MLL1-N-reactive material falls 1.2-fold.** At sites carrying MLL1-N without menin, the MLL1-N signal does not move at all (1.0422). **Two readings, and the data does not choose between them.** 1. **Dilution.** The fusion is evicted with menin, and the MLL1-N track's flatness is wild-type MLL1 remaining behind. Under this reading, D is the dilution factor and the fusion's true share ratio is nearer menin's 0.20. 2. **Retention.** An MLL1-N-containing complex genuinely stays after menin leaves - i.e. the fusion, or a fusion fragment, remains bound by a menin-independent contact (AT-hooks, CXXC, LEDGF/PSIP1). Reading 2 is `hyp-the-fusion-stays-on-chromatin-and-stops-working`, which this graph already holds. **This beat cannot arbitrate, and that is the result**: the experiment that would is a tagged or degron-tagged KMT2A::AFF1 in RS4;11.

Go deeper
The same measurement in the three lines at TSS windows, where it also holds

Instrument A (the deposited ROSE TSS table, 21,461 windows, top 1% = 215 windows), measured at each line's MENIN-DMSO top 1% 2:

linefusionMENINMLL1nDOT1LD = MLL1n/MENIN
RS4;11KMT2A::AFF10.66970.84870.61931.27x
MOLM13KMT2A::MLLT30.25130.61950.66382.46x
ML-2KMT2A::AFDN0.35170.62710.36221.78x

And the ceiling-free version - the fraction of each track's own top 1% still in its own top 1% under drug:

lineMENINMLL1nratio
RS4;110.2050.5492.68x
MOLM130.1300.6655.12x
ML-20.3720.6371.71x

D fires in MOLM13 (2.46x, above the pre-registered 1.8-2.2 band), sits inside the band in ML-2 (1.78x), and does not fire in RS4;11 (1.27x) on instrument A. On instrument B, in RS4;11, it fires at 3.54-4.00x. The discrepancy is the denominator, not the biology: a share computed over 21,461 promoter windows - all of them menin-enriched - has almost no unoccupied space to lose share into, so it compresses every effect. See concept-q118-a-share-ratio-is-comparable-only-if-the-denominator-is.

Why this is the most transferable thing in the beat

Q117 named this as its own open question and wrote it into its next steps: "MLL1 ChIP in the same cells under the same three conditions is the direct test of the attribution problem. If a shared-epitope antibody masks fusion loss, MLL1 and FLAG should diverge..." It could not run it because GSE260742's MLL1 samples were not downloaded.

This beat runs the equivalent test in human cells and puts a number on the masking: 3.5-4.0x at genome scale in the index-fusion line. Any published claim of the form "the menin inhibitor does not globally displace MLL1/the fusion from chromatin", made from an MLL1-N or an MLL1-C ChIP, is now quantified as compatible with a 5-fold menin loss at the same sites.

This graph holds such a claim: mechanism-vtp-50469-does-not-globally-displace-the-fusion-from-chromatin. That entry is not wrong - the MLL1-N track really does not globally displace - but its subject is MLL1-N-reactive material, and the number needed to read it is in this entry.

Facts and figures
scriptanalyses/q118-menin-at-the-human-index-fusion/binsC.py
statusrun
result summaryShared-epitope masking measured: D = 3.54x at MLL1n's top bins and 4.00x at MENIN's top bins in RS4;11 genome-wide; 2.68x / 5.12x / 1.71x by retention-in-own-top in RS4;11 / MOLM13 / ML-2 at TSS windows.
antibody mll1nanti-MLL1n, Bethyl A300-086A lots 5 or 6 - an MLL1 N-TERMINUS epitope, retained in the fusion
antibody meninanti-Menin, Bethyl A300-105A Lot 8
masking D genomewide rs4114
cell lineRS4;11, MOLM13, ML-2
genome buildhg19
shown for fusionKMT2A::AFF1 (RS4;11), KMT2A::MLLT3 (MOLM13), KMT2A::AFDN (ML-2) - and in none of them is the fusion distinguished from wild-type MLL1, which is the entry's subject
inputs
[
  "GSM3635693/94 RS4;11 MENIN",
  "GSM3635695/96 RS4;11 MLL1n",
  "GSM3635699 RS4;11 input",
  "GSE127507 ROSE TSS signal table"
]
How this could bite us back
  • If reading 1 is right, the field's standard pharmacodynamic readout systematically under-reports target engagement four-fold, and a compound that looked insufficiently active in an MLL1-N ChIP may have been active.
  • If reading 2 is right, the menin inhibitor leaves an MLL1-N-containing complex on chromatin at 83% of its baseline share at the co-occupied sites. A complex that stays but stops working is a different drug target from a complex that leaves, and it would need a degrader rather than an interface inhibitor.
  • Either way, a trial biomarker built on MLL1 ChIP is measuring a quantity whose relationship to fusion occupancy is unknown by a factor of four, and this is now a measured factor rather than a theoretical concern.
Still open
  1. Which of the two readings is right? A FLAG-, HA- or degron-tagged KMT2A::AFF1 in RS4;11, ChIPed under the same conditions, decides it in one experiment, and it is written up as exp-q118-a-tagged-kmt2a-aff1-chip-in-rs4-11-under-vtp-50469.
  2. What fraction of the MLL1-N signal in RS4;11 is fusion and what fraction is wild-type MLL1? RS4;11 is heterozygous for t(4;11) and retains a wild-type KMT2A allele, so the two species are both present, but no published measurement apportions the ChIP signal between them.
  3. An MLL1 C-TERMINUS antibody sees wild-type MLL1 and NOT the fusion. A C-terminal ChIP in the same cells would give the wild-type-only track that would let the fusion's contribution be subtracted arithmetically, and no such sample exists in GSE127507.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    In RS4;11 2 kb bins, at MLL1n's own top 0.1% the MENIN share ratio is 0.2560 and the MLL1n share ratio is 0.9069 (3.54-fold), and at MENIN's own top 0.1% they are 0.1999 and 0.7996 (4.00-fold); partitioned, the 692 bins in both held sets give MLL1n 0.8305 against MENIN 0.2005, the 698 MLL1n-only bins give MLL1n 1.0422, and the 698 MENIN-only bins are only 31.4% CpG islands against 80.8% for the MLL1n-only bins.

    What this evidence cannot show

    NEITHER ANTIBODY IN THIS SERIES IS FUSION-SPECIFIC AND THE ANTIBODY NAME SAYS SO. MLL1n is anti-MLL1 N-terminus (Bethyl A300-086A, lots 5 or 6); the N-terminus is retained in KMT2A::AFF1, so that track is the sum of the fusion and wild-type MLL1 in unknown proportion. MENIN is Bethyl A300-105A Lot 8; menin binds the fusion, wild-type MLL1 and non-KMT2A partners. Q117 could attribute its numbers because its fusion was FLAG-tagged; nothing here can. Every occupancy statement in this beat is about MLL1-N-reactive material or menin-occupied sites, never about the fusion. No scale factor exists for any sample. The depositors state that a Drosophila spike-in was run on one batch, that fly/human ratios were 'found to be small', and that NO spike-in normalisation was applied; no number for that measurement is reported and none is deposited. Every cross-condition figure here is therefore a WITHIN-SAMPLE share or a within-sample rank, which cancel an unknown per-sample scale factor exactly, and never a fold change. n = 1 per antibody per condition per line - there is no replicate anywhere in GSE127507 - so no number here carries a spread and none is precise. The MENIN and MLL1n ChIPs are different experiments with different efficiencies, so D is a ratio of two share ratios and not a measurement of how much wild-type MLL1 there is. Every denominator was printed and asserted non-zero. The MLL1n-AND-MENIN intersection is a PROXY for fusion-occupied chromatin, not an attribution: wild-type MLL1 also binds menin, so that set contains wild-type complexes too.

    Evidence that argues against this (1)
    • D is exactly what you would see if the fusion were NOT evicted and menin simply dissociated from an otherwise intact, chromatin-bound complex. Under that reading nothing is being masked and the MLL1-N number is the correct one. The graph already holds that model as hyp-the-fusion-stays-on-chromatin-and-stops-working, and this entry cannot distinguish it from dilution by wild-type MLL1. Krivtsov AV et al. 2019
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    At TSS windows in all three lines, the fraction of each track's own top 1% still in its own top 1% under VTP-50469 is 0.205 / 0.130 / 0.372 for MENIN and 0.549 / 0.665 / 0.637 for MLL1n in RS4;11 / MOLM13 / ML-2 - so MLL1-N-reactive material is retained 1.7 to 5.1-fold more than menin in every line.

    What this evidence cannot show

    NEITHER ANTIBODY IN THIS SERIES IS FUSION-SPECIFIC AND THE ANTIBODY NAME SAYS SO. MLL1n is anti-MLL1 N-terminus (Bethyl A300-086A, lots 5 or 6); the N-terminus is retained in KMT2A::AFF1, so that track is the sum of the fusion and wild-type MLL1 in unknown proportion. MENIN is Bethyl A300-105A Lot 8; menin binds the fusion, wild-type MLL1 and non-KMT2A partners. Q117 could attribute its numbers because its fusion was FLAG-tagged; nothing here can. Every occupancy statement in this beat is about MLL1-N-reactive material or menin-occupied sites, never about the fusion. The ROSE table is input-subtracted by the depositors with a cell-line-matched input; for RS4;11 and ML-2 that is the DMSO input for both conditions, because no VTP input exists for those lines. The MOLM13 MENIN pair is a 4-day exposure while everything else is 3 days, so the MOLM13 MENIN figure is not exposure-matched to its own MLL1n figure. n = 1 per antibody per condition per line - there is no replicate anywhere in GSE127507 - so no number here carries a spread and none is precise.

    Evidence that argues against this (1)
    • MOLM13's MENIN samples are day 4 and its MLL1n samples are day 3. The 5.12-fold MOLM13 masking figure therefore partly measures one extra day of drug, and the honest bound on it is the 3-day RS4;11 figure of 2.68-fold on the same statistic. NCBI Gene Expression Omnibus 2019
Connected to