GenestrongQ46-cooperating-lesions

KRAS - the commonest co-mutation in a disease that has almost none, and it is usually in a minority of the cells

also called KRAS · KRAS2 · RASK2 · K-Ras 2 · Ki-Ras · c-K-ras · P01116

KRAS is a switch that tells a blood cell to grow, and in infant leukaemia driven by a KMT2A fusion it is the extra mutation found most often. But in most patients only a minority of the leukaemia cells carry it, and experiments show that switching KRAS on is not enough to start this leukaemia - it changes where the cells travel rather than whether they are cancerous.

The same thing, for a scientist

KRAS (P01116, 189 aa) is the most frequently mutated single gene besides KMT2A itself in infant KMT2A-rearranged ALL, and until now the graph covered it only inside a pathway-level entry. Three things about it are specific to this disease and are not inherited from RAS biology in general. **Frequency.** Recurrent in 4 of 22 whole-genome-sequenced infant MLL-R ALL cases, within an overall kinase/PI3K/RAS mutation rate of 22/47 (47%) [E1]; about 24% of t(4;11) infants by Sanger sequencing of a 109-patient Interfant cohort [E3]; and in an independent 124-patient Interfant-99/06 series recurrent mutation was found *exclusively* in K-RAS and N-RAS [E2]. **Clonality.** 65% (20/31) of the activating kinase/PI3K/RAS mutations sat at mutant allele frequency below 30% despite 92% mean tumour purity, and where a case carried two such mutations, in 6 of 8 cases *each* was below 30% - they were in different subclones [E1]. **Sufficiency.** KRAS(G12V) plus MLL-AF4 in human cord-blood CD34+ HSPCs did not immortalise or initiate leukaemia; it enhanced extramedullary engraftment, leukocytosis and CNS infiltration, with a migration transcriptional signature [E5]. On that evidence a KRAS mutation here is better read as a **dissemination and maintenance modifier carried by a subclone** than as a second driver. What keeps that reading honest is the clinical data: RAS mutation was an independent predictor of poor outcome (HR 3.194, 95% CI 1.211-8.429, P = 0.019) in the Interfant cohort [E3], and KRAS mutation is an independent adverse factor in adult KMT2A-rearranged AML (HR 2.0, P = 0.005) [E6].

What it looks like

loading AF-P01116-F1…

AlphaFold AF-P01116-F1 ↗

colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)

I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper

How the accession was resolved, and why this entry asserts a length

gene_exact:KRAS AND organism_id:9606 AND reviewed:true, requesting gene_primary and length, returns exactly one reviewed human record:

accession	primary	synonyms	protein name	length
P01116	KRAS	KRAS2 RASK2	GTPase KRas (K-Ras 2) (Ki-Ras) (c-K-ras)	189

One row, primary name equal to the symbol asked for. That is the clean case. The sibling entry gene-nras in this beat is not clean and says so, and gene-cdkn2a returns two reviewed records that both carry the requested symbol as primary. Accession and length are both asserted in data so verify-accessions.py fails a swap that reading a protein name back would not catch.

No domain coordinates are recorded. UniProt P01116 carries no Domain features; the G-domain and the switch regions are annotated as regions and nucleotide-binding sites. The project rule is to leave a domain out rather than guess a boundary, because a domain with a wrong or null coordinate is invisible to every figure while looking present in the data.

Frequency in three cohorts, and what each instrument could see

cohort	n	assay	KRAS finding
St. Jude PCGP discovery 1	22 infant MLL-R ALL	WGS, mean haploid coverage 39x	KRAS recurrent, n = 4; NRAS n = 2; FLT3, NF1, PTPN11, PIK3R1 non-recurrent
St. Jude combined 1	47 infant MLL-R ALL	plus a 232-gene targeted panel on 43 further infant ALL	any kinase/PI3K/RAS activating mutation in 22/47 (47%)
Interfant 99/06 3	109 infant ALL	Sanger sequencing of NRAS, KRAS, BRAF	RAS mutation about 14% overall, about 24% in t(4;11) (P = 0.04)
Interfant 99/06 2	124 infant B-ALL	genome-wide multi-layer	recurrent mutation found exclusively in K-RAS and N-RAS

The instruments differ in a way that matters and is usually elided. Conventional Sanger sequencing 3 has a detection floor around 15-20% variant allele fraction, so a cohort screened that way cannot see the subclones that WGS-plus-deep-capture 1 and duplex sequencing 4 report. A frequency of 14-24% from Sanger and 47% from deep sequencing are not in conflict; they answer different questions, and only one of them is about what fraction of patients carry any RAS-pathway lesion at all.

Driver, passenger, or subclonal marker

The honest answer is none of the three cleanly, and the disagreement is the informative part.

Against driver. KRAS(G12V) does not cooperate with MLL-AF4 to immortalise or transform human cord-blood CD34+ HSPCs, and intrabone transplantation of the two together did not initiate leukaemia 5. Two thirds of these mutations are subclonal at diagnosis 1. In matched diagnosis/relapse pairs the pathway mutation was lost in 2 of 5 cases and fell from 39% to 15% MAF in a third 1, and an independent cohort states plainly that the KRAS/NRAS mutations were subclonal and frequently lost at relapse 2. A lesion the leukaemia discards under therapy is not maintaining it.

Against passenger. Every activating kinase/PI3K/RAS mutant allele was expressed, irrespective of its allele frequency, against only 38-48% of other non-silent mutations in the same cases 1; gene-set enrichment in the MLL-AFF1 cases showed RAS-activation expression signatures 1. A passenger is not selected for expression. Clinically, RAS mutation carried HR 3.194 for poor outcome in Interfant 3, RAS-mutant infants had higher white cell counts and greater in-vitro glucocorticoid resistance 3, and in adult KMT2A-r AML KRAS mutation is independently adverse 6.

For subclonal marker with a phenotype. 5 is the observation that reconciles them. KRAS activation changed where the cells went - extramedullary engraftment, leukocytosis, CNS infiltration, a migration signature - rather than whether they were leukaemic. A mutation conferring dissemination on a fraction of cells would produce exactly this pattern: subclonal, expressed, prognostically adverse, and dispensable at relapse.

A KRAS-KMT2A reciprocal fusion, reported once

In the same WGS cohort, one of the predicted in-frame reciprocal products involved KRAS: KRAS-MLL 1. This is a structural event, not a point mutation, and it is reported as a WGS prediction in a single case. It is recorded here because a reader enumerating KRAS lesions in this disease from point-mutation tables would never find it, and because this graph holds a reciprocal-product beat that it connects to. It is not evidence of a recurrent mechanism and nothing should be built on it.

What a KRAS mutation licenses therapeutically, and what it does not

MEK inhibitors (trametinib, selumetinib, MEK162/binimetinib) severely impair primary RAS-mutant KMT2A-rearranged infant ALL cells in vitro, reduce p-ERK, induce apoptosis, and synergise with prednisolone - in both RAS-mutant and RAS-wild-type cells 7. That last clause weakens the biomarker logic rather than strengthening it: if the synergy does not require the mutation, the mutation is not what selects the patient.

The strongest test so far is in AML, not infant ALL: revumenib plus selumetinib was synergistic in KMT2A-r RAS-mutant AML lines and PDXs, reduced leukaemia burden in vivo, and did not improve overall survival over menin inhibition alone 8. The authors state the combination will require further optimisation. That is the cleanest available answer to whether adding a RAS-pathway drug to a menin inhibitor buys anything, and today the answer is: not yet, in that model.

And the route to the patients is narrow. The one open paediatric MEK-inhibitor protocol for relapsed or refractory haematological malignancy sets a minimum age of 1 year 9, which is above the modal age of infant KMT2A-rearranged ALL.

Facts and figures

uniprot	P01116
gene symbol	KRAS
length AA	189
alphafold ID	AF-P01116-F1
cytoband	12p12.1
resolution note	gene_exact:KRAS AND organism_id:9606 AND reviewed:true with fields=accession,gene_primary,gene_synonym,protein_name,length returned exactly one row: P01116, primary gene name KRAS, 189 aa. Fetched from the UniProt REST API on 2026-09-20. Synonyms KRAS2, RASK2.
domains note	UniProt P01116 carries no Domain features; the G-domain and switch I/II are annotated as regions and binding sites. Per the project join rule no domain is recorded rather than a guessed boundary.
druggability	Indirectly drugged in this setting. No KRAS G12C or G12D allele-selective inhibitor has been reported in infant KMT2A-rearranged ALL, and the mutant alleles in this disease are not predominantly G12C. The practical route is downstream MEK inhibition, which works ex vivo irrespective of RAS genotype and has one open paediatric trial with a 1-year age floor.
subnuclear localization	NOT APPLICABLE, AND THIS IS A BOUNDED REJECTION RATHER THAN A GAP. KRAS is a lipid-anchored membrane-associated GTPase: UniProt P01116 annotates cell membrane with a LIPID-ANCHOR topology, endomembrane system and cytoplasm, and NO nuclear location at all - not experimental, not inferred, not by similarity [E_q77_loc1]. Asking for a subnuclear compartment for KRAS is the wrong question, and this entry says so rather than recording an empty field or approximating one. THE RIGHT SPATIAL QUESTION FOR KRAS IS ONE COMPARTMENT OUT: which membrane, and how it gets there. That question has real, drug-relevant answers - PDE-delta sustains KRAS's spatial organisation by shuttling the farnesylated protein, and blocking that interaction suppresses oncogenic signalling BY ALTERING LOCALISATION TO ENDOMEMBRANES [E_q77_loc2]; and GPR31 acts as a secretory-pathway chaperone whose silencing partially MISLOCALISES KRAS4B and slows KRAS-dependent tumour growth [E_q77_loc3]. Two independent programmes in which the therapeutic mechanism IS a localisation change.
shown for fusion	NO for the PROTEIN - KRAS's subcellular localisation has never been measured in a KMT2A-rearranged cell, an MLL-AF4 cord-blood model or any other fusion system. PARTIAL for the CELL, and the distinction is deliberate: in human cord-blood CD34+ HSPCs, KRAS(G12V) plus MLL-AF4 enhanced extramedullary engraftment, leukocytosis and CNS infiltration with a migration signature [E5], which is a statement about where the leukaemic cells go, not about where the protein sits. Nobody has imaged KRAS in a blast.

domains

[]

cooperating lesion profile

{
  "disease": "infant KMT2A-rearranged B-ALL",
  "recurrence_wgs_discovery": "4 of 22 cases",
  "pathway_level_frequency": "22/47 (47%) with any kinase/PI3K/RAS activating mutation",
  "sanger_frequency_t_4_11": "about 24% of t(4;11) infants in a 109-patient cohort",
  "clonality": "65% (20/31) of pathway mutations at MAF below 30% with 92% mean tumour purity",
  "relapse_behaviour": "lost in 2 of 5 matched pairs, declined in a third; stated as frequently lost at relapse in an independent 124-case cohort",
  "sufficiency": "KRAS(G12V) plus MLL-AF4 does not initiate leukaemia in human cord-blood CD34+ HSPC",
  "verdict": "subclonal dissemination and maintenance modifier rather than an initiating second driver"
}

essentiality

{
  "note": "No DepMap figure is quoted in this entry. A dependency score for KRAS across leukaemia lines would measure dependence in lines that carry a clonal KRAS mutation, which is the opposite of the subclonal situation this entry is about, and quoting it here would read as support for a claim it cannot support."
}

depmap kmt2a selectivity

{
  "chronos_mean_kmt2a_rearranged": -0.7084,
  "chronos_mean_other_leukaemia": -0.6426,
  "difference": -0.0657,
  "pan_essentiality_all_lines": -0.7249,
  "selectivity_rank": 2464,
  "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."
}

localization

{
  "subcellular": "PLASMA MEMBRANE BY LIPID ANCHOR, plus endomembrane system and cytoplasm - and no nuclear pool of any kind. UniProt P01116 fetched 2026-09-21 annotates 'Cell membrane' with topology LIPID-ANCHOR at ECO:0000269 on four references (PubMed:22431598, 23524970, 23698361, 29239724), 'Endomembrane system' at ECO:0000269 (PubMed:29239724) and 'Cytoplasm' at ECO:0000269 (PubMed:23524970, 23698361). A separate ISOFORM 2B comment annotates cell membrane with a lipid anchor at ECO:0000269 (PubMed:28619714). Two of the four references for the main membrane call are not localisation studies [E_q77_loc4], and one is specifically about the OTHER splice isoform, K-Ras4A [E_q77_loc5].",
  "subnuclear": "NOT APPLICABLE, AND THIS IS A BOUNDED REJECTION RATHER THAN A GAP. KRAS is a lipid-anchored membrane-associated GTPase: UniProt P01116 annotates cell membrane with a LIPID-ANCHOR topology, endomembrane system and cytoplasm, and NO nuclear location at all - not experimental, not inferred, not by similarity [E_q77_loc1]. Asking for a subnuclear compartment for KRAS is the wrong question, and this entry says so rather than recording an empty field or approximating one. THE RIGHT SPATIAL QUESTION FOR KRAS IS ONE COMPARTMENT OUT: which membrane, and how it gets there. That question has real, drug-relevant answers - PDE-delta sustains KRAS's spatial organisation by shuttling the farnesylated protein, and blocking that interaction suppresses oncogenic signalling BY ALTERING LOCALISATION TO ENDOMEMBRANES [E_q77_loc2]; and GPR31 acts as a secretory-pathway chaperone whose silencing partially MISLOCALISES KRAS4B and slows KRAS-dependent tumour growth [E_q77_loc3]. Two independent programmes in which the therapeutic mechanism IS a localisation change.",
  "how_shown": "biochemical screening, structure-based optimisation and imaging of KRAS distribution after PDE-delta inhibition [E_q77_loc2]; a genome-wide screen for genes required for KRAS4B membrane association, with ER-retention and silencing experiments [E_q77_loc3]; NMR fragment screening for the SOS-inhibition reference and an epidermal-differentiation study for the fourth, neither of which is a localisation study [E_q77_loc4]; fatty-acylation and SIRT2 experiments on K-Ras4A [E_q77_loc5]",
  "systems": "human pancreatic ductal adenocarcinoma, colorectal and other solid-tumour lines, plus keratinocytes for one reference. NO haematopoietic system, no infant ALL material, nothing KMT2A-rearranged.",
  "shown_for_fusion_localisation": "NO - nobody has measured KRAS protein localisation in any fusion context. There IS a fusion-context SPATIAL result and it is about the CELLS, not the protein, and the distinction is worth keeping: KRAS(G12V) with MLL-AF4 in human cord-blood CD34+ HSPCs enhanced EXTRAMEDULLARY engraftment, leukocytosis and CNS infiltration with a cell-migration transcriptional signature, without initiating leukaemia [E5]. That is a tissue-localisation phenotype, not a subcellular one, and a graph that stores both under the word 'localisation' will conflate them.",
  "node": "localization-plasma-membrane-leukaemic-surface-antigens",
  "the_comparator_already_in_this_graph": "FLT3-ITD is the worked precedent and it sits one entry away: localization-flt3-itd-intracellular-retention-er-versus-golgi records a receptor whose oncogenic signalling changes because the mutant protein is RETAINED IN AN INTRACELLULAR COMPARTMENT rather than reaching the surface. KRAS is the same class of problem approached from the drug side: PDE-delta and GPR31 inhibition work by moving KRAS off the plasma membrane. FLT3 and KRAS are the two commonest signalling co-mutations in this disease, and for BOTH of them the actionable spatial variable is membrane trafficking rather than anything nuclear."
}

cell type

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "KRAS is P01116, 189 aa, primary gene name KRAS, synonyms KRAS2 and RASK2 (UniProt, 2026-09-21, one clean reviewed row). data/symbol-collisions.json was grepped BEFORE the first query and THE RELEVANT ROW ALREADY EXISTED under the token HRAS1: KRAS, NRAS and HRAS are ALL EXACTLY 189 aa, HRAS1 is NRAS's only listed synonym while also being HRAS's synonym, and the registry marks length_defence_usable FALSE and block_autogeneration TRUE. Brief rule 68 in full force - only gene_primary discriminates, and a paper saying 'RAS mutation' without a gene is a PATHWAY claim. Everything in this block that says KRAS is P01116.",
  "expressed_in": {
    "class": 1,
    "summary": "ubiquitous. As with TP53, the cell-type question for KRAS in this disease is not where it is expressed but in WHICH CELLS THE MUTANT ALLELE SITS - and this entry's clonality block already answers that: 65% of the pathway mutations are below 30% mutant allele frequency at 92% tumour purity, so the mutation is in a MINORITY SUBCLONE in most patients."
  },
  "required_in": {
    "class": 2,
    "claim": "NOT REQUIRED, AND DELIBERATELY NOT QUOTED FROM DEPMAP. This entry's own essentiality block already refuses a DepMap figure for a stated reason - a dependency score measures lines with a CLONAL KRAS mutation, which is the opposite of the subclonal situation here - and that refusal is correct and is left in place. What is established instead is a SUFFICIENCY negative in the right cell type: KRAS(G12V) does not cooperate with MLL-AF4 to immortalise or transform HUMAN CORD-BLOOD CD34+ HSPCs, and intrabone transplantation of the two together did not initiate leukaemia [E5].",
    "pmids": [
      "26837759"
    ],
    "system_named_in_the_claim": "human cord-blood CD34+ haematopoietic stem and progenitor cells with MLL-AF4 and KRAS(G12V), xenografted - the closest available model of the index cell of origin",
    "nodes": [
      "cell-type-hsc-bone-marrow"
    ]
  },
  "context": "human cord-blood HSPC xenograft; infant KMT2A-rearranged patient cohorts for the clonality"
}

cell state

{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "maintains_state": {
    "class": 3,
    "states": [
      "cell-state-early-lymphocyte-precursor",
      "cell-state-cns-niche-adapted"
    ],
    "claim": "TWO STATES, AND NEITHER IS A STATE KRAS INSTALLS - KRAS MARKS ONE AND MODIFIES THE OTHER. (1) RAS BURDEN TRACKS THE MOST IMMATURE AND MOST REFRACTORY SUBGROUP: in 84 infants with KMT2A-rearranged leukaemia profiled by RNA-seq, methylation array, whole-exome and targeted deep sequencing, five integrative clusters correspond to stages of B-lymphopoietic and early hemato-endothelial development, the NUMBER of RAS-pathway mutations predicts prognosis, and the most refractory cluster - IC2, IRX-type UNDIFFERENTIATED - carries RAS-pathway mutations at 100% FREQUENCY and with the heaviest burden [E_q77_cs1]. (2) KRAS ACTIVATION MOVES CELLS INTO AN EXTRAMEDULLARY AND CNS COMPARTMENT: KRAS(G12V) with MLL-AF4 enhances extramedullary engraftment, leukocytosis and CNS infiltration with a cell-migration transcriptional signature, without initiating leukaemia [E5].",
    "pmids": [
      "36042201",
      "26837759"
    ],
    "system_named_in_the_claim": "84 human infants with KMT2A-rearranged leukaemia for the cluster result; human cord-blood CD34+ HSPC xenograft for the dissemination result - both in the index disease, which is rare in this graph"
  },
  "where_this_conclusion_already_lives": "Per brief rule 82, PMID 36042201 is cited in 17 entries of this graph and is load-bearing. Its integrative-cluster conclusion is already written in cell-state-early-lymphocyte-precursor, and this block deliberately POINTS AT that entry rather than restating it. What is added here is the specific KRAS-side reading: the association runs from RAS-mutation BURDEN to the undifferentiated cluster, which is a dose statement about a subclonal lesion, and it is the only place in this entry where a RAS lesion is tied to a differentiation state rather than to an outcome.",
  "the_direction_is_not_established": "100% RAS-mutation frequency in the most undifferentiated cluster is an association in 84 patients. It does not show that RAS mutations CAUSE the undifferentiated state, and the sufficiency experiment in this entry points the other way - KRAS(G12V) plus MLL-AF4 did not transform cord-blood HSPCs at all. The reading this entry can defend is that an undifferentiated, IRX-type background tolerates or selects for RAS-pathway lesions, not that the lesion produces the background.",
  "confidence": "moderate",
  "caveats": "The cluster association is cross-sectional in 84 patients with no functional arm; IC2 is one of five clusters so the 100% figure rests on a subgroup whose size is not restated here. The dissemination result is a xenograft with over-expressed KRAS(G12V) in cord-blood cells, which models the index cell of origin better than anything else available and is still an over-expression system. Neither study measures KRAS protein localisation. And this entry's own clonality data is the standing caveat on every state claim about KRAS here: in most patients the mutation is in a minority of cells, so a bulk state association is an association between a subclone's genotype and the whole sample's phenotype.",
  "counter_evidence": [
    {
      "note": "The strongest counter-argument to reading RAS burden as a driver of the refractory immature state is inside this entry already: the mutations are subclonal in two thirds of cases, were LOST in 2 of 5 matched diagnosis-relapse pairs and fell from 39% to 15% allele frequency in a third, and an independent 124-case Interfant cohort states plainly that they are frequently lost at relapse. A lesion the leukaemia discards under therapy is not maintaining the state that makes it refractory. The reconciliation the data supports is the dissemination reading - KRAS changes where a subclone goes, not what the bulk population is - and that is a claim about tissue compartments, not about differentiation stage.",
      "citation": {
        "title": "Activated KRAS Cooperates with MLL-AF4 to Promote Extramedullary Engraftment and Migration of Cord Blood CD34+ HSPC But Is Insufficient to Initiate Leukemia.",
        "authors": "Prieto C, Stam RW, Agraz-Doblas A, Ballerini P, Camos M, Castaño J, Marschalek R, Bursen A, Varela I, Bueno C, Menendez P",
        "year": 2016,
        "venue": "Cancer Res",
        "pmid": "26837759",
        "doi": "10.1158/0008-5472.CAN-15-2769",
        "type": "paper"
      }
    }
  ]
}

How this could bite us back

A subclonal target is the worst kind of target. Inhibiting MEK in a patient whose KRAS mutation sits in 15% of cells applies selection pressure to the 85% that do not carry it, and the clone that survives is by construction the one the drug was not chosen for.
Ex-vivo MEK sensitivity does not track RAS genotype - RAS-wild-type samples were both sensitive and resistant, with no measurable downstream over-activation to explain either - so genotype-based selection for a MEK inhibitor is not supported by the data usually cited for it.
The prednisolone synergy that makes MEK inhibition attractive occurs in RAS-wild-type cells too, so the combination's benefit, if real, would not be attributable to the co-mutation and would not require the co-mutation to be present.
Adding a MEK inhibitor to a menin inhibitor reduced leukaemia burden without extending survival in KMT2A-r RAS-mutant AML xenografts. A combination that moves a burden endpoint and not a survival endpoint has the shape of a transient response.
The patients are infants. The one open paediatric MEK-inhibitor protocol starts at 1 year of age, above the modal age of this disease, and the one MAPK master protocol carrying a SHP2 inhibitor was withdrawn for 'Sponsor decision.' at an 18-year floor. The pharmacology exists; the trial route to these patients does not.
MEK inhibition in infants carries a developmental-exposure question the adult literature cannot answer: RAS/MAPK signalling is required for normal growth, and the germline RASopathies show what chronic pathway perturbation does to a developing child.

Still open

Is the KRAS-mutant subclone the clone that seeds CNS disease? The cord-blood model predicts it, no patient dataset has tested it, and CSF-compartment or single-cell genotyping would answer it directly.
Do the KRAS mutations 'lost at relapse' by bulk sequencing persist below 1% by duplex sequencing, or are they genuinely eliminated? The two published answers used instruments whose detection floors differ roughly a hundredfold.
Does MEK inhibition select for the KRAS-wild-type clone in vivo? Nothing published tracks clone-level allele frequency under MEK inhibition in this disease.
Why is KRAS clonal and prognostically heavy in adult KMT2A-rearranged AML and subclonal in infant KMT2A-rearranged ALL? The difference may be cell of origin rather than anything about the gene.

Every source behind this entry (15)

E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In whole-genome sequencing of infant MLL-rearranged ALL, KRAS was recurrently mutated (4 of 22 discovery cases), activating kinase/PI3K/RAS mutations were present in 22/47 (47%) of infant MLL-R cases, 65% (20/31) of those mutations had mutant allele frequency below 30% despite 92% mean tumour purity, and all were expressed.
The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. · Andersson AK, Ma J, Wang J, et al. · Nature Genetics 2015
PMID 25730765 DOI
What this evidence cannot show
The discovery cohort is 22 cases at 39x mean haploid coverage; the 47% figure combines WGS with a 232-gene targeted panel, so it is a panel-bounded estimate for genes on that panel and not a genome-wide rate. Deep custom-capture confirmation of subclonality was performed on five cases only, at an average of 437x. Full text read via efetch db=pmc (PMC4553269) after Europe PMC fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- Targeted duplex sequencing of 20 infants and 7 children with KMT2A-r ALL found that 62% of identified cancer-associated mutations were below the resolution of standard NGS, and that only 33 of 87 exceeded 2% cellular prevalence - so both the burden and the clonality structure reported at 39x WGS are instrument-bounded rather than final. Pilheden M et al. 2022
E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 124 uniformly diagnosed and treated de novo infant B-ALL cases from Interfant-99/06, recurrent mutations were found exclusively in K-RAS and N-RAS, were subclonal, and were frequently lost at relapse.
Unraveling the cellular origin and clinical prognostic markers of infant B-cell acute lymphoblastic leukemia using genome-wide analysis. · Agraz-Doblas A, Bueno C, Bashford-Rogers R, et al. · Haematologica 2019
PMID 30679323 DOI
What this evidence cannot show
Abstract read for this entry. The phrase 'frequently lost at relapse' appears without the number of matched pairs in the abstract, so the strength of that specific claim rests on the concordant matched-pair count in [E1] (lost in 2 of 5). The cohort is infant B-ALL and predominantly t(4;11), so it says nothing about KMT2A-rearranged AML.
Evidence that argues against this (1)
- The same Lund group's longitudinal work shows relapse clones persisting as minor subclones from diagnosis at 0.06-0.34% of cells, so 'lost at relapse' measured by bulk sequencing can mean 'fell below the detection floor' rather than 'eliminated'. Pilheden M et al. 2022
E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Sanger sequencing of NRAS, KRAS and BRAF in 109 infant ALL patients found RAS mutations in about 14% overall and about 24% of t(4;11) cases, and RAS mutation was an independent predictor of poor outcome (HR 3.194, 95% CI 1.211-8.429, P = 0.019), with higher white cell counts at diagnosis and greater in-vitro glucocorticoid resistance.
Frequencies and prognostic impact of RAS mutations in MLL-rearranged acute lymphoblastic leukemia in infants. · Driessen EM, van Roon EH, Spijkers-Hagelstein JA, et al. · Haematologica 2013
PMID 23403319 DOI
What this evidence cannot show
Conventional Sanger sequencing has a variant-allele-fraction floor of roughly 15-20%, so this cohort is blind to exactly the subclonal mutations that dominate the deep-sequencing picture; its 14-24% is a clonal-mutation frequency, not a total. The hazard ratio comes from a subgroup of a retrospective cohort, with a wide confidence interval and no correction for multiple testing.
Evidence that argues against this (1)
- In the St. Jude cohort the trend to worse EFS and OS was present but did not reach significance, and was concentrated in cases with MAF above 30%; the authors also report that within MLL-AFF1 cases those with an activating mutation were on average YOUNGER, which is the opposite direction of confounding from the Interfant analysis. Andersson AK et al. 2015
E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Targeted duplex sequencing of 20 infants and 7 children with KMT2A-rearranged ALL found that the mutations exceeding 2% cellular prevalence were almost all in PI3K/RAS genes (31 of 33), typically KRAS or NRAS, and that drug-resistant clones present at 0.06-0.34% of diagnostic cells expanded at relapse.
Duplex Sequencing Uncovers Recurrent Low-frequency Cancer-associated Mutations in Infant and Childhood KMT2A-rearranged Acute Leukemia. · Pilheden M, Ahlgren L, Hyrenius-Wittsten A, et al. · HemaSphere 2022
PMID 36204688 DOI
What this evidence cannot show
27 patients, targeted panel rather than genome-wide, so the mutation set is bounded by panel design and cannot establish total burden. Duplex sequencing achieves a very low error floor but cannot by itself prove that a 0.1% variant sits in a leukaemic cell rather than a normal haematopoietic one. Abstract read; full text not retrieved for this entry.
Evidence that argues against this (1)
- The same paper reports that infants had FEWER such cancer-associated mutations than the older children, so a deeper instrument did not erase the age-specific quietness of the infant genome - it lowered the floor without reversing the direction. Pilheden M et al. 2022
E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Activated KRAS(G12V) did not cooperate with MLL-AF4 to immortalise human cord-blood CD34+ HSPCs or to initiate leukaemia after intrabone marrow transplantation, but it enhanced extramedullary engraftment, leukocytosis and CNS infiltration with a cell-migration transcriptional signature.
Activated KRAS Cooperates with MLL-AF4 to Promote Extramedullary Engraftment and Migration of Cord Blood CD34+ HSPC But Is Insufficient to Initiate Leukemia. · Prieto C, Stam RW, Agraz-Doblas A, et al. · Cancer Research 2016
PMID 26837759 DOI
What this evidence cannot show
A negative result in one model system. Cord-blood CD34+ HSPCs are not fetal-liver HSPCs, which this graph's own cell-of-origin work favours as the target cell; lentiviral MLL-AF4 is expressed from a heterologous promoter rather than the endogenous locus; and follow-up in immunodeficient mice may be short relative to the latency of a weakly transforming combination. Absence of leukaemia in this model is not proof that KRAS cannot cooperate in the right cell.
Evidence that argues against this (1)
- Prognostic data run the other way: RAS mutation was independently associated with poor outcome in an Interfant cohort, which is hard to explain if the lesion is biologically inert in this disease. Driessen EM et al. 2013
E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In 185 adult KMT2A-rearranged AML patients with mutational data from the HARMONY database, KRAS mutation was an independent adverse prognostic factor for overall survival in intensively treated patients (HR 2.0, P = 0.005), and in de novo AML under 60 KRAS or TP53 mutation predicted a lower complete remission rate (50% vs 86%, P < 0.001) and shorter survival (median 7 vs 30 months).
Rearrangements involving 11q23.3/KMT2A in adult AML: mutational landscape and prognostic implications - a HARMONY study. · Hernandez-Sanchez A, Gonzalez T, Sobas M, et al. · Leukemia 2024
PMID 38965370 DOI
What this evidence cannot show
Adults, AML, panel-based NGS on a registry cohort assembled across trials - none of which is infant KMT2A-rearranged ALL. It is included here because the age contrast is itself the finding: the same gene is clonal and prognostically heavy in adults and subclonal in infants. Abstract read; full text not retrieved.
Evidence that argues against this (1)
- Paediatric AML across nearly 1,000 COG participants shows the age split from the other side: DNMT3A and TP53, common in adults, were conspicuously absent from virtually all paediatric cases, so adult KMT2A-r co-mutation patterns must not be transferred to children. Bolouri H et al. 2018
E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MEK inhibitors trametinib, selumetinib and MEK162 impaired primary RAS-mutant KMT2A-rearranged infant ALL cells in vitro, reduced p-ERK and induced apoptosis, and synergised with prednisolone in both RAS-mutant and RAS-wild-type cells.
MEK inhibition is a promising therapeutic strategy for MLL-rearranged infant acute lymphoblastic leukemia patients carrying RAS mutations. · Kerstjens M, Driessen EM, Willekes M, et al. · Oncotarget 2017
PMID 27588400 DOI
What this evidence cannot show
Ex-vivo primary-cell cytotoxicity on small numbers of samples with no in-vivo or clinical arm. The authors report both sensitive and resistant samples among RAS-wild-type cases with no measurable downstream over-activation to explain either, which means ex-vivo MEK sensitivity is not predicted by RAS genotype. PubMed gives the publication year as 2017; several sibling entries in this graph record 2016 for the same DOI.
Evidence that argues against this (1)
- The prednisolone synergy occurred in RAS-wild-type cells too, so this result does not support RAS genotype as the patient-selection biomarker it is routinely cited as supporting. Kerstjens M et al. 2017
E8moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In 1,750 patients enrolled on Children's Oncology Group trials, RAS-pathway mutations were present in 43% of AML cases and, in KMT2A-rearranged AML, were associated with a lower complete remission rate, worse event-free and overall survival and early relapse; combining the MEK1/2 inhibitor selumetinib with the menin inhibitor revumenib was synergistic in vitro and reduced leukaemia burden in KMT2A-r RAS-mutant PDX models in vivo, but did not improve overall survival compared with menin inhibition alone.
Combining menin and MEK inhibition to target poor prognosis KMT2A-rearranged RAS pathway-mutant acute myeloid leukemia. · Scheidegger N, Schneider C, Alexe G, et al. · Blood Advances 2026
PMID 42085603 DOI
What this evidence cannot show
AML, not infant ALL, and the clinical association is a retrospective analysis of trial cohorts (NCT00070174, NCT00372593, NCT01371981, NCT00002798). The in-vivo survival result is the load-bearing one and it is negative; the synergy is a viability and cell-cycle readout. Abstract read; full text not retrieved for this entry.
Evidence that argues against this (1)
- In patient-derived xenografts from multiply pretreated KMT2A::AFF1 ALL, RAS-pathway and TP53 comutations emerged alongside acquired menin-inhibitor resistance, yet neither was sufficient to induce menin-inhibitor resistance in vitro - so RAS co-mutation is not established as the thing that needs co-targeting. Mahdavi L et al. 2026
E9established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.trial_registry
HEM iSMART subprotocol D, trametinib with dexamethasone, cyclophosphamide and cytarabine for relapsed or refractory paediatric haematological malignancy, sets a minimum enrolment age of 1 year with a planned enrolment of 26.
International Proof of Concept Therapeutic Stratification Trial of Molecular Anomalies in Relapsed or Refractory HEMatological Malignancies in Children, Subprotocol D: Trametinib + Dexamethasone + Cyclophosphamide and Cytarabine in Pediatric Patients With Relapsed or Refractory Hematological Malignancies · HEM iSMART consortium · ClinicalTrials.gov 2023
NCT05658640
What this evidence cannot show
Registry record retrieved from the ClinicalTrials.gov v2 API on 2026-09-20: status RECRUITING, phase 1/2, ages 1-21 years, start 2023-11-14, whyStopped null. A registry states eligibility, not practice - a trial with a 1-year floor does not exclude compassionate or named-patient use below it, and eligibility can be amended. The claim that this is the only such open trial is bounded by the query actually run (condition 'acute lymphoblastic leukemia', intervention 'trametinib'), which returned three studies, the other two being a non-randomised drug-sensitivity assay study and an adult-only terminated study.
Evidence that argues against this (1)
- The one registered trial of MAPK-pathway agents in haematological malignancy that carried a SHP2 inhibitor (ERAS-601) was WITHDRAWN with the registry reason 'Sponsor decision.', at a minimum age of 18 years - so the adult route to SHP2 inhibition in leukaemia is not open either. Erasca et al. 2022
E_q77_loc1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
KRAS is the only protein in this tranche with NO nuclear annotation of any kind. UniProt P01116 (fetched 2026-09-21) annotates exactly three unscoped locations - 'Cell membrane' with topology LIPID-ANCHOR, 'Endomembrane system' and 'Cytoplasm' - all at ECO:0000269, plus one isoform-scoped comment for isoform 2B (cell membrane, lipid anchor, ECO:0000269). There is no nucleus line, experimental, inferred or by similarity.
UniProtKB P01116 (RASK_HUMAN) - SUBCELLULAR LOCATION comment block with evidence codes, including the isoform 2B comment · UniProt Consortium 2026
What this evidence cannot show
A measurement of the database, read from the comment array with evidence codes and molecule scoping attached on 2026-09-21. The absence of a nuclear annotation is recorded as an absence in the record, not as proof that no nuclear pool exists anywhere in the literature - the distinction this project's rule about fallback paths insists on. What it does license is the framing: for KRAS the subnuclear question is not the question, and this entry states that as a bounded rejection rather than leaving an empty field.
Evidence that argues against this (1)
- An ECO:0000269 code on all three compartments makes this record look better evidenced than it is. Two of the four references behind the main membrane call are not localisation studies - an NMR fragment-screening paper about SOS-mediated nucleotide exchange, and an epidermal-differentiation paper about desmoglein-1 and Erbin - and a third is specific to the K-Ras4A splice isoform rather than to the canonical 4B. The compartment is not in doubt; the provenance of the annotation is thinner than the code implies. Maurer T et al. 2012
E_q77_loc2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KRAS's spatial organisation is itself a drug target and the mechanism is explicitly a localisation change: correct localisation and signalling by farnesylated KRAS is regulated by the prenyl-binding protein PDE-delta, which sustains KRAS's spatial organisation by facilitating its diffusion in the cytoplasm, and small molecules that bind the PDE-delta prenyl pocket with nanomolar affinity suppress oncogenic RAS signalling BY ALTERING ITS LOCALISATION TO ENDOMEMBRANES, inhibiting proliferation of KRAS-dependent pancreatic ductal adenocarcinoma cells in vitro and in vivo.
Small molecule inhibition of the KRAS-PDEδ interaction impairs oncogenic KRAS signalling. · Zimmermann G, Papke B, Ismail S, Vartak N, Chandra A, Hoffmann M, Hahn SA, Triola G, Wittinghofer A, Bastiaens PI, Waldmann H · Nature 2013
PMID 23698361 DOI
What this evidence cannot show
Human pancreatic ductal adenocarcinoma, not leukaemia, and the KRAS alleles are the clonal solid-tumour ones rather than the subclonal lesions this entry is about. Cited here for the SHAPE of the result - that a KRAS-directed therapeutic can work by moving the protein between membranes - not for the compound. No PDE-delta inhibitor has been tested in KMT2A-rearranged leukaemia of any kind.
Evidence that argues against this (1)
- Importing a solid-tumour KRAS strategy into infant KMT2A-rearranged ALL runs straight into this entry's clonality finding: in most patients the KRAS mutation is in a minority subclone and is frequently lost at relapse, so a KRAS-localisation drug would be treating a fraction of the disease. The same objection already sank the biomarker logic for MEK inhibition here, where the synergy with prednisolone did not require the mutation. Prieto C et al. 2016
E_q77_loc3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A genome-wide screen for genes required for KRAS4B membrane association identified the G protein-coupled receptor GPR31, which associates with KRAS4B on cellular membranes in a FARNESYLATION-DEPENDENT fashion; retaining GPR31 on the endoplasmic reticulum inhibits delivery of KRAS4B to the plasma membrane, and silencing GPR31 partially MISLOCALISES KRAS4B, slows growth of KRAS-dependent tumour cells and blocks KRAS-stimulated macropinocytosis - so GPR31 acts as a secretory-pathway chaperone for KRAS4B.
The G protein-coupled receptor GPR31 promotes membrane association of KRAS. · Fehrenbacher N, Tojal da Silva I, Ramirez C, Zhou Y, Cho KJ, Kuchay S, Shi J, Thomas S, Pagano M, Hancock JF, Bar-Sagi D, Philips MR · J Cell Biol 2017
PMID 28619714 DOI
What this evidence cannot show
This is the reference UniProt attaches to the ISOFORM 2B comment, and the paper is about KRAS4B, so the isoform scoping in the record is correct and meaningful - which is the opposite of the TP53 case in this same tranche, where every isoform comment is uncoded. Solid-tumour cell lines; no haematopoietic system. Recorded because a chaperone-dependent delivery step is a second, independent way to make KRAS localisation druggable.
Evidence that argues against this (1)
- Mislocalisation strategies have a documented failure mode in this very graph: FLT3-ITD is RETAINED in the endoplasmic reticulum and signals aberrantly FROM there, so moving an oncoprotein off the plasma membrane does not reliably switch it off. Whether endomembrane-resident KRAS is silent or differently active is precisely the question the PDE-delta and GPR31 papers each answer in their own system and neither answers in a blast. Zimmermann G et al. 2013
E_q77_loc4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Two of the four ECO:0000269 references UniProt gives for KRAS's 'Cell membrane' location are not localisation studies: Maurer et al. 2012 is an NMR fragment-screening paper reporting small-molecule ligands binding a distinct pocket in Ras and inhibiting SOS-mediated nucleotide exchange, and Harmon et al. 2013 is an epidermal-differentiation paper about desmoglein-1/Erbin suppression of ERK activation.
Small-molecule ligands bind to a distinct pocket in Ras and inhibit SOS-mediated nucleotide exchange activity. · Maurer T, Garrenton LS, Oh A, Pitts K, Anderson DJ, Skelton NJ, Fauber BP, Pan B, Malek S, Stokoe D, Ludlam MJ, Bowman KK, Wu J, Giannetti AM, Starovasnik MA, Mellman I, Jackson PK, Rudolph J, Wang W, Fang G · Proc Natl Acad Sci U S A 2012
PMID 22431598 DOI
What this evidence cannot show
Rule 91, and this is the clean statement of it: a curator read AN experiment in each paper, and neither paper is about where KRAS is. The annotation is not wrong - KRAS is lipid-anchored at the plasma membrane and that is not in dispute - but anyone counting four independent experimental sources for the compartment would be counting two studies whose subjects are nucleotide exchange and keratinocyte differentiation. The founding observations for RAS membrane anchoring predate both by decades and are not cited in this record.
Evidence that argues against this (1)
- The two references that ARE localisation studies - the PDE-delta paper and, for isoform 2B, the GPR31 paper - fully support the compartment, so this is a provenance correction and not a challenge to the fact. It is recorded because the same audit on MED26, HOXA10, STAG2, HDAC5 and HDAC9 in this tranche finds records with no experimental code at all, and the contrast only means something if the coded records are read as carefully as the uncoded ones. Zimmermann G et al. 2013
E_q77_loc5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
One of the four membrane references is isoform-specific to the OTHER splice product: Jing et al. 2017 report that SIRT2 and lysine fatty acylation regulate the transforming activity of K-Ras4A, not of the canonical K-Ras4B that P01116 represents.
SIRT2 and lysine fatty acylation regulate the transforming activity of K-Ras4a. · Jing H, Zhang X, Wisner SA, Chen X, Spiegelman NA, Linder ME, Lin H · Elife 2017
PMID 29239724 DOI
What this evidence cannot show
KRAS produces two proteins by alternative fourth exons - 4A and 4B - with different C-terminal membrane-targeting sequences and different trafficking. A reference about 4A supporting the canonical entry's membrane annotation is defensible at the level of 'KRAS is membrane-anchored' and not at the level of 'this is how the canonical protein gets there'. Solid-tumour and biochemical systems; no haematopoietic material. Flagged because isoform conflation is the identifier trap this tranche keeps finding, and here it appears inside a correctly coded record.
Evidence that argues against this (1)
- UniProt does scope the GPR31 reference to isoform 2B, showing the record CAN carry isoform-resolved localisation properly - so the unscoped placement of the K-Ras4A reference is an inconsistency within one record rather than a limitation of the format. The practical consequence for this graph is small, because no claim here turns on 4A versus 4B, but it would not be small for anyone reasoning about C-terminal membrane-targeting chemistry. Fehrenbacher N et al. 2017
E_q77_cs1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 84 INFANTS with KMT2A-rearranged leukaemia profiled by RNA sequencing, methylation array, whole-exome and targeted deep sequencing with single-sample and single-cell inference of haematopoietic differentiation, five robust integrative clusters map onto stages of B-lymphopoietic and early hemato-endothelial development; the NUMBER of RAS-pathway mutations predicts prognosis, and the most refractory cluster - IC2, IRX-type UNDIFFERENTIATED - carries RAS-pathway mutations at 100% frequency with the heaviest burden.
Multi-omics analysis defines highly refractory RAS burdened immature subgroup of infant acute lymphoblastic leukemia. · Isobe T, Takagi M, Sato-Otsubo A, Nishimura A, Nagae G, Yamagishi C, Tamura M, Tanaka Y, Asada S, Takeda R, Tsuchiya A, Wang X, Yoshida K, Nannya Y, Ueno H, Akazawa R, Kato I, Mikami T, Watanabe K, Sekiguchi M, Seki M, Kimura S, Hiwatari M, Kato M, Fukuda S, Tatsuno K, Tsutsumi S, Kanai A, Inaba T, Shiozawa Y, Shiraishi Y, Chiba K, Tanaka H, Kotecha RS, Cruickshank MN, Ishikawa F, Morio T, Eguchi M, Deguchi T, Kiyokawa N, Arakawa Y, Koh K, Aoki Y, Ishihara T, Tomizawa D, Miyamura T, Ishii E, Mizutani S, Wilson NK, Göttgens B, Miyano S, Kitamura T, Goyama S, Yokoyama A, Aburatani H, Ogawa S, Takita J · Nat Commun 2022
PMID 36042201 DOI
What this evidence cannot show
Cross-sectional association in 84 patients with no functional arm - it does not show that RAS mutations cause the undifferentiated state, and this entry's own sufficiency experiment argues they do not. IC2 is one of five clusters and the subgroup size behind the 100% figure is not restated here. Brief rule 82 applies: this paper is cited in 17 entries of this graph and its cluster conclusion is already written in cell-state-early-lymphocyte-precursor, so this item adds only the KRAS-side reading and points at the existing entry. The study is HUMAN and in the INDEX DISEASE, which is rare enough in this graph to be worth noting.
Evidence that argues against this (1)
- This entry's clonality data is the standing objection and it is severe here: 65% of the pathway mutations sit below 30% mutant allele frequency at 92% tumour purity, they are lost in a substantial fraction of matched relapse pairs, and an independent 124-case cohort calls them subclonal and frequently lost. A 100% frequency in a bulk-profiled cluster therefore means 100% of PATIENTS in that cluster have a RAS lesion SOMEWHERE in the sample, not that the cluster's cells carry it. Reading it as a property of the undifferentiated state would be the subclone-to-bulk error this entry exists to prevent. Prieto C et al. 2016

Connected to

related to →KRAS as a co-mutation in KMT2A-r leukemia — the single most frequently co-mutated gene, and it still cannot start the disease
gene-level dossier for the co-mutation entry: accession, clonality, sufficiency
part of →Kinase/PI3K/RAS pathway activation — the only common co-lesion class in KMT2A-r leukemia, and it is mostly subclonal
one of the four RAS-pathway members the pathway-level entry summarises
co-occurs with →Infant acute lymphoblastic leukaemia (ALL diagnosed under 1 year of age)
recurrent in 4/22 WGS cases; subclonal in two thirds
co-occurs with →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
KRAS(G12V) does not cooperate with MLL-AF4 to initiate leukaemia; it confers dissemination
co-occurs with →KMT2A-rearranged acute myeloid leukaemia in children and infants
clonal and independently adverse in adult KMT2A-r AML (HR 2.0), unlike in infants
← co-occurs withNRAS - the other RAS, the one whose 0.3% clone came back, and a UniProt record that lists HRAS1 as its own synonym
correlated in paediatric B-ALL (phi 0.19); the only two recurrently point-mutated genes in infant KMT2A-r ALL
← depends onThe RAS-mutant subclone is where the leukaemia goes, not what it is - a dissemination modifier that a marrow aspirate systematically under-counts
built on the KRAS sufficiency and clonality evidence
← related toTreating a subclonal lesion is selection pressure applied to the clone you did not aim at
why a 15% KRAS mutation is a worse target than no KRAS mutation
related to →The early lymphocyte precursor (ELP) state — what KMT2A-rearranged infant B-ALL scores as against a developing bone marrow reference, and why it is earlier than the pro-B arrest this graph also records
RAS-pathway mutation burden is heaviest, at 100% frequency, in IC2 - the IRX-type undifferentiated, most refractory cluster of infant KMT2A-rearranged leukaemia
related to →The CNS-adapted state — leukaemia-propagating cells reprogrammed by the central nervous system niche, and worse at restarting systemic disease afterwards
KRAS(G12V) with MLL-AF4 in cord-blood CD34+ HSPCs adds extramedullary engraftment and CNS infiltration with a migration signature, without initiating leukaemia
related to →FLT3-ITD signals from inside the cell, not from the surface - and whether it is the ER or the Golgi decides which pathway fires
the two commonest signalling co-mutations in this disease both have membrane-trafficking as their actionable spatial variable, and neither has anything nuclear
related to →The plasma membrane - the one compartment this graph targets with drugs and had never written down
KRAS is the one protein in this tranche with no nuclear pool at all - lipid-anchored at the plasma membrane, so the subnuclear question is rejected rather than left empty
← related toSMYD3 - a lysine methyltransferase with three claimed histone marks and one robust non-histone substrate, sitting on an HOXA9 feed-forward loop in MLL-AF9 AML and a fatty-acid-oxidation axis in CML stem cells, with zero interventional trials
SMYD3 methylates MAP3K2 at K260, blocking PP2A binding and potentiating Ras/Raf/MEK/ERK; SMYD3 catalytic activity is required for Ras-driven pancreatic and lung tumour development in mice and synergises with a MEK inhibitor