Oncofusion

start / graph / line-sem-kmt2a-aff1

Explainerestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q21-index-fusion-models

SEM (CVCL_0095) — the workhorse, from a five-year-old, and microsatellite unstable

also called CVCL_0095 · ACH-000782

SEM is the cell line most of this field's KMT2A::AFF1 experiments are done in. It came from a five-year-old girl, not a baby, and its DNA repair is broken in a way the real infant disease almost never is.

The same thing, for a scientist

SEM (CVCL_0095, DepMap ACH-000782, DSMZ ACC-546) is the single most used human KMT2A::AFF1 model and, in this project's own provenance audit, the only route by which any resolved unstated system reached the index fusion. It was established from the **peripheral blood of a 5-year-old Caucasian female**; the founding report describes it as biphenotypic and interleukin-7-responsive. Cellosaurus annotates KMT2A-AFF1, TP53 p.Arg248Gln heterozygous, CDKN2A p.His83Tyr heterozygous, and — the fact this entry exists to surface — **microsatellite instability: Instable (MSI)**, from the CCLE next-generation characterisation. It is a member of the LL-100 blood-cancer panel and has shRNA-screen, proteomic, phosphoproteomic, methylation, exome, microarray and RNA-seq data. Its subclone SEM-K2 (CVCL_S906, ACH-001735) is a separate DepMap entry from the same patient.

Go deeper
Three things about SEM that its ubiquity hides

1. It is not infant material. Age at derivation 5 years 12. Infant KMT2A::AFF1 ALL and childhood KMT2A::AFF1 ALL differ in their gene-expression programme: the fetal programme is maintained in infant cases and not in childhood cases 4. SEM is on the childhood side of that line.

2. It is microsatellite unstable 3. Infant KMT2A-rearranged ALL has among the lowest somatic mutation burdens of any human cancer and carries essentially no mismatch-repair phenotype. A line with MSI has a mutational process the disease does not, which matters for any DNA-damage, replication-stress or Fanconi-pathway claim — and this graph has an open Fanconi signal whose binding objection is explicitly line artefact.

3. SEM and SEM-K2 are one patient. Both are in the screened DepMap panel as separate lines 2. A dependency reproduced 'in SEM and SEM-K2' has been reproduced in one child's leukaemia, in the same way that MOLM-13 and MOLM-14 are one adult.

Facts and figures
accessionCVCL_0095
depmap IDACH-000782
resolved byCellosaurus accession lookup 2026-09-19; primary name read back as SEM
age at derivation5 years (age at sampling)
infant derivedfalse
sexFemale
populationCaucasian
sitePeripheral blood
lineageB-precursor, described as biphenotypic in the founding report
fusionKMT2A::AFF1
msiInstable (MSI)
in depmaptrue
depmap crispr gene effecttrue
men1 gene effect-0.959
same individual asCVCL_S906 (SEM-K2, ACH-001735, also in the screened panel)
repositoryDSMZ ACC-546
obtainableYes — DSMZ catalogue line
co lesions
[
  "TP53 p.Arg248Gln het",
  "CDKN2A p.His83Tyr het"
]
How this could bite us back
  • MSI means an elevated mutational input, so any resistance mutation that emerges under drug pressure in SEM may reflect the line's mutator phenotype rather than a realistic clinical escape route.
  • SEM and SEM-K2 are one patient counted twice in the screened DepMap panel, which inflates the apparent independence of every panel-level statistic this graph has computed.
  • TP53 and CDKN2A are both mutated here and are both wild-type in most infant KMT2A::AFF1 disease, so apoptosis and cell-cycle results in SEM start from a different baseline than the patients do.
Still open
  1. Was SEM derived at diagnosis or at relapse, and at what age was the patient diagnosed? If she was an infant, a large part of this beat's verdict changes.
  2. Is SEM's MSI phenotype present in any primary KMT2A-rearranged ALL, or is it a culture acquisition? A mismatch-repair phenotype in the disease would be a major and currently undocumented finding.
  3. How many of this graph's SEM-derived conclusions would survive replication in an infant-derived line? Nobody has run that comparison for any conclusion.
Every source behind this entry (4)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The acute lymphoblastic leukaemia cell line SEM, carrying the t(4;11) chromosomal rearrangement, is biphenotypic and responsive to interleukin-7.

    PMID 8199015DOI
    What this evidence cannot show

    Bibliographic record fetched from PubMed on 2026-09-19; the paper was not read and the abstract was not read in full. The 5-year age and the peripheral-blood site are taken from Cellosaurus [E2], not from this report.

    Evidence that argues against this (1)
    • 'Biphenotypic' in 1994 usage is not the modern MPAL definition and does not establish lineage plasticity in the sense this graph uses. IL-7 responsiveness likewise reflects the culture conditions of the era; SEM is now grown without IL-7, so a property in the founding report is not necessarily a property of the cells anyone uses today. Drexler HG et al. 2004
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Cellosaurus accession CVCL_0095 records SEM as a Caucasian female human cancer cell line aged 5 years at sampling, derived in situ from peripheral blood, with a KMT2A-AFF1 gene-fusion annotation, TP53 p.Arg248Gln and CDKN2A p.His83Tyr heterozygous mutations, DepMap identifier ACH-000782, DSMZ catalogue ACC-546, and a child cell line SEM-K2 (CVCL_S906, DepMap ACH-001735) derived from it.

    Cellosaurus cell line SEM (CVCL_0095) · Bairoch A, et al. (SIB Swiss Institute of Bioinformatics) · Cellosaurus 2025
    What this evidence cannot show

    Fetched by accession 2026-09-19 with the primary name read back. Cellosaurus records the age at sampling, and for a line established at relapse the age at sampling and the age at diagnosis differ; the record does not state which this is, so '5 years' is an upper bound on age at diagnosis rather than the diagnosis age itself.

    Evidence that argues against this (1)
    • If SEM was derived at relapse then the patient was diagnosed earlier than 5 years, possibly considerably earlier, and the line could be closer to infant disease than the record implies. Nothing in the Cellosaurus entry settles this, and it is the single fact that would most change how this line should be read. Bairoch A et al. 2025
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the next-generation characterisation of the Cancer Cell Line Encyclopedia, SEM is classified as microsatellite instable (MSI) and RS4;11 as MSI-high.

    Next-generation characterization of the Cancer Cell Line Encyclopedia · Ghandi M, Huang FW, Jané-Valbuena J, et al. · Nature 2019
    PMID 31068700DOI
    What this evidence cannot show

    The MSI calls for these two specific lines were read from their Cellosaurus records, which attribute them to this paper (and, for RS4;11, additionally to Sanger). The paper itself was not read and the abstract was not read in full, so the classifier's threshold and error rate are not established here. An MSI call from sequencing signatures is a computational classification, not a mismatch-repair assay.

    Evidence that argues against this (1)
    • MSI status in a long-passaged line can be acquired in culture rather than inherited from the patient, in which case it says nothing about the disease and everything about the model — which is the interpretation this entry adopts. But the converse would be more interesting and is not excluded: if MSI was present in the patient, then a subset of KMT2A-rearranged ALL has a mismatch-repair phenotype that the low-mutation-burden consensus would have missed. Tomoyasu C et al. 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Fetal-specific gene-expression programmes are maintained in MLL-AF4 infant ALL but not in MLL-AF4 childhood ALL, so a KMT2A::AFF1 line derived from a child is not transcriptionally equivalent to the infant disease even though it carries the same fusion.

    PMID 34824279DOI
    What this evidence cannot show

    Abstract read; full text not read. The comparison is between primary infant and childhood MLL-AF4 ALL, not between cell lines, so applying it to SEM assumes the line retains its donor's age-appropriate programme through many passages — which is exactly what long-term culture might erase in either direction.

    Evidence that argues against this (1)
    • If culture erases the age-specific programme, then SEM being from a 5-year-old is irrelevant, and an infant-derived line in culture would be no better. That would make the argument for infant lines much weaker and would redirect the whole model problem toward primary material and xenografts rather than toward new lines. Cruickshank MN et al. 2017
Connected to