Oncofusion

start / graph / gene-hnrnpa2b1

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q6-interactome-rna-layer

HNRNPA2B1 — the panel's strongest KMT2A-selective RNA-layer gene after PTBP1, with ZERO records against MLL or KMT2A

also called hnRNP A2/B1 · hnRNPA2 · RNPA2 · HNRPA2B1 · heterogeneous nuclear ribonucleoprotein A2/B1

HNRNPA2B1 reads the chemical tag on RNA in the nucleus and helps process new messages. It is one of the genes KMT2A-rearranged cells need most in this survey, and not one published paper has ever paired it with KMT2A.

The same thing, for a scientist

HNRNPA2B1 is a nuclear m6A reader and hnRNP-family processing factor. In this project's 79-gene scan it is the strongest RNA-layer gene after PTBP1: index-arm differences of -0.122 (B-lymphoid) and -0.275 (myeloid), **lineage-concordant**, 1.7th percentile of a 466-gene null, with 31.1% of 1,208 lines depleted so it is well clear of floor compression [E1]. And the literature is empty: `(MLL OR KMT2A) AND (HNRNPA2B1 OR hnRNP A2/B1 OR hnRNPA2)` returns **0 records** over three spellings, against 20 for `HNRNPA2B1 AND (leukemia OR hematopoietic stem)` [E2]. The pairing of a real dependency signal with an absent literature is the clearest gap this beat found.

Go deeper
The gap, stated precisely

Two facts, both checkable.

The dependency signal is among the best in the panel. Only PTBP1 ranks higher, and PTBP1 was added to the panel after reading the paper that nominates it. Of the genes enumerated blind, HNRNPA2B1 is the top hit. It is lineage-concordant, it is not floor-compressed, and it sits at the 1.7th percentile of the null 1.

Nobody has published on it in this disease. Zero records over three spellings 2.

What this does not mean is that HNRNPA2B1 is a target. The effect size is -0.12 to -0.28 Chronos against menin's -0.87, on four and seven cell lines, selected as the best of 79 genes - which is a multiple-comparisons setting. It is a lead, and the honest description is that it is the one gene in the beat where an unbiased scan points somewhere nobody has looked.

Why a gene with 20 leukaemia records has none with KMT2A

HNRNPA2B1 is not obscure. It has 20 records against leukaemia or haematopoietic stem cells 2, including an appearance in the integrated molecular analysis of adult T-cell leukaemia/lymphoma 3. The field knows the gene.

What it has never been is a KMT2A question. The likeliest reason is enumeration: hnRNP-family proteins are treated as general processing machinery and are rarely the subject of subtype-specific screens, which is the same reason this beat's own panel nearly missed PTBP1 - another hnRNP-adjacent protein, and the gene that outranks it.

That two of the top three RNA-layer hits are hnRNP-family or hnRNP-adjacent proteins, in a panel built mostly from m6A and spliceosome vocabulary, is worth recording as a pointer for the next enumeration.

Facts and figures
uniprotP22626
length AA353
gene symbolHNRNPA2B1
uniprot resolutiongene_exact:HNRNPA2B1 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as HNRNPA2B1; length 353 aa checked
rna functionnuclear m6A reader and hnRNP-family processing factor
status in beatthe strongest blindly-enumerated RNA-layer hit and the largest literature gap; a lead, not a target
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": -0.3579,
  "pct_lines_depleted": 31.1,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": -0.1216,
  "kmt2a_r_myeloid_diff": -0.2753,
  "null_percentile": 1.72,
  "lineage_discordant": false,
  "floor_compressed": false
}
pubmed counts
{
  "HNRNPA2B1 AND (leukemia OR hematopoietic stem)": 20,
  "(MLL OR KMT2A) AND (HNRNPA2B1 OR hnRNP A2/B1 OR hnRNPA2)": 0
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.6256,
  "chronos_mean_other_leukaemia": -0.4062,
  "difference": -0.2194,
  "pan_essentiality_all_lines": -0.3579,
  "selectivity_rank": 206,
  "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
  • Depleted in 31.1% of all DepMap lines: a substantial general requirement underlies whatever subtype bias exists.
  • An hnRNP-family protein with no catalytic pocket is a hard target, and its paralogues within the hnRNP A/B family are numerous.
  • The entry's whole case is one internally-generated scan with no independent replication. It should be read as a lead and would not survive being quoted as a finding.
Still open
  1. Does HNRNPA2B1 depletion do anything in a KMT2A::AFF1 line? Nobody has run the experiment and the scan says it is the place to look.
  2. Is the signal reproducible in an independent dependency dataset, or is it the chance extreme of a 79-gene panel?
  3. Why are two of the top three RNA-layer hits hnRNP-family or hnRNP-adjacent proteins in a panel built from m6A and spliceosome vocabulary?
Every source behind this entry (4)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives HNRNPA2B1 a mean gene effect of -0.358 across 1,208 lines with 31.1% depleted below -0.5, and lineage-concordant index-arm differences of -0.122 (four B-lymphoid) and -0.275 (seven myeloid) against 43 other leukaemia lines, at the 1.7th percentile of a 466-gene random null (SD 0.075), not floor-compressed. It is the strongest RNA-layer gene in the panel after PTBP1, and the strongest of those enumerated before any paper nominated them.

    What this evidence cannot show

    Four B-lymphoid and seven myeloid index lines. Selected as the best of 79 genes, so the percentile is a multiple-comparisons statistic. The effect size is roughly a third of menin's at most.

    Evidence that argues against this (2)
    • Picking the top gene out of a 79-gene panel scored on eleven index lines is exactly the setting in which a chance extreme appears, and this beat has no replication of the observation in any independent dataset - unlike PTBP1, whose ranking is corroborated by a published domain-focused screen. Cao Z et al. 2026
    • At 31.1% of lines depleted HNRNPA2B1 is substantially required across cancer generally, so part of the index-arm difference sits on top of a broad requirement rather than replacing it. Oncofusion knowledge graph (internal analysis) 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence: `(MLL OR KMT2A) AND (HNRNPA2B1 OR hnRNP A2/B1 OR hnRNPA2)` returns 0 PubMed records, against 20 for `HNRNPA2B1 AND (leukemia OR hematopoietic stem)`.

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

    Three spellings covered including the slashed form PubMed tokenises differently. A paper framing a result at hnRNP-family level would be invisible to all three, and hnRNP nomenclature is unusually varied.

    Evidence that argues against this (1)
    • Twenty leukaemia records and zero KMT2A records is a statement about what has been asked, not about what is true. An absence of testing is not a negative result, and here it runs in the beat's favour rather than against it, which is a reason to be sceptical of how the absence is being framed. Oncofusion knowledge graph (internal analysis) 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    HNRNPA2B1 appears in the integrated molecular analysis of adult T-cell leukaemia/lymphoma, establishing that the gene is on the field's radar in leukaemia generally.

    PMID 26437031DOI
    What this evidence cannot show

    Abstract-level; adult T-cell leukaemia/lymphoma, a virally driven adult disease with no relation to KMT2A-rearranged infant leukaemia. Cited only to establish that the gene is studied in leukaemia, not to transfer any biology.

    Evidence that argues against this (1)
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB P22626 (ROA2_HUMAN) is human HNRNPA2B1, 353 amino acids; resolved by gene_exact query with the primary gene name read back as HNRNPA2B1 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