Oncofusion

start / graph / gene-alkbh5

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q6-interactome-rna-layer

ALKBH5 — the second m6A eraser: a cancer-stem-cell self-renewal role elsewhere, one incidental KMT2A record, and a flat dependency row

also called ABH5 · alkB homolog 5, RNA demethylase · alpha-ketoglutarate-dependent dioxygenase alkB homolog 5

ALKBH5 is the other enzyme that strips the chemical tag off RNA. It helps cancer stem cells renew themselves in some settings, but almost no cancer cell line depends on it and nothing has been published about it in KMT2A-rearranged disease.

The same thing, for a scientist

ALKBH5 is the second m6A demethylase. It selectively promotes tumorigenesis and cancer stem-cell self-renewal [E1], and in haematopoiesis it modulates stem and progenitor energy metabolism through m6A-mediated RNA stability control [E2]. Against this graph's index disease it is effectively absent: `(MLL OR KMT2A) AND (ALKBH5 OR ABH5)` returns 1 record, which title-checking shows to be the normal-haematopoiesis metabolism paper rather than a KMT2A study [E3]. DepMap gives it a Chronos mean of -0.146 with 1.4% of 1,208 lines depleted and lineage-discordant index-arm differences of +0.094 / -0.002, the 76.6th null percentile [E4].

Go deeper
Both erasers are flat, and that is one observation rather than two

FTO is depleted in 0.2% of 1,208 lines; ALKBH5 in 1.4% 4. Neither has any index-arm enrichment. The m6A writers in the same panel are a different picture entirely - METTL3 79.5% of lines depleted, METTL14 78.8%, WTAP 66.6%.

That asymmetry is the single most robust pattern the scan produced about m6A. Depositing the mark is broadly required; removing it is essentially never required in a proliferation assay. Two readings, both worth carrying: the erasers may be redundant with each other, or removal of the mark may simply not be rate-limiting for cell growth in culture.

The asymmetry also constrains the therapeutic reading. Every m6A-pathway drug concept in this beat's literature - STM2457 on METTL3, the YTHDC1 inhibitor, FTO inhibitors - assumes a specific node is rate-limiting, and the screen says the writers are much closer to that than the erasers are.

A one-record count that title-checks to zero

(MLL OR KMT2A) AND (ALKBH5 OR ABH5) returns 1 record. Title-checked, it is a study of ALKBH5 in normal haematopoietic stem and progenitor energy metabolism 2, not a KMT2A study; the terms co-occur incidentally.

So the accurate count is zero KMT2A-specific ALKBH5 studies - the same shape as METTL14 in this beat, where a raw count of one also title-checked to zero. Two of this beat's counted absences would have been reported as thin literatures rather than absent ones without the title check.

Facts and figures
uniprotQ6P6C2
length AA394
gene symbolALKBH5
uniprot resolutiongene_exact:ALKBH5 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as ALKBH5; length 394 aa checked
rna functionm6A demethylase (eraser); alpha-ketoglutarate dependent dioxygenase
writer eraser asymmetryerasers FTO 0.2% and ALKBH5 1.4% of lines depleted; writers METTL3 79.5%, METTL14 78.8%, WTAP 66.6%
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": -0.1456,
  "pct_lines_depleted": 1.4,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": 0.0944,
  "kmt2a_r_myeloid_diff": -0.0024,
  "null_percentile": 76.61,
  "lineage_discordant": true
}
pubmed counts
{
  "ALKBH5 AND leukemia": 43,
  "(MLL OR KMT2A) AND (ALKBH5 OR ABH5)": "1 raw, 0 after title check"
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.0899,
  "chronos_mean_other_leukaemia": -0.1227,
  "difference": 0.0328,
  "pan_essentiality_all_lines": -0.1456,
  "selectivity_rank": 13170,
  "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."
}
How this could bite us back
  • ALKBH5 is required for normal haematopoietic stem and progenitor energy metabolism, which is an on-target toxicity argument in a disease of infants.
  • A gene depleted in 1.4% of 1,208 lines supplies no dependency rationale.
  • If the two erasers are redundant, inhibiting either changes bulk m6A very little and the pharmacology would need to hit both.
Still open
  1. Do FTO and ALKBH5 buffer each other, which would explain why both m6A erasers are near-flat while every writer is broadly required?
  2. Has anyone measured m6A in KMT2A::AFF1 B-lymphoid blasts at all, given that no record pairs this fusion with the modification?
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The RNA demethylase ALKBH5 selectively promotes tumorigenesis and cancer stem cell self-renewal.

    PMID 32402250DOI
    What this evidence cannot show

    Abstract-level. The setting is cancer stem cells and tumorigenesis rather than KMT2A-rearranged leukaemia, and 'selectively' in the title refers to selectivity over normal cells rather than over disease subtypes.

    Evidence that argues against this (1)
    • ALKBH5 is depleted in 1.4% of 1,208 DepMap lines with a mean Chronos of -0.146 and no index-arm enrichment. A self-renewal claim and a near-null proliferation dependency can both be true, but nothing in the dependency data supports a requirement in KMT2A-rearranged cells. Oncofusion knowledge graph (internal analysis) 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    ALKBH5 modulates haematopoietic stem and progenitor cell energy metabolism through m6A modification-mediated RNA stability control.

    PMID 37742191DOI
    What this evidence cannot show

    Abstract-level; normal haematopoiesis rather than leukaemia. This is the single record returned by the MLL/KMT2A alias query, and it is an incidental co-occurrence rather than a KMT2A study.

    Evidence that argues against this (1)
    • A role in normal haematopoietic stem and progenitor metabolism is a toxicity argument, not a therapeutic one: an eraser required for normal HSPC energy metabolism is a poor target in an infant. Shen C et al. 2020
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence: `(MLL OR KMT2A) AND (ALKBH5 OR ABH5)` returns 1 PubMed record which on title check is a normal-haematopoiesis metabolism study, so the KMT2A-specific count is zero; `ALKBH5 AND leukemia` returns 43.

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

    Two aliases covered. As with METTL14 in this beat, the raw count of one would have read as a thin literature and title-checking makes it an absent one.

    Evidence that argues against this (1)
    • Forty-three leukaemia records with none touching KMT2A means the field has studied ALKBH5 in leukaemia extensively and never asked this subtype question. An absence of testing is not a negative result. Shen C et al. 2020
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives ALKBH5 a mean gene effect of -0.146 across 1,208 lines with 1.4% depleted below -0.5, and lineage-discordant index-arm differences of +0.094 (four B-lymphoid) and -0.002 (seven myeloid), at the 76.6th percentile of a 466-gene null. The other m6A eraser FTO is similarly flat at 0.2% of lines depleted, while the writers METTL3, METTL14 and WTAP are depleted in 79.5%, 78.8% and 66.6% of lines.

    What this evidence cannot show

    The writer-versus-eraser asymmetry is a robust pattern across five genes; the per-gene index-arm numbers rest on four and seven lines and are not results individually.

    Evidence that argues against this (1)
    • Two flat erasers is equally consistent with mutual redundancy, which a single-gene screen cannot test. The asymmetry would then be an artefact of the writers having no equivalent backup rather than of demethylation being dispensable. Oncofusion knowledge graph (internal analysis) 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q6P6C2 (ALKB5_HUMAN) is human ALKBH5, 394 amino acids; resolved by gene_exact query with the primary gene name read back as ALKBH5 and the length checked.

    What this evidence cannot show

    Single reviewed human hit whose primary gene name equals the query symbol. This project has logged six cases of gene_exact returning a protein that merely lists the queried symbol as a synonym, so the read-back is the point, not the query.

    Evidence that argues against this (1)
    • An accession establishes identity and length. It says nothing about whether this protein matters in KMT2A-rearranged leukaemia, which is what the rest of this entry is for. UniProt Consortium 2026
Connected to