The immunophenotype beat's own ledger: what was enumerated, what was written, what was dropped and why, and what could not be read
also called Q45 enumeration · marker bounded rejection register · immunophenotype beat ledger
This is the working record of the beat that wrote the surface-marker entries: which markers were considered, which got an entry, which were deliberately left out and the reason, and which papers could be found but not read. It exists so that someone repeating this work next year repeats the decisions and not just the searches.
The same thing, for a scientist
The standing brief requires **bounded rejection** as well as bounded absence: when a marker, a paper or a laboratory would reasonably be expected in a beat and is not there, the reason must leave a trace. This is that trace for the Q45 immunophenotype beat. **Markers enumerated: 19. Written: 10.** The nine dropped, with reasons, are in `data.dropped`. Two of them are judgement calls a future cycle may reverse: - **CD371 / CLEC12A** is the strongest omission. It is **the most sensitive marker of transient myelomonocytic switch** in the largest relevant paediatric cohort (1,812 BCP-ALL) [E1], and it is named repeatedly in three of this beat's entries. It was dropped because the CD371 literature is a **DUX4-fusion and general BCP-ALL** story rather than a KMT2A one, and the beat's remit was the KMT2A-rearranged markers. **If a future cycle writes one more marker entry, it should be CLEC12A/CD371, which has no entry yet.** - **CD15 and CD65 (FUT4 / fucosyltransferase antigens)** carry more of the aberrant-myeloid diagnostic signal in this disease than CD33 does - Schwartz's 100%-accuracy subgroup is CD65s+ and/or CD15+ [E2], and Behm's NG2-positive cases were CD15/CD65-positive [E3]. They were dropped because the CD15 and CD65s antigens are **carbohydrate epitopes** without a clean single-gene mapping, so a `gene-` entry would misrepresent them; the right entry is a `concept-` on carbohydrate myeloid antigens and it is not written. **Counted absences recorded in this beat, with their exact queries**, are in `data.counted_absences`. The load-bearing one is NG2/CSPG4 against menin inhibition: **zero records across three query forms with querytranslation inspected**, and the repository grepped. **Papers found, on topic, and unreadable after all five routes: two**, both recorded in `data.unreadable` with the routes tried. One of them, PMID 32876512, asserts in its title exactly the finding this beat was sent to look for. **One methodological finding worth carrying forward.** `gene_exact:TdT AND organism_id:9606 AND reviewed:true` returns **P04053, whose primary gene name is DNTT** - a synonym match that resolves correctly here but is the rule-62 shape, and TdT appears in this beat's marker lists throughout. **CD antigens are an alias minefield in a specific way this project had not logged:** the gene symbol and the CD number are different vocabularies for eight of the nineteen markers enumerated, and in the worst case (`MME` / `NEP_HUMAN` / neprilysin / CD10 / CALLA) there are **four non-overlapping names for one molecule**.
Bounded rejections: what a reader would expect to find cited here and will not
Laboratories that publish repeatedly in this area and are cited thinly or not at all:
- The iBFM Flow Network / AIEOP-BFM group (Dworzak, Buldini, and colleagues). Cited twice in this beat, and they are the dominant group in paediatric ALL flow MRD standardisation. Their MRD methodology papers were judged out of scope because this beat is about diagnostic markers rather than MRD gating, and this graph already has an MRD beat. A future cycle joining the two beats should start here.
- The Menendez / Bueno group (Barcelona). Cited four times and they produced the NG2 CNS work, the NG2/MLL-AF4/NR3C1 mechanism, the CD34+CD19-CD22+ escape reservoir and the CD133 CAR myeloablation result. This is a concentration risk: four of this beat's load-bearing findings come from one laboratory, and the beat could not find independent replication of any of them.
- Mesothelin and JL1 in acute leukaemia (PMIDs 41629741, 30809981, found by title). Both are surface antigens with reported KMT2A or paediatric-ALL relevance. Dropped as too far from the established diagnostic panel to earn an entry in this beat; recorded so a future antigen-discovery cycle can find them.
A paper whose title looked squarely on-beat and was not: PMID 42720167, "Cuplike nuclei and bimodal CD34 expression in adult KMT2A-rearranged pro-B acute lymphoblastic leukemia" - an images item with no abstract and no PMCID, adult, single case or small series. It is cited in gene-cd34 at weak confidence for its title assertion only, and it should not be read as evidence about infants.
| script | manual enumeration; UniProt REST and PubMed E-utilities queries run 2026-09-20 and recorded in data.counted_absences and evidence E4 |
|---|---|
| status | run |
| result summary | 19 markers enumerated and resolved against UniProt; 10 written as gene entries; 9 dropped with recorded reasons; 3 counted absences recorded with exact queries; 2 on-topic papers unreadable after all five routes |
| alias findings | 8 of 19 markers have a gene symbol disjoint from the clinical CD name: CSPG4/NG2, IL3RA/CD123, MME/CD10, PROM1/CD133, KIT/CD117, ANPEP/CD13, MS4A1/CD20, NCAM1/CD56. Worst case MME: gene MME, UniProt entry NEP_HUMAN, protein neprilysin, clinical CD10 and CALLA - four non-overlapping vocabularies. gene_exact:TdT resolves to DNTT by synonym match (rule 62 shape). |
inputs
[ "scripts/check-named-but-unwritten.py queue (coordinator, 2026-09-20)", "UniProt REST gene_exact over 19 symbols", "PubMed E-utilities esearch/efetch" ]
written
[ "gene-cd34", "gene-cspg4", "gene-cd33", "gene-il3ra", "gene-cd22", "gene-mme", "gene-prom1", "gene-cd38", "gene-kit", "gene-cd7" ]
dropped — the numbers behind the chart
[
{
"marker": "CD371 / CLEC12A",
"uniprot": "not resolved in this beat",
"reason": "the literature is DUX4-fusion and general BCP-ALL, not KMT2A; the beat's remit was KMT2A markers",
"what_would_change_my_mind": "any measurement of CD371 in KMT2A-rearranged disease, or its use as an early switch detector in this genotype. THIS IS THE STRONGEST OMISSION."
},
{
"marker": "CD15 and CD65s",
"reason": "carbohydrate epitopes without a clean single-gene mapping; a gene- entry would misrepresent them",
"what_would_change_my_mind": "a concept- entry on carbohydrate myeloid antigens; they carry MORE diagnostic weight in this disease than CD33 does"
},
{
"marker": "ANPEP / CD13",
"uniprot": "P15144",
"length_aa": 967,
"reason": "myeloid specificity 0.36, the weakest marker measured in this beat, and absent in MLL-AF4 B-ALL in a 283-case series (PMID 17434163)",
"what_would_change_my_mind": "a use for CD13 as a negative discriminator in this genotype, which the 17434163 finding actually suggests"
},
{
"marker": "MS4A1 / CD20",
"uniprot": "P11836",
"length_aa": 297,
"reason": "a mature B marker, generally negative in pro-B KMT2A-r disease; appears in the t(11q23) flow score but contributes negatively",
"what_would_change_my_mind": "rituximab or CD20 CAR data in KMT2A-rearranged infants"
},
{
"marker": "CD4",
"uniprot": "P01730",
"length_aa": 458,
"reason": "named in 20 entries but as a T-cell subset marker and a monocytic marker, not as a KMT2A blast antigen; would be a T-cell-biology entry",
"what_would_change_my_mind": "CD4-dim as a monocytic lineage-switch marker, which is how it would earn a place here"
},
{
"marker": "CD79A",
"uniprot": "P11912",
"length_aa": 226,
"reason": "a lineage-assignment marker in EGIL/WHO criteria rather than a KMT2A-specific one; better placed in an MPAL entry",
"what_would_change_my_mind": "the CD117/cytoplasmic-CD79a/CD7 triple being followed up in KMT2A-rearranged ambiguous-lineage disease"
},
{
"marker": "DNTT / TdT",
"uniprot": "P04053",
"length_aa": 509,
"reason": "lower MFI in KMT2A-r than KMT2A-germline B-ALL (PMID 38468826) but no mechanism and no therapeutic angle found",
"what_would_change_my_mind": "a link between TdT level and the VDJ repertoire question in this graph's MRD beat, which is plausible and unexamined"
},
{
"marker": "MPO",
"uniprot": "P05164",
"length_aa": 745,
"reason": "the WHO/ICC myeloid-lineage-defining marker, so it belongs in a lineage-assignment concept entry rather than a marker entry",
"what_would_change_my_mind": "MPO kinetics at lineage switch, which is exactly what exp-marker-order-at-lineage-switch-paired-flow-reanalysis should measure and which this beat failed to put in that panel"
},
{
"marker": "NCAM1 / CD56",
"uniprot": "P13591",
"length_aa": 858,
"reason": "co-expressed with 7.1 in 33/38 NG2-positive AML (PMID 10914547) and an internal control in the CD7 study, but its own literature is T-ALL and NK biology",
"what_would_change_my_mind": "CD56 as part of an NG2-positive myeloid signature in KMT2A-r AML, for which PMID 10914547 is already suggestive"
},
{
"marker": "Mesothelin; JL1",
"pmids": [
"41629741",
"30809981"
],
"reason": "found by title; too far from the established diagnostic panel for this beat",
"what_would_change_my_mind": "an antigen-discovery cycle, where both belong"
}
]counted absences
[
{
"claim": "no published intersection of NG2/CSPG4 with menin inhibition",
"queries": [
"NG2[tiab] AND (menin[tiab] OR revumenib[tiab] OR SNDX-5613[tiab] OR VTP-50469[tiab]) -> 0",
"CSPG4[tiab] AND (menin[tiab] OR \"menin inhibitor\"[tiab]) -> 0",
"(CSPG4 OR NG2) AND leukemia AND (menin OR revumenib OR \"menin inhibitor\") -> 0"
],
"querytranslation_checked": true,
"repo_grepped": true,
"not_run": [
"7.1 (unusable token)",
"MCSP",
"ziftomenib",
"bleximenib",
"enzomenib",
"DS-1594",
"JNJ-75276617"
],
"date": "2026-09-20"
},
{
"claim": "no published marker-by-marker ORDER for lineage switch in KMT2A-rearranged disease",
"basis": "Project EVOLVE describes 10 lineage-drift cases without naming the changed markers; the 2026 Leukemia review on detection issues definitions and recommendations without an order; PMID 32876512 asserts a harbinger phenotype in its title and is unreadable",
"strength": "an inference from three abstracts, two of whose full texts were not read - the weakest absence in this beat",
"date": "2026-09-20"
},
{
"claim": "no CD22 site-density measurement in infant KMT2A-rearranged blasts",
"basis": "AALL1621 (ages 1-21) and ITCC-059 cohort 3 (ages >1 to <18) both exclude infants while ITCC-059 names KMT2A::AFF1 as a qualifying high-risk lesion",
"date": "2026-09-20"
}
]unreadable
[
{
"pmid": "32876512",
"doi": "10.1080/10428194.2020.1815018",
"title": "Aberrant leukemia-associated immunophenotype as potential harbinger of lineage switch in KMT2A-rearranged leukemia: a case series.",
"routes": "R1 EPMC pmcid null isOA N abstract 0 chars; efetch cross-check no pmc ArticleId so R2-R4 unavailable; R5 DOI 403 (Taylor & Francis), 25 chars stripped",
"date": "2026-09-20"
},
{
"pmid": "41925057",
"doi": "10.1002/1545-5017.70289",
"title": "Successful Management of Lineage Switch Infant Leukemia With Revumenib and Hematopoietic Cell Transplantation.",
"routes": "R1 EPMC pmcid null isOA N abstract 0 chars; efetch cross-check no pmc ArticleId so R2-R4 unavailable; R5 DOI 403 (Wiley), 25 chars stripped",
"date": "2026-09-20"
},
{
"pmid": "42720167",
"doi": "10.1182/blood.2026034862",
"title": "Cuplike nuclei and bimodal CD34 expression in adult KMT2A-rearranged pro-B acute lymphoblastic leukemia.",
"routes": "no abstract in PubMed; EPMC no pmcid; efetch no pmc ArticleId; DOI route not attempted (ashpublications, not open access)",
"date": "2026-09-20"
}
]- Four of this beat's load-bearing findings - NG2 in CNS infiltration, the NG2/MLL-AF4/NR3C1 mechanism, the CD34+CD19-CD22+ escape reservoir and the CD133 CAR myeloablation - come from a single laboratory group, and no independent replication of any of them was found. Any strategy built on more than one of them is more correlated than it looks.
- The marker selection for this beat came from a symbol-frequency queue, and the ledger shows that produced a defensible but skewed set: CD33 got an entry on weak diagnostic evidence because it is a gene symbol, while CD15 and CD65s carry more diagnostic weight and got none because they are carbohydrate epitopes. A frequency queue over HGNC symbols cannot see antigens that are not genes.
- Recording a bounded rejection makes the beat auditable and also makes it look more complete than it is. Nine markers were dropped and the reasons are honest, but 'out of scope' did most of the work, and scope was set by the dispatch brief rather than by the biology.
- Should an entry for CLEC12A/CD371 be written (it does not exist yet; the id would be gene-clec12a)? It is the single strongest omission from this beat and it is the marker most likely to answer the early-switch-detection question.
- Is there a KMT2A-specific CD371 measurement anywhere? The published work is DUX4-fusion and general BCP-ALL, and the KMT2A fraction is unreported.
- Should a `concept-` entry cover the carbohydrate myeloid antigens CD15 and CD65s, which no gene entry can properly represent?
- Can any of the four Menendez-group findings be independently replicated from public data? A concentration of that size in one laboratory is a structural risk to several of this graph's conclusions.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Among 1,812 children with newly diagnosed BCP-ALL in AIEOP-BFM ALL 2009, 160 (8.8%) were CD371-positive; these cases showed a transient myelomonocytic switch in up to 65.4% of samples at day 15 and inferior early response, with 5-year event-free survival of 88.3%. Among 420 validation-cohort patients, 27 of 28 (96.4%) DUX4-fusion-positive cases were CD371-positive. CD371 was the most sensitive marker of transient myelomonocytic switch.
CD371-positive pediatric B-cell acute lymphoblastic leukemia: propensity to lineage switch and slow early response to treatment. · Buldini B, ..., Dworzak MN · Blood 2024What this evidence cannot show
Abstract read in full from raw efetch XML on 2026-09-20; full text not read. Cited here to justify the bounded rejection of CD371 rather than to make a claim about KMT2A-rearranged disease. **The abstract does not report the KMT2A-rearranged fraction of the CD371-positive group**, which is precisely why the marker was dropped: its established relevance is to DUX4-fusion and general BCP-ALL, and its relevance to this graph's index disease is unmeasured.
Evidence that argues against this (1)
- The reason for dropping CD371 is also the reason it may be the most important omission: a marker whose KMT2A relevance is unmeasured is a gap, not an absence, and 'out of scope' and 'nobody has looked' are different justifications that this ledger deliberately does not conflate. Silbert SK et al. 2025
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In 313 adult ALL patients, NG2 expression was significantly associated with a CD10-/CD34-/CD24-/CD65s+/CD15+/CD13-/CD33- phenotype, and NG2 predicted MLL rearrangement with 100% accuracy among patients disclosing a CD65s-positive and/or CD15-positive immunophenotype, with overall sensitivity 0.89, specificity 0.89 and positive predictive value 0.93.
Expression of the human homologue of rat NG2 in adult acute lymphoblastic leukemia: close association with MLL rearrangement and a CD10(-)/CD24(-)/CD65s(+)/CD15(+) B-cell phenotype. · Schwartz S, ..., Thiel E · Leukemia 2003What this evidence cannot show
Abstract read in full from raw efetch XML on 2026-09-20; full text not read. Adult ALL. Cited here to justify the bounded rejection of CD15 and CD65s: **this is the single strongest quantitative statement in the beat about those two antigens, and it shows they carry more diagnostic weight than CD33, which DID get an entry.** The asymmetry is a defect of the beat's marker selection, made from the coordinator's symbol-based queue, and is recorded as such.
Evidence that argues against this (1)
- CD15 and CD65s are carbohydrate epitopes whose synthesis involves several fucosyltransferases, so writing them as gene entries would falsely imply a one-gene mapping. The correct form is a concept entry on carbohydrate myeloid antigens, and its absence is the real gap rather than the absence of a gene entry. Behm FG et al. 1996
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Among 104 consecutive children with ALL screened prospectively with monoclonal antibody 7.1, NG2-positive cases were significantly more likely to be infants, to have hyperleukocytosis and CNS involvement, to be CD10-negative, and to express the myeloid-associated antigens CD15 and CD65.
Human homologue of the rat chondroitin sulfate proteoglycan, NG2, detected by monoclonal antibody 7.1, identifies childhood acute lymphoblastic leukemias with t(4;11)(q21;q23) or t(11;19)(q23;p13) and MLL gene rearrangements. · Behm FG, ..., Bernstein ID · Blood 1996What this evidence cannot show
Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **Nine NG2-positive cases.** Cited here only to corroborate that CD15 and CD65 travel with the KMT2A-rearranged phenotype in children as well as adults, which is the paediatric half of the bounded-rejection argument. PubMed returns no DOI for this record.
Evidence that argues against this (1)
- A 1996 series of nine positive cases is thin support for a marker-selection decision, and the decision to drop CD15 and CD65 from this beat was made on the vocabulary problem rather than on the strength of their evidence. The evidence for them is, if anything, better than for CD7, which did get an entry on six patients. Bommannan K et al. 2022
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Nineteen human surface-marker genes were resolved against UniProt on 2026-09-20 with `gene_exact:<SYMBOL> AND organism_id:9606 AND reviewed:true`, requesting accession, entry name, gene_primary, gene_synonym, protein_name and length. Every query returned exactly one reviewed human record whose primary gene name matched the symbol requested, with the single exception that `gene_exact:TdT` returned P04053 whose primary gene name is DNTT, a synonym match. Resolved: CD34 P28906 385 aa; CSPG4 Q6UVK1 2322 aa; CD33 P20138 364 aa; IL3RA P26951 378 aa; CD22 P20273 847 aa; MME P08473 750 aa; PROM1 O43490 865 aa; CD38 P28907 300 aa; KIT P10721 976 aa; CD7 P09564 240 aa; ANPEP P15144 967 aa; MS4A1 P11836 297 aa; CD4 P01730 458 aa; CD79A P11912 226 aa; DNTT P04053 509 aa; CD3E P07766 207 aa; MPO P05164 745 aa; NCAM1 P13591 858 aa.
UniProtKB REST search, gene_exact queries over 19 human surface-marker symbols with gene_primary and length requested, run 2026-09-20 · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026What this evidence cannot show
All 19 queries were run in one session on 2026-09-20 and every row was read, per rules 62, 63 and 65. **Single-row returns mean the multi-row hazard did not arise for these symbols; it does not mean it cannot.** The CD-number forms (CD117, CD123, CD10, CD133, CD13, CD20) were **not** used as UniProt queries because they are not gene symbols - which is itself the finding: for 8 of 19 markers here the clinical name and the gene symbol are disjoint vocabularies. ANPEP, MS4A1, CD4, CD79A, DNTT, CD3E, MPO and NCAM1 were resolved but not written up; their accessions are recorded so a future cycle need not re-fetch.
Evidence that argues against this (1)
- That all 19 resolved cleanly is a weak reassurance: the paralogue-collision cases this project has logged (CSNK2A1/CSNK2A2, HDAC7/HDAC9, TRAF2/TANK, TCF3/TCF7L1) all involve gene families where a synonym is another member's primary symbol. CD22/SIGLEC2 and CD33/SIGLEC3 are exactly that shape and happened not to collide; a future beat touching other SIGLECs should not assume the same. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
the beat ledger: 19 markers enumerated and resolved against UniProt, 10 written, 9 dropped with reasons, 3 counted absences, 3 unreadable papers
records that CD33 got an entry on weak diagnostic evidence while CD15 and CD65s, which carry more weight, got none, because an HGNC-symbol queue cannot see antigens that are not genes
- related to →Which marker moves first? A re-analysis of flow files that already exist, to put an order on lineage switch
the ledger records that MPO was dropped and should have been in that experiment's panel
adds three identifiers with the routes tried and the stripped lengths measured