UROS: searched and empty — the scan's rank-15 hit has no leukemia literature, and no entry was written
also called UROS negative result · the false positive · heme biosynthesis scan hit
A scan run by this project said that leukemia cells with the KMT2A fusion depend on UROS, an enzyme that makes the red pigment in blood. A search of the medical literature found nothing at all connecting this enzyme to leukemia, so no entry was written about it — and that empty result is recorded here on purpose.
The same thing, for a scientist
UROS (uroporphyrinogen III synthase, UniProt P10746, 265 aa) ranked 15 of 4,589 in the coordinator's unbiased DepMap KMT2A-selectivity scan and had no entry in this graph. Targeted PubMed searches for UROS with leukemia, with KMT2A/MLL, with dependency and with myeloid differentiation returned **no relevant primary literature**: the only hits were a 1985 paper and a congenital erythropoietic porphyria (CEP) treatment study. On the brief's rule — write an entry only where the literature supports one — **no gene entry for UROS was created**, and this node records the negative so that the next person does not repeat the search. Two nearby findings were located and are recorded because they bound what could be true: heme biosynthesis as a whole is an apoptotic modifier in AML and heme-pathway genes are upregulated at diagnosis in pediatric AML non-responders [E2], and SETD1A controls transcriptional pause release of heme biosynthesis genes in leukemia [E3]. Neither is about UROS and neither rescues the hit. The most economical explanation of the scan signal is a **toxic-intermediate artefact**: UROS loss causes accumulation of non-physiological porphyrin isomer I, which is the disease mechanism of CEP [E4], so knocking out UROS harms any cell with upstream haem flux regardless of its fusion status.
What was searched, and what came back
PubMed, via the MCP tools, on 2026-09-18:
| Query | Total hits | Relevant |
|---|---|---|
| UROS uroporphyrinogen III synthase leukemia | 2 | none (one 1985 paper; one CEP therapy study) |
| uroporphyrinogen III synthase UROS cancer dependency | 2 | none (a CEP study; a CRISPR-truncation methods paper) |
| uroporphyrinogen III synthase myeloid differentiation heme leukemia cells | 0 | — |
| UROS deficiency uroporphyrin I accumulation phototoxicity cells | 2 | background on CEP only |
There is no paper connecting UROS to KMT2A-rearranged leukemia, to MLL fusions, to infant leukemia, or to leukemia dependency of any kind. This is a clean negative, not an inconclusive one 1.
The nearest real findings, and why they do not rescue the hit
Two things in the neighbourhood are real and worth recording, because someone will find them and be tempted.
1. Heme biosynthesis is an apoptotic modifier in AML. A CRISPR knockout screen for metabolic genes altering commitment to apoptosis in BCL-2-dependent AML lines singled out heme biosynthesis: reducing heme biosynthesis potentiates apoptosis by lowering electron-transport-chain activity, depolarising the mitochondrial membrane and raising the propensity to die. In 240 pediatric patients across three Children's Oncology Group AML trials, patients who would later fail to achieve remission after two inductions expressed higher heme biosynthetic genes at diagnosis, and heme biosynthesis had the highest proportion of significantly upregulated genes of any metabolic pathway in non-responders 2.
That points the opposite way from the scan. It says losing heme biosynthesis makes leukemia cells easier to kill — a sensitiser, not a dependency — and it is about the pathway, not about UROS, which is never named in the paper's text (zero occurrences of 'uroporphyrin' in the full text read here).
2. SETD1A, a KMT2-family methyltransferase, controls heme biosynthesis gene transcription in leukemia. Rapid degradation of SETD1A in AML lines immediately downregulates DNA repair and heme biosynthesis transcripts, through a non-enzymatic, domain-dependent cyclin K function that governs RNA polymerase II pause release 3.
This is the most seductive of the near-misses because SETD1A is already in this graph and is a KMT2 family member. But SETD1A is not KMT2A, the paper is not about KMT2A-rearranged leukemia, and nothing in it says UROS specifically. Reading a KMT2A-fusion connection into it would be exactly the kind of chain of plausible steps this project has been burned by before.
The best explanation of the signal: a toxic intermediate, not a dependency
UROS is the fourth enzyme of haem biosynthesis. When it is absent, the upstream substrate hydroxymethylbilane cyclises non-enzymatically to the isomer I series instead of the physiological isomer III series, and those non-physiological porphyrins accumulate. That accumulation is congenital erythropoietic porphyria: skin photosensitivity and chronic haemolytic anaemia whose severity tracks porphyrin overload, and which can be relieved by reducing flux through the pathway upstream at ALAS2 4.
So a UROS knockout is not primarily a loss-of-heme experiment. It is a gain-of-toxic-metabolite experiment, and how much it hurts a cell depends on how much flux that cell is pushing into the pathway, not on what fusion it carries. A panel of 11 lines that happened to have higher haem flux than the comparison panel would produce exactly the signal the scan reports, with no fusion biology involved at all.
This is a hypothesis about the artefact, not a demonstration of it. But it is a cheap and falsifiable one: compare ALAS1/ALAS2 expression and baseline porphyrin content between the two DepMap panels, and the explanation stands or falls.
Why recording an empty result is worth a node
Three reasons, and they are the reasons this node exists rather than a silence.
- It stops the search being repeated. The next agent that reads
ana-depmap-kmt2a-selective-scanwill see UROS at rank 15 with no entry and will do exactly what was done here. This node costs less than that search. - It calibrates the scan. Four novel hits went into this beat. One (CCND3) had strong independent literature, one (ZEB2) had strong literature that partly contradicted the scan's framing, one (GFI1) had literature pointing the opposite way on direction, and one has nothing. A one-in-four clean miss is useful information about how much a rank in this scan is worth, and it can only be recorded if the miss is written down.
- A speculative node would have been worse. It would have been easy to write
gene-urosaround haem metabolism, mitochondrial function and the venetoclax connection, cite the Cell Metab paper, and produce something that looked like knowledge. The brief specifically forbids that, and it was right to.
| script | analyses/depmap-kmt2a-selective-scan/run.py |
|---|---|
| status | run |
| result summary | UROS, rank 15 of 4,589 in the unbiased KMT2A-selectivity scan, has no leukemia literature. Four PubMed query formulations returned nothing connecting UROS to leukemia, MLL/KMT2A, infant leukemia or cancer dependency. No gene entry was created, on the brief's rule that an entry is written only where the literature supports one. The searched-and-empty result is recorded here instead. |
| gene | UROS |
| uniprot | P10746 |
| length AA | 265 |
| accession resolution | Resolved by fetching https://rest.uniprot.org/uniprotkb/P10746.json on 2026-09-18 and reading back the primary gene name, which is UROS with no synonyms listed; sequence length 265 aa. Note the UniProtKB ID is HEM4_HUMAN, which does not contain the gene symbol — a case where trusting the entry name rather than the primary gene name would have looked like a mismatch. Not resolved by a gene_exact: search, which matches synonyms. |
| scan rank | 15 |
| scan n genes | 4589 |
| kmt2a selectivity difference | -0.3866 |
| chronos mean kmt2a rearranged | -0.8226 |
| chronos mean other leukemia | -0.436 |
| pan essentiality all lines | -0.1983 |
| n all lines | 1208 |
| entry created | false |
| reason no entry | No credible published connection between UROS and KMT2A-rearranged or any other leukemia. Writing a gene node would have required building it out of pathway-level and adjacent findings, which is the speculative node the brief forbids. |
| cheapest falsifier | Compare ALAS1/ALAS2 expression and baseline porphyrin content between the 11 KMT2A-rearranged and 43 other leukemia DepMap lines. If heme flux differs between the panels, the UROS signal is a metabolite-toxicity artefact and not a dependency. |
queries run
[ "UROS uroporphyrinogen III synthase leukemia (2 hits, none relevant)", "uroporphyrinogen III synthase UROS cancer dependency (2 hits, none relevant)", "uroporphyrinogen III synthase myeloid differentiation heme leukemia cells (0 hits)", "UROS deficiency uroporphyrin I accumulation phototoxicity cells (2 hits, CEP background only)", "heme biosynthesis acute myeloid leukemia dependency (93 hits, triaged; 2 relevant to the pathway, none to UROS)" ]
- Recording this as a false positive could itself be the error. The scan was designed to find things the literature has not reached, and dismissing an unstudied gene because it is unstudied would defeat the purpose of running it. The toxic-intermediate mechanism is the reason for scepticism here, not the empty search.
- The toxic-intermediate account has a weakness of its own: UROS's pan-essentiality is only -0.198, which is milder than a broadly flux-dependent toxicity would predict. The account should be tested, not assumed.
- If haem biosynthesis really does modify apoptotic commitment in AML, and this graph's interest in venetoclax is live, then the pathway may matter in this disease for reasons that have nothing to do with the fusion — and dismissing UROS should not be read as dismissing that.
- Do the 11 KMT2A-rearranged DepMap lines carry higher haem biosynthetic flux than the 43 comparison lines? That single comparison would convert the toxic-intermediate hypothesis from plausible to settled, in either direction.
- Are other enzymes of the haem pathway — ALAD, HMBS, UROD, CPOX, PPOX, FECH — also selective in the scan? If several are, the signal is a pathway-level flux difference and definitely not about UROS. If only UROS is, the toxic-intermediate account gains weight, because UROS is the step whose loss produces the toxic isomer.
- Given that heme-gene expression at diagnosis marks pediatric AML chemotherapy non-response, is there anything to test in KMT2A-rearranged patients specifically? This is a question about the pathway, not about UROS, and would need TARGET or St. Jude Cloud expression data split by fusion.
- How many of the scan's other high-ranking non-pan-essential genes would also come back empty on a literature search? One clean miss in four is a small sample, and the scan's calibration would be much better understood from twenty.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset
UROS ranks 15 of 4,589 protein-coding genes by KMT2A-rearranged selectivity in this project's unbiased DepMap scan (difference -0.387; mean Chronos -0.823 across 11 KMT2A-rearranged leukemia lines versus -0.436 across 43 other leukemia lines; pan-essentiality -0.198 across 1,208 lines), and targeted PubMed searches on 2026-09-18 for UROS with leukemia, with cancer dependency, with myeloid differentiation and with porphyrin phototoxicity returned no primary literature connecting UROS to leukemia of any kind.
DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API, 4,589 HGNC protein-coding genes split by Cellosaurus KMT2A gene-fusion annotation; analyses/depmap-kmt2a-selective-scan/results.json; plus PubMed searches run through the PubMed MCP tools on 2026-09-18 · Broad Institute DepMap; Cellosaurus (SIB); Open Targets Platform; PubMed/NCBI; searches and analysis run by this project · Databases and derived analysis 2026What this evidence cannot show
A negative literature search is bounded by the queries run. Four query formulations were used, all in PubMed, all in English, and none searched conference abstracts, preprint servers or patents. PubMed's automatic term expansion also over-ANDs multi-concept queries and can return zero for a query that a looser search would answer, which is why several formulations were used rather than one. The DepMap numbers are this project's own unreviewed computation on 11 lines, none infant-derived, over 24% of the protein-coding genome, without lineage conditioning.
Evidence that argues against this (1)
- Absence of a PubMed hit is not absence of biology. UROS could be a genuine dependency that nobody has studied, and a scan's job is precisely to surface things the literature has not reached. The argument for calling this a probable false positive is the toxic-intermediate mechanism below, not the empty search on its own. Blouin JM et al. 2020
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
A CRISPR/Cas9 knockout screen for metabolic genes influencing cellular commitment to apoptosis in BCL-2-dependent AML lines identified heme biosynthesis as an apoptosis-modifying pathway, with reduced heme biosynthesis potentiating apoptosis through loss of electron-transport-chain activity and baseline mitochondrial depolarisation; in RNA-seq from 240 pediatric patients across three Children's Oncology Group AML trials, patients who later failed to achieve remission after two rounds of induction expressed higher heme biosynthetic genes at diagnosis than responders, and heme biosynthesis contained the highest proportion of significantly upregulated genes of any metabolic pathway in non-responders.
Systematic Dissection of the Metabolic-Apoptotic Interface in AML Reveals Heme Biosynthesis to Be a Regulator of Drug Sensitivity · Lin KH, Xie A, Rutter JC, Ahn YR, Lloyd-Cowden JM, Nichols AG, Soderquist RS, Koves TR, Muoio DM, MacIver NJ, Lamba JK, Pardee TS, McCall CM, Rizzieri DA, Wood KC · Cell Metabolism 29:1217-1231.e7 2019What this evidence cannot show
Full text read via PubMed Central (PMC6506362). UROS is never mentioned: the string 'uroporphyrin' occurs zero times in the retrieved full text, and the genetic validation used ALAS1, ALAD and FECH. The direction is also wrong for the scan's hit — this paper says losing heme biosynthesis makes AML cells MORE apoptosis-prone, which is a sensitiser relationship, not a dependency. AML only, no KMT2A-rearranged analysis, and the pediatric cohort is COG AML, not infant or KMT2A-selected.
Evidence that argues against this (1)
- The pediatric COG association runs the other way from the drug rationale: high heme-gene expression marks chemotherapy non-response, which would make heme biosynthesis a resistance marker rather than a target, and neither reading says anything about UROS specifically. Hoshii T et al. 2022
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Rapid degron-mediated degradation of the histone methyltransferase SETD1A in AML cell lines immediately downregulates transcripts associated with DNA repair and heme biosynthesis; CRISPR functional analysis and metabolomics show an essential role for SETD1A in maintaining mitochondrial respiration in AML cells; and the effect runs through a non-enzymatic, SETD1A domain-dependent cyclin K function that controls RNA polymerase II promoter-proximal pause release.
SETD1A regulates transcriptional pause release of heme biosynthesis genes in leukemia · Hoshii T, Perlee S, Kikuchi S, Rahmutulla B, Fukuyo M, Masuda T, Ohtsuki S, Soga T, Nabet B, Kaneda A · Cell Reports 41:111727 2022What this evidence cannot show
Read as abstract only. SETD1A is a KMT2-family methyltransferase but it is NOT KMT2A, and this paper is about AML generally, not KMT2A-rearranged leukemia. UROS is not named in the abstract. It is recorded here only because it is the closest thing to a chromatin-to-heme-biosynthesis link in leukemia and because someone will otherwise find it and build a chain of inference on it. That chain should not be built.
Evidence that argues against this (1)
- If heme-biosynthesis gene transcription in leukemia is controlled by SETD1A rather than by KMT2A or its fusions, the observation argues against, not for, a KMT2A-fusion-specific heme dependency. Lin KH et al. 2019
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Congenital erythropoietic porphyria is an inborn error of heme synthesis resulting from uroporphyrinogen III synthase (UROS) deficiency and the accumulation of non-physiological porphyrin isomer I metabolites, combining skin photosensitivity and chronic hemolytic anemia whose severity is related to porphyrin overload; reducing flux through the pathway by downregulating the upstream rate-limiting enzyme ALAS2, including by iron chelation with deferiprone, reduces porphyrin production in UROS-deficient erythroid cell lines, in patient-derived CD34+ erythroid cultures and in a CEP mouse model, reversing photosensitivity and correcting anemia.
Iron chelation rescues hemolytic anemia and skin photosensitivity in congenital erythropoietic porphyria · Blouin JM, Ged C, Lalanne M, Lamrissi-Garcia I, Morice-Picard F, Costet P, Daher R, Moreau-Gaudry F, Bedel A, Puy H, Gouya L, Karim Z, Richard E · Blood 136:2457-2468 2020What this evidence cannot show
Read as abstract only. This is a therapy paper for a rare inherited disease and has nothing to do with leukemia; it is cited solely to establish what UROS loss does to a cell — accumulate toxic isomer I porphyrins whose damage scales with upstream flux. The inference from that to 'the DepMap signal is a flux artefact' is this Queen's, not the paper's, and is stated in the sections as a falsifiable hypothesis rather than a finding.
Evidence that argues against this (1)
- The toxic-intermediate account predicts that UROS knockout should hurt any high-flux cell and therefore show up as broadly essential rather than KMT2A-selective. Its pan-essentiality of -0.198 is mild, which is not obviously what that account predicts, so the artefact hypothesis is not free of problems either. analysis run by this project 2026
- ← refutesThe whole protein-coding genome: menin is still first, and a metabolic pathway appears (run)
the toxic-metabolite model predicts upstream loss is protective, but ALAS1 and ALAD loss are selectively deleterious; four pathway members rank in the top 47
neither restores nor removes the clean negative: the haem cluster stays open in both directions
UROS at rank 15 has no leukemia literature at all; one clean miss in four novel hits, recorded as calibration for the scan
the nearest near-miss: SETD1A controls pause release of heme biosynthesis genes in leukemia, but SETD1A is not KMT2A and UROS is not named
- related to →BCL2 — the one co-dependency in this beat that already has an approved drug pointed at it
heme biosynthesis loss potentiates apoptosis in BCL-2-dependent AML, which is a sensitiser relationship pointing opposite to the scan's dependency reading