Oncofusion

start / graph / gene-khdrbs2

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q6-interactome-rna-layer

KHDRBS2 (SLM1) — a counted absence: zero leukaemia records over five aliases, in a family that regulates the MLL-AF4 message

also called SLM1 · SLM-1 · Sam68-like mammalian protein 1 · T-STAR family · KH domain containing, RNA binding, signal transduction associated 2

KHDRBS2 is one of the three close relatives of Sam68, the protein that holds down production of the MLL-AF4 fusion. Nobody has published on it in any leukaemia at all, which is worth recording because the family as a whole is now implicated in this disease.

The same thing, for a scientist

KHDRBS2/SLM1 is one of three KH-domain STAR-family paralogues (with KHDRBS1/Sam68 and KHDRBS3/SLM2) sharing the domain architecture that recognises AU-rich RNA. The family collectively inhibits MLL-AF4 translation through ribosomal stalling [E1], but the abstract naming that result names the *family*, not which member does the work. Searched this session, `(KHDRBS2 OR SLM1) AND (leukemia OR leukaemia)` returns **0 records**, and `(MLL OR KMT2A) AND (KHDRBS2 OR SLM1 OR SLM-1)` returns **0** [E2]. DepMap gives it a Chronos mean of +0.069 with 0.1% of 1,208 lines depleted - effectively no dependency anywhere [E3]. This entry is a bounded absence with its queries stated, written because the family's involvement makes the absence interesting rather than merely empty.

Go deeper
What is and is not being claimed

The claim is narrow and checkable: no PubMed record pairs KHDRBS2 or SLM1 with leukaemia of any kind, and none pairs it with MLL or KMT2A. The queries are written out in 2 so the next reader can rerun them rather than re-derive them.

The claim is not that KHDRBS2 is irrelevant. It is that nobody has looked. Those are different statements and this graph has had to correct entries that conflated them.

The absence is worth a file precisely because of the family result 1. Before that paper, KHDRBS2 would have been an arbitrary gene. After it, it is one of three candidates for a demonstrated activity on the fusion's own transcript, and two of the three have never been examined in this disease.

Why a symbol-keyed search could still be wrong here

This project has twice been burned by absence claims that were artefacts of the symbol set. Two checks were run.

Retired and alternative symbols: SLM1 and SLM-1 are the used alternatives and both were included; the aliases are not word-like or calendar-colliding, so the count needs no title-check deflation of the sort SEPT* or RD required.

The complex-name route: the family term KHDRBS was searched as its own token and returns exactly one leukaemia record, which is 1 itself. So unlike the NSL case this project logged in cycle 45, the family phrase here is reachable - and it reaches one paper, which does not name individual members in its abstract. The honest summary is that the family is studied in this disease at family resolution only, and no member-level work exists for KHDRBS2.

Facts and figures
uniprotQ5VWX1
length AA349
gene symbolKHDRBS2
uniprot resolutiongene_exact:KHDRBS2 AND organism_id:9606 AND reviewed:true; one hit; primary gene name read back as KHDRBS2; length 349 aa checked
absence claim
{
  "queries_run": [
    "(KHDRBS2 OR SLM1) AND (leukemia OR leukaemia)",
    "(MLL OR KMT2A) AND (KHDRBS2 OR SLM1 OR SLM-1)",
    "KHDRBS AND leukemia"
  ],
  "records": [
    0,
    0,
    1
  ],
  "date": "2026-09-20",
  "note": "the single family-term record is PMID 36335100 itself"
}
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": 0.0694,
  "pct_lines_depleted": 0.1,
  "n_lines": 1208,
  "kmt2a_r_lymphoid_diff": 0.0676,
  "kmt2a_r_myeloid_diff": 0.0412
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": 0.102,
  "chronos_mean_other_leukaemia": 0.0512,
  "difference": 0.0508,
  "pan_essentiality_all_lines": 0.0694,
  "selectivity_rank": 14805,
  "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
  • An entry whose content is a zero is the easiest kind to mis-cite as evidence of irrelevance. It is evidence that nobody looked.
  • If the family is redundant on the fusion transcript, no single-member manipulation - including the ones that have not been done - would show anything, and the absence would never be resolved by member-level work.
Still open
  1. Which KHDRBS family member actually carries the translational-repression activity on the MLL-AF4 transcript, given that the founding paper reports it at family level?
  2. Is KHDRBS2 expressed at all in B-lymphoid progenitors and in t(4;11) blasts, which would decide whether its literature absence is an oversight or appropriate?
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    RNA-binding proteins of the KHDRBS family inhibit MLL-AF4 translation through ribosomal stalling at AU-rich regulatory sites, and disrupting that association by synonymous mutation restores translation and causes leukaemic transformation in vivo.

    PMID 36335100DOI
    What this evidence cannot show

    The abstract names the KHDRBS family collectively. It does not state which of KHDRBS1, KHDRBS2 or KHDRBS3 carries the activity, so attributing it to KHDRBS2 specifically would be unsupported. Abstract read; full text not read.

    Evidence that argues against this (1)
    • KHDRBS2 has markedly restricted expression compared with the ubiquitously expressed KHDRBS1, so a family-level result is most likely carried by the member actually present in a B-lymphoid progenitor. That would make KHDRBS2's absence from the literature appropriate rather than an oversight. UniProt Consortium 2026
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Counted absence, queries run this session against PubMed: `(KHDRBS2 OR SLM1) AND (leukemia OR leukaemia)` returns 0 records; `(MLL OR KMT2A) AND (KHDRBS2 OR SLM1 OR SLM-1)` returns 0 records. The family term `KHDRBS AND leukemia` returns 1 record, which is the MLL-AF4 translational-control paper itself.

    What this evidence cannot show

    A count is only as good as its symbol set. Five spellings were covered. PubMed matches multi-word [Title/Abstract] terms as unstemmed phrases, so a paper using a phrase not in this list would be missed, and a paper discussing the family without naming the member is invisible to every member-keyed query by construction.

    Evidence that argues against this (1)
    • The generalisable failure this project has logged is that bounded, well-written absence claims are the ones that look most trustworthy while being wrong - the septin case under-counted tenfold by keying to current symbols. Here the family-phrase route was also run and returned one record, so the absence is bounded at member resolution rather than at topic resolution. Okuda H et al. 2022
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.analysis

    DepMap CRISPR (Chronos) gives KHDRBS2 a mean gene effect of +0.069 across 1,208 lines with 0.1% of lines depleted below -0.5, and a positive difference in the KMT2A-rearranged index arm (+0.068 B-lymphoid, +0.041 myeloid) against 43 other leukaemia lines.

    What this evidence cannot show

    A flat DepMap row cannot distinguish 'not required' from 'not expressed in these lines' from 'buffered by a paralogue'. Given KHDRBS2's restricted expression the second is the most likely reading and the screen does not test it.

    Evidence that argues against this (1)
    • A dropout screen cannot detect a negative regulator in either direction, which is the activity the family paper reports. So a flat row here carries almost no information about the published mechanism. Okuda H et al. 2022
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProtKB Q5VWX1 (KHDR2_HUMAN) is human KHDRBS2, 349 amino acids; resolved by gene_exact query with the primary gene name read back as KHDRBS2 and the length checked.

    What this evidence cannot show

    Single reviewed human hit with a matching primary gene name. The accession establishes identity and nothing about disease relevance.

    Evidence that argues against this (1)
    • At 349 aa against KHDRBS1's 443 the paralogues are not interchangeable in size, and sequence similarity within a KH/STAR family is not the same as shared substrate preference, which is the property that would matter for the AU-rich sites on the fusion transcript. UniProt Consortium 2026
Connected to