Niche proteins enumerated and NOT written in this tranche, and four symbol traps measured while doing it - three of them new to the registry

[
  "gene-cxcl12",
  "gene-cxcr4",
  "gene-itga4",
  "gene-vcam1",
  "gene-kitlg",
  "gene-lepr",
  "gene-spp1",
  "gene-sele",
  "drug-plerixafor",
  "hyp-cxcr4-function-is-a-hoxa-meis1-output-of-the-fusion",
  "ana-cxcr4-under-menin-inhibition-geo-reanalysis",
  "absence-cxcr4-axis-never-measured-under-menin-inhibition",
  "ana-is-niche-biology-measured-in-kmt2a-rearranged-disease-a-census",
  "ana-q33-niche-protein-layer-rejected-candidates"
]

[
  {
    "candidate": "ACKR3 / CXCR7",
    "terms": "ACKR3[tiab] OR CXCR7[tiab]",
    "intersection": 0,
    "disposition": "rejected as an entry, flagged as a resistance liability",
    "reason": "The second CXCL12 receptor, not antagonised by plerixafor. Zero papers in this disease and zero mentions of either token anywhere in this corpus, so there is nothing to build an entry on - but that is exactly what makes it the unexamined bypass for any CXCR4-directed programme. Recorded in gene-cxcr4 and drug-plerixafor instead. Name list did not include the previous symbols RDC1 or CMKOR1, so the zero may be too tight (rule 69)."
  },
  {
    "candidate": "ITGB1",
    "terms": "ITGB1[tiab]",
    "intersection": 0,
    "disposition": "rejected, zero is informative",
    "reason": "The beta chain of VLA-4 has no KMT2A literature while its alpha chain has one paper, although every VLA-4 result is jointly about both. Recorded in gene-itga4 as the reason a VLA-4 result must not be attributed to ITGA4 alone."
  },
  {
    "candidate": "SELL / SELP / P-selectin / L-selectin",
    "terms": "SELL[tiab] OR SELP[tiab] OR P-selectin[tiab] OR L-selectin[tiab]",
    "intersection": 0,
    "disposition": "rejected outright",
    "reason": "The whole selectin family has zero KMT2A-fusion literature. E-selectin was written despite also being zero, because of the strength of the AML vascular-niche mechanism and the existence of a clinical antagonist; P- and L-selectin have neither."
  },
  {
    "candidate": "JAG1 / jagged",
    "terms": "JAG1[tiab] OR jagged[tiab]",
    "intersection": 2,
    "disposition": "rejected for this tranche, referred to a NOTCH beat",
    "reason": "NOTCH ligand biology in KMT2A-rearranged disease is a topic in its own right that this graph already touches through mech-kmt2a-maml2-notch-dominant-negative. Writing JAG1 as a niche ligand would mis-frame it."
  },
  {
    "candidate": "ANGPT1 / TEK",
    "terms": "inherited from ana-marrow-niche-rejected-candidates (cycle 40): 3 records, none niche-mechanistic",
    "intersection": null,
    "disposition": "rejection INHERITED and labelled as inherited",
    "reason": "Not re-queried in this beat. The cycle-40 ledger's reasoning included that ANGPTL2 is not an angiopoietin despite the name, so the ANGPT1/TIE2 axis proper has no KMT2A-specific literature. Recorded as inherited rather than re-presented as a new measurement."
  },
  {
    "candidate": "nestin / NES",
    "terms": "nestin[tiab]",
    "intersection": 6,
    "disposition": "DEFERRED, not rejected",
    "reason": "Six records exist and cell-type-nestin-msc already carries the biology. A gene-nes entry is defensible; it was not written because the cell-type entry covers the evidence and because the corpus token NES means NUCLEAR EXPORT SIGNAL here, not nestin. If written it must be titled for the intermediate filament, must search as 'nestin', and must state the collision in its first line. Note also the unresolved tension between Ding 2012, where Scf deletion from nestin-cre cells did NOT affect HSCs, and cell-type-nestin-msc's leukaemia-supporting result."
  },
  {
    "candidate": "FUT4 / FUT7 / ST3GAL4 / ST3GAL6",
    "terms": "not queried",
    "intersection": null,
    "disposition": "named as an open question in gene-sele, not written",
    "reason": "The glycosyltransferases that build E-selectin ligands. Whether any is under KMT2A-fusion control is the question that would convert a general AML niche mechanism into a fusion-specific one, and it is recorded as gene-sele's second open question rather than as four thin entries."
  },
  {
    "candidate": "S100A8",
    "terms": "appeared in the SPP1 intersection (PMID 40920990, 'Critical role of niche S100A8 for acute myeloid leukemia progression and hematopoiesis regeneration')",
    "intersection": null,
    "disposition": "not pursued",
    "reason": "A niche factor in AML progression that surfaced incidentally. Not read beyond the title and not pursued for lack of a KMT2A-specific claim in the title; recorded so a future cycle knows it was seen."
  }
]

[
  {
    "token": "SELE",
    "class": "short_symbol_is_a_prefix_of_a_common_word",
    "corpus_substring_files": 89,
    "corpus_wordboundary_files": 1,
    "matches": {
      "SELECTIVE": 52,
      "SELECTIVITY": 36,
      "SELEX": 15,
      "SELECTED": 12,
      "SELECTIVELY": 7,
      "SELECTION": 3,
      "SELECTS": 1,
      "SELECT": 1,
      "SELE": 5
    },
    "rule": "grep with -w, or search as E-selectin",
    "in_registry_at_read_time": false,
    "family": "same as the existing WAS, RPE and CPM rows"
  },
  {
    "token": "SCF",
    "class": "acronym_means_a_complex_not_this_gene",
    "this_graph_means": "SKP1-CUL1-F-box ubiquitin ligase (SCF-Skp2 degrading MLL; SCF-betaTrCP substrates)",
    "uniprot_means": "UniProt entry name SCF_HUMAN and curated synonym for KITLG (P21583, stem cell factor)",
    "corpus_wordboundary_files": 41,
    "corpus_occurrences": 153,
    "rule": "search KITLG or 'KIT ligand'; never a bare SCF",
    "in_registry_at_read_time": false,
    "note": "2 of the 153 are the string SCFR, which is KIT's own alias - a third sense of the same three letters"
  },
  {
    "token": "LEPR",
    "class": "gene_exact_returns_several_reviewed_records_wrong_one_first",
    "wrong_first": {
      "accession": "O15243",
      "entry": "OBRG_HUMAN",
      "primary_gene": "LEPROT",
      "length_aa": 131
    },
    "correct": {
      "accession": "P48357",
      "entry": "LEPR_HUMAN",
      "primary_gene": "LEPR",
      "length_aa": 1165
    },
    "discriminator": "length 131 vs 1165; but read the PRIMARY GENE NAME back, which always works",
    "rule": "rule 62 shape - would pass verify-accessions.py because a synonym match is not a typo",
    "in_registry_at_read_time": false,
    "sibling": "gene_exact:SPP1 also returns two records (P10451 osteopontin 314 aa, Q9BX95 SGPP1 441 aa) but the correct one is first today"
  },
  {
    "token": "VCAM-1 vs VCAM1",
    "class": "hyphenation_is_a_different_string",
    "measured": "VCAM1[tiab] AND (disease half) = 2; VCAM-1[tiab] AND the identical half = 0, valid translation, 'No items found.'",
    "rule": "adhesion papers overwhelmingly write VCAM-1; search both spellings",
    "in_registry_at_read_time": false
  }
]

{
  "class": "empty_querytranslation_is_not_a_zero",
  "occurrences_this_session": 4,
  "approx_queries_run": 40,
  "affected_queries": [
    "CXCR4[tiab] AND (disease half) - true value 11",
    "CXCR4[tiab] AND infant[tiab] AND leukemia[tiab] - true value 2",
    "(SPP1[tiab] OR osteopontin[tiab]) AND (disease half) - true value 4",
    "(SELE[tiab] OR E-selectin[tiab] OR uproleselan[tiab]) AND (disease half) - true value 0"
  ],
  "signature_of_a_genuine_zero": "non-empty querytranslation PLUS warninglist.outputmessages containing 'No items found.'",
  "signature_of_a_failure": "empty or absent querytranslation and no count field",
  "why_it_matters": "Two of the four were CXCR4 queries, the central molecule of this beat, and recording either as written would have published a false zero aligned with the hypothesis being tested."
}