Oncofusion

start / graph / ana-histone-chaperone-dependency-lineage-split

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q37-histone-chaperones

Is any histone chaperone selectively required in KMT2A-rearranged leukaemia? Twenty-nine genes, lineage-split - and the answer is no, with one number worth a null and a positive tail nobody expected

also called chaperone DepMap scan · the beat's own dependency triage

We asked whether cells carrying this fusion need any of the histone-packaging genes more than other leukaemia cells do. Across twenty-nine genes the answer is no - and the genes they appear to need LEAST are ones this project had assumed were central to how the fusion works.

The same thing, for a scientist

Twenty-nine histone chaperone, variant and variant-exchange genes scored on DepMap CRISPR (Chronos) gene effect across **11 KMT2A-rearranged leukaemia lines** (7 myeloid, 4 lymphoid) versus **43 other leukaemia lines**, with pan-essentiality across 1,208 lines reported for every gene and the index arm split by lineage against the matching half of the comparator. Raw effect sizes only; no z-scores, after this project's cycle-45 correction. **Result: nothing in the chaperone layer is a selective dependency of this disease.** Every gene with a difference past the project's -0.15 convention is disqualified on one of two grounds: | gene | difference | why it does not count | |---|---:|---| | CHAF1B | -0.187 | pan-essentiality **-2.314** - floor-compressed | | H3-3A | -0.227 | pan-essentiality -1.454, **and the sign REVERSES across the lineage split** (myeloid -0.472, lymphoid +0.173) | | H2AZ1 | -0.184 | pan-essentiality -0.770, past the -0.5 convention | | NASP | -0.163 | pan-essentiality -0.463 - **the only gene that passes both conventions**, and it has no null | | DAXX | -0.305 | passes both, and a 50,000-label permutation null in the sibling analysis already killed it (p = 0.0128, fails Bonferroni over 254) | **Two findings beyond the negative.** 1. **HIRA's sign reverses across the lineage split** (-0.147 myeloid, +0.032 lymphoid) while **ATRX's and DAXX's signals live entirely in the lymphoid half** (-0.321 and -0.601, against -0.0003 and -0.117 myeloid). The index disease is lymphoid. Four lines. 2. **The positive tail is the surprise.** The three genes LEAST required in KMT2A-rearranged lines are **SSRP1 (+0.265), EP400 (+0.121) and SRCAP (+0.118)**, with SUPT16H (+0.117) close behind. SSRP1 and SUPT16H are **FACT**, which this graph already carries as part of the fusion's proposed elongation mechanism.

Go deeper
The positive tail, and why this project has never looked at one

Every DepMap analysis in this project sorts ascending and reads the bottom: genes MORE required in the index arm. A difference of +0.265 for SSRP1 is the largest absolute difference in this scan after MEN1, DAXX, KAT7 and DOT1L - and it points the other way.

Three readings, in order of prior:

  1. Noise. Eleven lines, 29 genes, no null. A tail exists at both ends of any such scan.
  2. Lineage. The index arm is myeloid-enriched. The split does not rescue this reading: SSRP1 is +0.206 myeloid and +0.354 lymphoid, positive in both halves.
  3. Real. A fusion that supplies its own elongation drive through the super elongation complex may depend less on the general elongation and exchange apparatus than a leukaemia that does not.

Reading 3 would be a substantive negative for this graph's own mechanism entries and is recorded so it is not lost. The test is a permutation null run on the POSITIVE tail, which this project has never done and which costs one script.

The honest statement: FACT being less required in KMT2A-rearranged lines is an observation from one scan with no null, and this entry asserts it as an observation and nothing more.

What this analysis cannot do
  • No permutation null. The sibling triage's five-gene shortlist did not survive one, and the strongest gene in it (DAXX, -0.305) failed Bonferroni. Nothing here is larger than that.
  • No infant KMT2A::AFF1 line exists in DepMap. The 11 index lines are ALL-PO, HB11;19, KOPN-8, Karpas-45, ML-2, MOLM-13, MOLM-14, MV4-11, Mono-Mac-6, NOMO-1, OCI-AML-2, RS4;11, SEM, SEMK2, SHI-1 and THP-1 as annotated by Cellosaurus (16 lines, of which 11 have usable screens). None is infant-derived.
  • A knockout screen cannot see a rate-setting chaperone. Every claim in this beat about turnover concerns how FAST a mark is lost, and a two-week viability screen measures whether the cell dies. The SPT6 degron result shows that acute and chronic loss of the same chaperone produce different phenotypes; this instrument only ever measures the chronic one.
  • A knockout screen cannot see a partial-occupancy window. Halving Chaf1b in a mouse suppresses leukaemogenesis without harming normal haematopoiesis while CHAF1B's CRISPR score is -2.31.
  • A knockout screen cannot see a paralogue-buffered pair. ASF1A and ASF1B are each dispensable and the double knockout has not been run.
Facts and figures
scriptanalyses/histone-chaperone-dependency-lineage-split/run.py
statusrun
result summaryNo histone chaperone, variant or variant-exchange gene is a selective dependency of KMT2A-rearranged leukaemia lines once pan-essentiality and lineage are accounted for. NASP is the only gene passing both of this project's conventions and has no permutation null. The largest positive differences - genes LEAST required in the index arm - are SSRP1, EP400 and SRCAP, i.e. FACT and the H2A.Z exchange machinery.
n genes29
n index lines11
n index myeloid7
n index lymphoid4
n comparator lines43
n all lines1208
cellosaurus kmt2a lines found16
no permutation nulltrue
next steprun a 50,000-label permutation null on both tails, including the positive one, which this project has never done
inputs
[
  "DepMap CRISPR (Chronos) gene effect via Open Targets GraphQL",
  "Cellosaurus KMT2A-rearrangement annotation",
  "HGNC symbol-to-Ensembl mapping"
]
How this could bite us back
  • A clean negative from a screen is the easiest result to over-read: it says complete knockout of these genes does not preferentially kill KMT2A-rearranged lines, and it says nothing about partial inhibition, acute inhibition, rate effects, paralogue-buffered pairs or the in-vivo stem cell compartment - all of which are where the published chaperone phenotypes in leukaemia live.
  • If the positive tail is real, the graph's FACT entries need qualifying rather than extending, and this beat has no way to tell whether it is real.
Still open
  1. Run the permutation null on both tails. NASP is the only candidate and the positive tail (SSRP1, EP400, SRCAP) has never been tested in this project.
  2. Are FACT and the H2A.Z exchange machinery genuinely less required in KMT2A-rearranged lines? If so it is a negative result about this graph's own elongation mechanism and should be written as one.
  3. Why is the ATRX/DAXX signal confined to the lymphoid half while HIRA's reverses? With four lymphoid lines this may be nothing, but the index disease is lymphoid and the pattern is consistent across two genes that work together.
  4. Repeat with a double-knockout library for the paralogue pairs ASF1A/ASF1B, UBN1/UBN2 and H2AZ1/H2AZ2, none of which a single-gene screen can assess.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    Across 29 histone chaperone, variant and exchange genes scored on DepMap CRISPR gene effect in 11 KMT2A-rearranged versus 43 other leukaemia lines, no gene combines a difference past -0.15 with a pan-essentiality shallower than -0.5 except NASP (-0.163, -0.463) and DAXX (-0.305, -0.374); the three largest positive differences are SSRP1 +0.265, EP400 +0.121 and SRCAP +0.118; HIRA's difference reverses sign across the myeloid/lymphoid split of the index arm (-0.147 / +0.032) as does H3-3A's (-0.472 / +0.173), while ATRX and DAXX carry their entire signal in the 4-line lymphoid half (-0.321 and -0.601).

    What this evidence cannot show

    Run 2026-09-20. DepMap CRISPR (Chronos) gene effect via the Open Targets GraphQL API; KMT2A-rearrangement status from Cellosaurus; Ensembl ids from HGNC. **Moderate: this is a well-specified negative, and the negative is the robust part.** No permutation null was computed. The lymphoid half is 4 lines. No infant KMT2A::AFF1 line is in DepMap. The DepMap release version is not stated by the API. The -0.15 and -0.5 thresholds are this project's conventions, not derived cut-offs.

    Evidence that argues against this (1)
    • The sibling analysis that ranked 254 unwritten interactors by the same statistic produced a five-gene shortlist that a 50,000-label permutation null subsequently killed; its strongest gene, DAXX, reached p = 0.0128 and failed Bonferroni correction. Nothing in this scan is larger, so no positive claim from it can be stronger than one already refuted. Oncofusion knowledge graph (internal analysis) 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A pooled CRISPR knockout screen cannot detect a therapeutic window that exists only at partial reduction: reducing Chaf1b by heterozygous deletion or by a CAF-1 dominant-negative allele suppresses leukaemogenesis in vivo without impairing normal haematopoiesis, while CHAF1B's DepMap pan-essentiality is -2.314 across 1,208 lines.

    A CHAF1B-Dependent Molecular Switch in Hematopoiesis and Leukemia Pathogenesis. · Volk A, Liang K, Suraneni P, et al. · Cancer Cell 2018
    PMID 30423293DOI
    What this evidence cannot show

    Abstract read from efetch 2026-09-20; full text not read (PMC6235627 exists). Mouse, myeloid leukaemia. The comparison drawn here between a germline heterozygote and a CRISPR knockout is this entry's argument, not the paper's. Graded strong for the window claim within the reported models; the pharmacological extension is an inference.

    Evidence that argues against this (1)
    • A germline heterozygote is a lifelong 50% reduction with full compensation available during development; a drug gives partial, fluctuating occupancy in an already-damaged marrow. The mouse result bounds what the screen can see, but it does not establish that the window is reachable pharmacologically. Dean ST et al. 2023
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Acute and chronic depletion of the same histone chaperone produce different phenotypes: acute SPT6 degradation causes loss of RNAPII processivity and readthrough transcription, while cryptic intragenic transcription appears only on long-term depletion and is attributable to accumulated epigenetic perturbation rather than to a direct function.

    Targeted protein degradation reveals a direct role of SPT6 in RNAPII elongation and termination. · Narain A, Bhandare P, Adhikari B, et al. · Molecular Cell 2021
    PMID 34233157DOI
    What this evidence cannot show

    Abstract read from efetch 2026-09-20; full text not read (PMC8354102 exists). Human cell line with an engineered degron; not leukaemic. Cited here as the methodological limit on this analysis - a two-week CRISPR screen is the long-term condition - and that framing is this entry's, not the paper's.

    Evidence that argues against this (1)
    • The acute/chronic distinction is about protein presence and does not map cleanly onto a screen's readout either; a pooled screen measures proliferation, which is a third thing again. The limitation is real but the size of it is unquantified. Oncofusion knowledge graph (internal analysis) 2026
Connected to