CTH (P32929) - gene_exact returns TWO reviewed records and the second is VSIG2, a third instance of rule 62 on the same symbol the Ensembl resolver already got wrong
also called cystathionine gamma-lyase · CGL_HUMAN · cystathionase
CTH finishes the job CBS starts, releasing cysteine so the cell can make the antioxidant glutathione. A menin inhibitor drops it about three-fold - and the symbol CTH is also a registered alias of a completely unrelated immune-type membrane protein, which has now tripped two different databases in this project.
The same thing, for a scientist
CTH (P32929, 405 aa, CGL_HUMAN) catalyses the last step of transsulfuration, cleaving L,L-cystathionine into L-cysteine, ammonia and 2-oxobutanoate, part of which cysteine supplies glutathione synthesis; it also produces hydrogen sulfide [E3]. **The identifier finding comes first because it is the most transferable thing here: `gene_exact:CTH AND organism_id:9606 AND reviewed:true` returns TWO reviewed human records** - P32929 (405 aa, cystathionine gamma-lyase) and **Q96IQ7 (327 aa, VSIG2_HUMAN, V-set and immunoglobulin domain-containing protein 2, which carries CTH as a gene synonym)** [E3]. `data/symbol-collisions.json` already records CTH/VSIG2 as an **Ensembl resolver** problem; this beat shows it is **also** a UniProt `gene_exact` problem, making it the **third** documented instance of rule 62's two-record trap after DDIT3 and SFN - and the only one where the same pair of genes had already burned a different instrument in this project. **Measured here:** RS4;11 day 7, VTP-50469 **-1.57** from a detected baseline of 4.35 log2 CPM, EPZ5676 -0.82, margin **-0.75**; but its day-2 value is **+0.105**, giving a NEGATIVE day-2 share, the only serine-arm gene with that profile, so it is the arm's least stable member [E1]. **And the cycle-56 MV4;11 entry lists CTH under 'did not replicate' at +0.24: its MV4;11 baseline is 0.29 log2 CPM, BELOW the detection floor. That is an absence of signal, not a non-replication** [E1]. DepMap: mean **+0.0699, 0.0% of 1,258 below -0.5, leukaemia mean +0.1016** - a positive gene effect, the only one in the panel [E2]. Germline loss causes **cystathioninuria, a benign biochemical phenotype** [E3] - the mildest germline consequence of any enzyme in this beat.
This is real experimental data — atoms placed by crystallography or cryo-EM, deposited by the people who did the experiment.
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.
The identifier trap, and why this instance is worth a rule refinement
Three facts, in the order they were established 3:
- The registry already knew about CTH/VSIG2 -
data/symbol-collisions.jsoncarries a row measured earlier this same cycle: the Ensembl symbol endpoint returns ENSG00000019102 (VSIG2) FIRST forCTH, and the correct id is ENSG00000116761. This beat read that row before declaring its panel, which is what rule 88 asks for, and it worked - the pre-registration names the collision in advance. - The registry's row describes the wrong instrument's failure as if it were the only one. It is written as an Ensembl-resolver problem.
gene_exact:CTH AND organism_id:9606 AND reviewed:trueon UniProt also returns two records, P32929 and Q96IQ7, and a beat that read the registry row would not expect that. - The length check happens to save you here and that is luck. 405 aa against 327 aa is 78 apart, well outside the SMC1A/SMC3 limit of 16 that rule 66 documents - but the CTH/VSIG2 pair is not a paralogue pair and there is no reason its lengths should differ. The defence that actually works is reading
gene_primaryback, which returns CTH for P32929 and VSIG2 for Q96IQ7.
The refinement worth recording: a collision row should name the OBJECT PAIR, not the instrument. A row saying "the Ensembl endpoint gets CTH wrong" invites a reader to trust UniProt for the same symbol. A row saying "CTH is simultaneously the approved symbol of P32929/ENSG00000116761 and a registered alias of VSIG2/Q96IQ7" travels to every resolver.
The literature side is cleaner than CBS's: CTH[tiab] returns 1,333, its single fusion-scoped hit is PMID 36177864, Triangular Mechanical Structure of the Proximal Femur, and the alias CGL[tiab] returns 944 with 0 fusion-scoped - so the genuine fusion-scoped count is zero 4.
A floor artefact in an existing entry, corrected here and not edited there
ana-the-serine-arm-replicates-in-a-third-aff1-line-from-another-lab-but-the-stress-output-genes-do-not lists CTH in its did_not_replicate_the_isr_output block as "+0.24 against -1.57 in RS4;11".
CTH's MV4;11 Safe3 DMSO baseline is 0.29 log2 CPM, below the 1.0 detection floor declared in this beat 1. The gene is not expressed in that cell line at that timepoint, so +0.24 is noise around zero counts and is not a failure to replicate. The honest statement is that CTH could not be tested in MV4;11.
That entry belongs to another beat and is reported rather than edited. The change it needs is one sentence: move CTH out of did_not_replicate_the_isr_output and into a not-detected category, with the baseline of 0.29 log2 CPM beside it. The source that settles it is that beat's own results file plus this beat's floor declaration.
The general point is this project's own standing rule about printing a detection floor beside every null, applied to a non-replication rather than to a null. A gene that fails to replicate and a gene that is not expressed look identical in a fold-change table and mean opposite things.
| gene symbol | CTH |
|---|---|
| uniprot | P32929 |
| length AA | 405 |
| entry name | CGL_HUMAN |
| protein name | Cystathionine gamma-lyase |
| ensembl gene ID | ENSG00000116761 |
| ensembl readback | display_name read back as CTH before use - REQUIRED for this symbol: the Ensembl symbol endpoint returns ENSG00000019102 (VSIG2) as its FIRST hit for CTH, and this project's own registry records that failure. |
| PDB note | All nine PDB cross-references of UniProt P32929, read off the live cross-reference list on 2026-09-22. None of the structures themselves was fetched and no chain, resolution or residue range is claimed. |
| accession resolution | RULE 62 VIOLATED FOR THIS SYMBOL. gene_exact:CTH AND organism_id:9606 AND reviewed:true returns TWO reviewed rows: P32929 (CGL_HUMAN, 405 aa, gene_primary CTH) and Q96IQ7 (VSIG2_HUMAN, 327 aa, gene_primary VSIG2, carrying CTH and CTXL as synonyms). Selected on gene_primary read-back, not on length - the 78 aa gap discriminates here but is coincidental. THIRD instance in this project after DDIT3 and SFN. |
| localization | CYTOPLASM, but UNCODED. UniProt P32929's single SUBCELLULAR LOCATION statement carries no ECO evidence code and no citation, so it is a curator assertion rather than a measurement - recorded as uncoded per rule 91 rather than discarded. No localisation experiment on CTH has been performed in any KMT2A-rearranged cell. Note for contrast that CBS, the enzyme immediately upstream in the same pathway, has coded cytoplasmic AND nuclear statements. |
| subcellular localization | CYTOSOLIC by uncoded UniProt annotation. A cytosolic inflammasome-associated function has been reported recently (PMID 42062561, title only, not read), which would be a compartment-specific claim this entry has not verified. |
| subnuclear localization | NONE, answered negative rather than left blank: no nuclear pool is annotated or reported for CTH, unlike CBS immediately upstream of it, which carries an ECO:0000269 Nucleus statement. |
| cell type | RS4;11 and SEM (KMT2A::AFF1 B-ALL), REH (KMT2A wild type - both probes detected and both rising, so a usable control here), MOLM13 (KMT2A::MLLT3 AML) measured here; NOT TESTABLE in MV4;11, where the baseline is 0.29 log2 CPM below floor. Per UniProt, highly expressed in liver and muscle and HARDLY DETECTED IN THYMUS, spleen and heart - a lymphoid-tissue expression statement worth noting in a lymphoid disease. NO primary infant KMT2A-rearranged sample and NO normal human pro-B cell. |
| cell state | Measured in the drug-treated arresting state at day 7 of continuous menin-inhibitor exposure under uncontrolled medium methionine and cystine supply, which for a transsulfuration enzyme is the reaction's own substrate context. CTH is the only serine-arm gene whose day-2 and day-7 values disagree in SIGN (+0.105 then -1.57), so its kinetics are uninterpretable here. The state with a human phenotype is a benign biochemical one: cystathioninuria, abnormal plasma cystathionine with increased urinary excretion and no reliable clinical syndrome. |
| epigenetic readout | NONE MEASURED at the CTH locus in any menin- or KMT2A-scoped study, and CTH is not among the SSP genes named in PMID 29672864. No H3K4me3, H3K79me2, KMT2A or menin ChIP exists. Bounded absence. Given the instability of this gene's transcript measurement, it is a low priority for the ChIP this beat's other entries request - stated so that a reader does not add loci indiscriminately. |
| shown for fusion | READOUT ONLY and WEAKLY. -1.57 in KMT2A::AFF1 RS4;11 with a sign-discordant day-2 value, -1.08/-0.94 on two array probes in SEM under the weak compound, -0.37 in KMT2A::MLLT3 MOLM13, and NOT TESTABLE in MV4;11 (baseline below floor). The genuine fusion-scoped literature count is zero once the femur-mechanics false positive is removed. No mechanism, chromatin measurement or dependency in any fusion. |
| absence class | measured-not-analysed |
| absence class note | The transcript response was in deposited data and unanalysed. The MV4;11 correction is not an absence at all - it is a misclassified floor value in an existing entry. |
PDB IDS
[ "2NMP", "3COG", "3ELP", "5EIG", "5TSU", "5TT2", "6NBA", "6OVG", "9KHN" ]
the measurement rs411 d7 log2fc
{
"vtp50469_330nM": -1.57,
"epz5676_1uM": -0.82,
"vtp_minus_epz": -0.752,
"epz_fraction_of_vtp": 0.52,
"dmso_baseline_log2cpm": 4.35,
"detected": true,
"floor_collision": false,
"d2_vtp": 0.105,
"d2_share_of_d7": -0.07,
"warning": "the day-2 and day-7 values disagree in sign; this is the least stable member of the serine arm"
}molm13 kmt2a mllt3
{
"vtp": -0.368,
"epz": -0.539,
"dmso_baseline_log2cpm": 4.85,
"detected": true
}mv411 independent lab
{
"value_reported_elsewhere": 0.24,
"dmso_baseline_log2cpm": 0.29,
"detected": false,
"correction": "BELOW the 1.0 log2 CPM floor. The cycle-56 entry lists CTH under 'did_not_replicate_the_isr_output' at +0.24 against -1.57; that is an absence of signal, not a non-replication, and CTH could not be tested in MV4;11. Reported for correction, not edited - this beat does not own that entry."
}the measurement array gse63664 every probe
{
"206085_s_at": {
"rs411_dmso_log2_rma": 6.85,
"rs411_mi22": -0.37,
"sem_dmso": 8.82,
"sem_mi22": -1.08,
"reh_dmso": 6.91,
"reh_mi22": 0.43
},
"217127_at": {
"rs411_dmso_log2_rma": 7.38,
"rs411_mi22": -0.38,
"sem_dmso": 9.42,
"sem_mi22": -0.94,
"reh_dmso": 7.46,
"reh_mi22": 0.47
},
"detection_floor_used": "6.0 log2 RMA; both probes above it in all three lines, so REH is a usable control for CTH and it moves in the opposite direction"
}essentiality
{
"depmap_release": null,
"depmap_release_note": "Rule 83: no release identifier in the Open Targets payload. Fetched 2026-09-22.",
"depmap_screens": 1258,
"depmap_tissue_blocks_flattened": 29,
"chronos_mean": 0.0699,
"chronos_median": 0.0731,
"pct_lines_below_minus_0_5": 0,
"leukaemia_screens": 57,
"leukaemia_mean_gene_effect": 0.1016,
"depmap_common_essential": "not asserted",
"reading": "the ONLY positive mean gene effect in the panel, and zero of 1,258 screens below -0.5. Not asserted as a growth-promoting knockout; read as a clean non-dependency under cysteine-replete conditions.",
"readback_was_load_bearing": "an unread resolver hit would have returned VSIG2 and the table would have looked plausible"
}identifier collision measured
{
"uniprot_gene_exact_rows": 2,
"rows": {
"P32929": "CGL_HUMAN, 405 aa, gene_primary CTH - the transsulfuration enzyme",
"Q96IQ7": "VSIG2_HUMAN, 327 aa, gene_primary VSIG2, synonyms CTH and CTXL - an Ig-domain membrane protein"
},
"ensembl_first_hit": "ENSG00000019102 = VSIG2; correct id ENSG00000116761",
"CTH[tiab]": 1333,
"fusion_scoped_raw": 1,
"fusion_scoped_genuine": 0,
"false_positive_named": {
"36177864": "Triangular Mechanical Structure of the Proximal Femur"
},
"CGL[tiab]": 944,
"cgl_unmeasured_risk": "CGL also abbreviates chronic granulocytic leukaemia in older haematology literature; that contamination was NOT measured here and would be dangerous in this corpus",
"registry_status": "data/symbol-collisions.json HAS a CTH/VSIG2 row, written earlier this same cycle, and it describes the ENSEMBL failure only. The UniProt gene_exact instance is new and is reported for the coordinator to add to the same row.",
"rule_refinement_proposed": "a collision row should name the OBJECT PAIR and both accessions, not the instrument that happened to fail - a row about the Ensembl endpoint invites a reader to trust UniProt for the same symbol"
}- **CTH's germline phenotype is benign** - cystathioninuria, a biochemical finding with no reliable clinical syndrome - which makes it the mildest germline consequence of any enzyme in this beat and the only one where an essential-function argument is weak. That is not a licence: the terminal transsulfuration step supplies cysteine for glutathione, and glutathione depletion is a ferroptosis mechanism in every proliferating tissue.
- **CTH is hardly detected in thymus and spleen per UniProt**, so a transsulfuration-directed agent would act on a pathway that lymphoid tissue barely runs - which cuts against, not for, a lymphoid-selective rationale.
- **This entry's transcript measurement is the weakest in the tranche and is graded weak overall.** It is written because the identifier finding is transferable and because it corrects a floor artefact in an existing entry, not because CTH is an interesting node in this biology.
- Does CTH move reproducibly at all? Day 2 is +0.105 and day 7 is -1.57 from a 4.35 log2 CPM baseline in one line, MOLM13 gives -0.37, and MV4;11 cannot test it. A qPCR time course in RS4;11 and SEM would establish whether this gene belongs in the serine arm; the beat's verdict does not depend on it either way.
- Is CTH an ATF4 target in this disease? An ATF4/CTH/SQOR axis is reported in renal cell ferroptosis (PMID 41759595, title only). If CTH is ATF4-proximal it should have a small compound margin like PSAT1's, and its measured -0.75 is mid-arm - so the prediction is testable and currently unmet.
- Does the CGL alias hide chronic-granulocytic-leukaemia contamination? CGL[tiab] returns 944 records and the overlap with CTH was not measured. In a leukaemia corpus that is a live hazard and one esearch intersection would settle it.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
On a pre-registered panel CTH falls 1.57 log2 in RS4;11 at day 7 under VTP-50469 330 nM from a detected baseline of 4.35 log2 CPM against 0.82 under EPZ5676 (margin -0.75), but its day-2 value is +0.105, giving a negative day-2 share of the day-7 effect and making it the only serine-arm gene with that profile; it is -0.37 in MOLM13 from a detected baseline of 4.85; on the array both probes fall in RS4;11 (-0.37 at baseline 6.85, -0.38 at 7.38) and more in SEM (-1.08 at 8.82, -0.94 at 9.42) and both RISE in REH (+0.43 at 6.91, +0.47 at 7.46); and in MV4;11 its Safe3 DMSO baseline is 0.29 log2 CPM, BELOW the 1.0 detection floor, so the +0.24 that a previous entry reports as a failure to replicate is an absence of signal rather than a non-replication.
What this evidence cannot show
One line per platform, single doses, NO dose-response, n=2-3 per arm, day 7 of continuous exposure so arrest is a competing explanation. No series records the amino-acid composition of the medium, and this is an amino-acid-response panel. MI-2-2 in the array arm is the WEAK compound, so the REH wild-type contrast has little dynamic range. CTH's RS4;11 baseline of 4.35 log2 CPM is the lowest of any gene carrying a margin in the serine arm, so its -1.57 is measured near the low end of the dynamic range. The negative day-2 share means the day-2 and day-7 measurements disagree in sign, which for a gene at this baseline is most simply read as noise - and the entry says so rather than constructing a late-response story.
Evidence that argues against this (1)
- A gene whose day-2 value is +0.105 and day-7 value is -1.57, measured from a 4.35 log2 CPM baseline in one line, is the weakest member of the serine arm and its -0.75 margin should carry the least weight in the beat's median of -0.55. Removing CTH from the arm entirely would move that median and the verdict does not depend on it, but a reader should know it is in there. this repository 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
CTH DepMap gene effect is mean +0.0699, median +0.0731, with 0.0% of 1,258 screens below -0.5 and a leukaemia mean of +0.1016 over 57 screens - the only POSITIVE mean gene effect in the pre-registered panel - with all 29 tissue blocks flattened and approvedSymbol read back as CTH, which is the readback that matters for this symbol.
Open Targets Platform GraphQL target.depMapEssentiality, all 29 tissue blocks flattened, panel-maximum payload guard PASS at 1,258 screens; analyses/q109-serine-one-carbon-depmap-dependency/results.json · Open Targets / DepMap 2026What this evidence cannot show
Rule 83: the payload carries NO release identifier, so depmap_release is recorded as null and absolute values drift quarterly; the cross-gene contrast is the durable part. Leukaemia screens are matched on the payload's own diseaseFromSource string against four substrings and are NOT split by KMT2A status. Rule 84: no median-split proportion is reported - the percentage is against an absolute -0.5 threshold. A CRISPR knockout in cysteine- and serine-replete medium understates a biosynthetic dependency. A positive mean gene effect is within noise and is NOT asserted here as a growth-promoting knockout phenotype. The readback is load-bearing for this gene specifically: an unread resolver hit would have returned VSIG2, and the printed table would have looked entirely plausible.
Evidence that argues against this (1)
- Cysteine is supplied in standard medium and imported by SLC7A11, so a knockout of the terminal transsulfuration enzyme is measured under conditions where its product is free. The opposing literature makes exactly this point with an FTO/SLC7A11/CBS/CTH axis controlling cysteine metabolism and survival in NSCLC - and SLC7A11 is in this beat's panel with an inverted compound-selectivity (+0.13 menin, -1.84 DOT1L) that nothing here explains. Kuganesan N et al. 2026
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
gene_exact:CTH AND organism_id:9606 AND reviewed:true returns TWO reviewed human records: P32929 (CGL_HUMAN, 405 aa, primary gene name CTH, recommended name Cystathionine gamma-lyase, nine PDB cross-references, annotated Cytoplasm with NO evidence code, catalysing the last step of the trans-sulfuration pathway from L-methionine to L-cysteine in a PLP-dependent manner by cleaving L,L-cystathionine into L-cysteine, ammonia and 2-oxobutanoate with part of the cysteine used for glutathione biosynthesis, highly expressed in liver and muscle and hardly detected in heart, pituitary, spleen, thymus and vascular tissue, CSE promoter activity regulated by oestrogen receptor alpha, and recording cystathioninuria, an autosomal recessive phenotype of abnormal plasma cystathionine accumulation with increased urinary excretion) and Q96IQ7 (VSIG2_HUMAN, 327 aa, primary gene name VSIG2, V-set and immunoglobulin domain-containing protein 2, carrying CTH and CTXL as gene synonyms, annotated Membrane with ECO:0000305).
UniProtKB search gene_exact:CTH AND organism_id:9606 AND reviewed:true, and entry P32929 (CGL_HUMAN), Cystathionine gamma-lyase · UniProt Consortium · UniProt 2026What this evidence cannot show
Run 2026-09-22. This is the THIRD documented instance in this project of gene_exact returning more than one reviewed human record, after DDIT3 (169 aa factor plus a 34 aa uORF-encoded peptide) and SFN (14-3-3 sigma plus mitochondrial oligoribonuclease). The length gap here is 78 aa so the length check happens to discriminate - but that is luck, not a defence, and the reliable test is reading gene_primary back. The Cytoplasm statement on P32929 is UNCODED, so CTH's localisation is a curator assertion. The nine PDB entries were not fetched.
Evidence that argues against this (1)
- The registry row for this collision describes it as an ENSEMBL resolver failure, which is how it was found. Reading that row would not have warned a beat about UniProt, and this entry is evidence that a collision row written around one instrument understates the problem. The generalisable form is to record the OBJECT PAIR - CTH is simultaneously the approved symbol of P32929/ENSG00000116761 and a registered alias of VSIG2/Q96IQ7/ENSG00000019102 - which travels to every resolver. this repository 2026
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PubMed 2026-09-22, one tag per term with translations read back: CTH[tiab] returns 1,333 records; CTH[tiab] AND the eight-operand KMT2A-fusion set returns 1, PMID 36177864, whose title was opened and is 'Triangular Mechanical Structure of the Proximal Femur' - a false positive, so the genuine fusion-scoped count is ZERO; CTH[tiab] AND the nine-operand menin/menin-inhibitor set returns 0; and the alias CGL[tiab] returns 944 with 0 fusion-scoped.
What this evidence cannot show
Bounded by titles and abstracts only. The 944 CGL records were not id-set-intersected with CTH's 1,333, so only CGL's fusion-scoped zero is used; CGL also abbreviates chronic granulocytic leukaemia in older haematology literature, which would be a dangerous contaminant in this corpus and was NOT measured here. The 1,333-record total is a positive count of a three-letter token and should not be quoted as 'the CTH literature'.
Evidence that argues against this (1)
- CTH is reported as a direct ATF4 target elsewhere - an ATF4/CTH/SQOR axis drives renal cell ferroptosis (PMID 41759595, title only, not read) - which would make CTH an ATF4-proximal gene like PSAT1 and predict a SMALL compound margin. Its measured margin is -0.75, mid-arm, and its day-2/day-7 sign disagreement means the data cannot adjudicate. So the one panel gene with a published ATF4 link outside PSAT1 is also the one whose measurement is least reliable. Pan W et al. 2026
CBS and CTH are the two steps of transsulfuration; CBS -4.07 and CTH -1.57 in RS4;11 day 7
an ATF4/CTH/SQOR axis is published in renal ferroptosis (PMID 41759595, title only); CTH mid-arm margin of -0.75 neither confirms nor refutes an ATF4-proximal position
its CTH did-not-replicate row is a detection-floor artefact: MV4;11 baseline 0.29 log2 CPM, below floor
- related to →KMT2A-rearranged acute lymphoblastic leukaemia
transcript falls 1.57 log2 under a menin inhibitor in RS4;11 with a sign-discordant day-2 value; the least stable member of the serine arm
the rule-62 two-record instance and the Ensembl/UniProt collision measured on this symbol
the least stable member of the serine arm, and the source of a floor-artefact correction to an earlier entry