The second ring — smaller KMT2A-fusion mechanism labs, and where the literature actually points

Lineage-switch relapses are rare, so the cohort is necessarily small and the recurrence claims rest on few events. Whether the chromatin-modifier changes cause the switch or accompany it is not settled by the study design.

• Shared fusion breakpoints establish common ancestry, not the direction of transition: a myeloid relapse sharing the breakpoint is equally consistent with switching of the lymphoid clone and with outgrowth of a myeloid sibling present from the start. Journal of Clinical Oncology 2025

n=36 for the discovery sequencing; subgroup percentages rest on small denominators. Relapse cohorts are selected by definition and cannot describe patients who were cured.

• Thirty-six relapse genomes and 257 longitudinal samples from 30 patients is the right design and a small denominator for subgroup claims. TP53 and IKZF1 alterations in 64% of drug-response gene changes is a proportion OF THE ALTERATIONS FOUND, which is not the same as a patient frequency, and relapse cohorts are assembled from patients who relapsed. Statistics in Medicine 2006

Single-group finding; 'normal hematopoietic cells remain unaffected' is a strong claim resting on in-vitro and murine readouts rather than on long-term in-vivo tolerance. Immunoproteasome inhibition has immunological consequences that a leukaemia-focused study will not capture.

• Two corrections to this item, both from one author list. (1) Affiliation. This entry and its section text place Florian Heidel at 'University Medicine Greifswald', taken from the 2023 PSMB8 paper. A 2025 Journal of Clinical Oncology paper lists 'Heidel, Florian H - Hannover Medical School, Hannover, Germany'. The Greifswald attribution is therefore stale and the index should carry Hannover, with the Greifswald affiliation retained as historical. (2) Competitive position, which matters more. The entry's premise for the whole second ring is that these groups are 'small enough that a reciprocal collaboration is realistic' with low competitive overlap. Heidel is a listed author on the registration-enabling phase 2 of KOMET-001, the ziftomenib trial in relapsed/refractory NPM1-mutant AML - that is, he is a clinical menin-inhibitor trialist as well as the source of the PSMB8 dependency claim. A group that sits in both the non-menin dependency literature and the menin-inhibitor trial network is not competitively neutral on either question, and a lab approaching him about PSMB8 as an alternative to menin should know that before it writes. None of this is a criticism; both facts are printed on the papers. Caveat: an author line establishes participation in a trial, not its terms; and this JCO paper does not carry a per-author disclosure in the record read here, so any financial relationship is not asserted. Read: title, abstract and author affiliations of the PubMed record; not full text. Wang ES et al. 2025

MLL-ENL (KMT2A::MLLT1) is not KMT2A::AFF1 and the metabolic phenotype may be partner-specific. The indexed affiliation for this paper reads only 'Department of Genetics', so the institution is NOT verified here.

• This is a different fusion in a different lineage. MLL-ENL and MLL-AF9 in murine or human myeloid systems are the models this dependency was found in, and the index disease here is human infant KMT2A::AFF1 B-lymphoid leukaemia. A dependency that has not been lineage-split in this project's own DepMap framework should be treated as a candidate. Oncofusion knowledge graph (internal analysis) 2026

Conditional knockout mouse models of MLL-AF9 AML; stage-dependent opposite effects are exactly the kind of result that does not transfer cleanly to a drug given continuously to a patient.

• The index is incomplete in a way that matters for the mechanism questions this graph cares about, and the omission is a whole mechanism class rather than a single laboratory. A group spanning Hunan University, the University of Virginia and Zhejiang University (Wei Li, Bi Shi, Hao Jiang) published in Nature Communications in January 2026 that the RNA-binding protein AKAP95 is required for MLL-rearranged leukemogenesis; that AKAP95 phase separation and RNA binding modulate Pol II recruitment into condensates genome-wide; that AKAP95 interacts with the MLL1 translocated fragment, with partial co-condensation strengthening AKAP95 binding at MLL-AF9 target genes; that AKAP95 loss downregulates MLL-AF9 targets and impairs MLL-AF9-driven leukemogenesis while being dispensable for normal and regenerative haematopoiesis; and that a designed peptide bridging AKAP95 to HSP70 disrupts its phase separation and inhibits MLLr leukaemia proliferation. This is a second-ring laboratory by every criterion this entry uses, it owns a narrow well-defined piece of KMT2A-fusion biology, and it is the only group identified anywhere in this beat working on fusion condensates - the mechanism class the strategy entries describe as having no owner. Its absence is a reason to treat this index as a snapshot of what one survey pass found rather than as a map of the second ring, and to re-run the survey against recent literature before any strategic conclusion is drawn from a group's absence. Caveats: MLL-AF9 and murine models, not KMT2A::AFF1 or infant B-ALL; a single 2026 paper, unreplicated; the peptide is a tool compound. Read: abstract and affiliations only. Yan X et al. 2026

Murine myelomonocytic transformation models; the abstract as indexed has gene symbols stripped, so the precise C/EBP family members involved should be read from the paper before being asserted. C/EBPalpha is itself essential for normal myelopoiesis, so the therapeutic window is questionable.

• This is a different fusion in a different lineage. MLL-ENL and MLL-AF9 in murine or human myeloid systems are the models this dependency was found in, and the index disease here is human infant KMT2A::AFF1 B-lymphoid leukaemia. A dependency that has not been lineage-split in this project's own DepMap framework should be treated as a candidate. The escapees are the informative finding: a requirement that rare cells evade by upregulating one gene is dose-dependent, which is the opposite of what a clean target looks like. Oncofusion knowledge graph (internal analysis) 2026

Very recent (2026) and unreplicated. Relevant to this project as a comparator fusion family in the same patient population rather than as KMT2A biology.

• NUTM1-rearranged leukaemia is a DIFFERENT entity, defined by the absence of KMT2A rearrangement. It is a useful comparator for what infant leukaemia looks like without the beat's lesion and is not evidence about KMT2A biology; global hypomethylation irrespective of partner is also the kind of genome-wide observation most sensitive to blast purity. Oncofusion knowledge graph (internal analysis) 2026

An RNA-based fusion-detection method identifies the transcript junction, not the genomic DNA breakpoint, so it does not substitute for long-distance inverse PCR for breakpoint-level questions or for patient-specific DNA MRD assay design. A 2017 method; the sequencing landscape has moved.

• A low-input anchored-multiplex assay is a methods contribution whose performance figures come from its own validation, usually on samples chosen to have known fusions. Sensitivity for the rare partners that motivate such an assay is the number that matters and is the hardest to establish, since the cases are scarce by definition. Blood Cells, Molecules and Diseases 2010