Mutation Profile
Five driver mutations across six oncogenic axes. N=1 myeloid case. MRD-negative, 35 months post-HSCT.
| Gene | Variant | VAF | Domain | Classification | Actionability |
|---|---|---|---|---|---|
| DNMT3A | R882H | 39% | Methyltransferase | Pathogenic | No direct inhibitor |
| IDH2 | R140Q | 2% | Active site | Pathogenic | Enasidenib (FDA 2017) |
| SETBP1 | G870S | 34% | SKI homology | Pathogenic | No direct inhibitor |
| PTPN11 | E76Q | 29% | N-SH2 | Pathogenic | SHP2 inhibitors (investigational, solid-tumour trials only) |
| EZH2 | V662A | 59% | SET domain | Pathogenic | Tazemetostat CONTRAINDICATED (LoF) |
| Mono 7 | -7 | — | — | Adverse | IPSS-M risk factor |
Cytogenetics Monosomy 7 ELN 2022
Adverse
IPSS-M
Very High (3.24)
Status
MRD-negative, full donor chimerism
0 matches in 8,655 screened patients
(GENIE myeloid patients with all five genes on panel).
The exact combination has no estimable frequency: zero in 8,655 patients
bounds it below 1 in 2,889 (95%, exact binomial).
In 2013, recurrent SETBP1 mutations were identified in the
SKI-homologous degron motif of atypical CML patients
1
[1] R 2013
Recurrent SETBP1 mutations in atypical chronic myeloid leukemia. Nat Genet (2013)
,
independently confirmed by Makishima et al.
2[2] H 2013
Somatic SETBP1 mutations in myeloid malignancies. Nat Genet (2013)
This portal extends the SETBP1 partnership landscape to 21,017 myeloid
samples in GENIE v19.0
17[17] Consortium 2017
AACR Project GENIE: Powering Precision Medicine through an International Consortium. Cancer Discov (2017)
, integrating co-occurrence
statistics with protein language model scoring
18[18] Z 2023
Evolutionary-scale prediction of atomic-level protein structure with a language model. Science (2023)
, molecular
docking, and formal ACMG evidence aggregation across 15 computational tools.
Patients Screened
8,655
GENIE v19.0 myeloid patients with all five genes on panel · HARMONY not queried
Quintuple Matches Zero
0
Exact 95% bound: below 1 in 2,889
IPSS-M Risk
Very High
Score: 3.24 · inputs DNMT3A, IDH2, EZH2, SETBP1, PTPN11, monosomy 7, marrow blasts 12%, diagnostic Hb and platelets
ELN 2022
Adverse
Median OS: 9.7 months (Lachowiez 2023)
Actual Survival Alive
28+ mo
MRD negative, full donor chimerism
DMS Validated
1/5
PTPN11 E76Q (enrichment 0.329), wet-lab PS3_Strong
Progressive Mutation Filtering
Zero matches in 8,655 screened patients.
That is the empirical finding in GENIE v19.0, the only database searched for all five
variants together (HARMONY could not be queried for co-occurrence). Earlier four-gene searches
without EZH2 in cBioPortal (46 studies) and GDC (24 projects) found no match either; ICGC (API
offline), ClinVar (no co-occurrence data) and 7 further databases were not searched. What each
cohort size counts is set out in the
denominator table.
What the zero supports: 0 of 8,655 GENIE v19.0 myeloid patients with all five genes on the sequencing panel carry this combination. None carries coding mutations in all five genes, and none carries EZH2 V662A.The exact five-variant combination has no estimable population frequency. Zero in 8,655 patients bounds it below 1 in 2,889 (95%, exact binomial).
Gene level: the five genes are not independent. 8 of 10 pairs are enriched (co-occur more than expected; patient-level O/E 31
What the zero supports: 0 of 8,655 GENIE v19.0 myeloid patients with all five genes on the sequencing panel carry this combination. None carries coding mutations in all five genes, and none carries EZH2 V662A.The exact five-variant combination has no estimable population frequency. Zero in 8,655 patients bounds it below 1 in 2,889 (95%, exact binomial).
Gene level: the five genes are not independent. 8 of 10 pairs are enriched (co-occur more than expected; patient-level O/E 31
[31] C 2013
Mutational landscape and significance across 12 major cancer types. Nature (2013)
): SETBP1+EZH2 (O/E=3.82), SETBP1+PTPN11 (O/E=2.63), DNMT3A+IDH2 (O/E=2.29), PTPN11+EZH2 (O/E=2.05), DNMT3A+PTPN11 (O/E=1.63), IDH2+PTPN11 (O/E=1.24), IDH2+EZH2 (O/E=1.22), IDH2+SETBP1 (O/E=1.12).
A pairwise maximum-entropy (Ising) model fitted to GENIE predicts mutations in all five genes in about 1 in 25,000 myeloid patients (95% bootstrap interval 1 in 13,000 to 1 in 65,000; sampling error only). Summed over the five four-gene subsets, the fitted pairwise model
predicts 6.4 patients where 2 are
observed, so the five-gene frequency is probably lower than this estimate. An in-house clonal
model on the PyClone-VI framework 16[16] S 2020
PyClone-VI: scalable inference of clonal population structures using whole genome data. BMC Bioinformatics (2020)
favours a linear
evolution model.
Source: AACR Project GENIE v19.0.
Counts are samples: first bar all 21,017 myeloid samples after the hypermutation filter
(>20 coding mutations excluded), later bars samples carrying every listed variant.
Clinical Annotation
OncoKB
Level 1
IDH2 R140Q: FDA-approved enasidenib
ACMG Classification
4 P + 1 LP
9-21 evidence points (research classification; EZH2 V662A is a VUS in ClinVar)
ClinGen Validity
7 Definitive
10 ClinGen curations, all germline syndromes; none for a myeloid disease
Novel Variant
EZH2 V662A
0 PubMed hits, unreported in literature
DMS Validated
1/5
PTPN11 E76Q, wet-lab PS3_Strong
One of five variants has a direct deep mutational scanning score
29
[29] Z 2025
Deep mutational scanning of the multi-domain phosphatase SHP2 reveals mechanisms of regulation and pathogenicity. Nat Commun (2025)
: PTPN11 E76Q
(enrichment 0.329, 99th percentile of 12,054 SHP2 variants scored).
For DNMT3A R882H, a paired DMS preprint (Garcia et al. 2025, bioRxiv) maps mutations that modify its
dominant-negative effect; no per-variant score from it is used here.
Database Links
Direct links to authoritative databases for each variant. One-click access for verification and cross-referencing.
Pathogenicity Profile
Figure 2 | Multi-tool pathogenicity concordance for five patient variants.
All scores normalized to 0–1. AlphaMissense, CADD PHRED (/40), ESM-2 |LLR| (/12), REVEL,
and site conservation shown per variant.
DNMT3A R882H
PP3_Strong
AM: 0.9953 | CADD: 28.5 | LLR: -8.383
IDH2 R140Q
PP3_Strong
AM: 0.9872 | CADD: 32.0 | LLR: -1.4782
SETBP1 G870S
PP3_Strong
AM: 0.9962 | CADD: 28.7 | LLR: -9.8042
PTPN11 E76Q
PP3_Strong
AM: 0.9972 | CADD: 28.6 | LLR: -1.8651
EZH2 V662A
PP3_Strong
AM: 0.9984 | CADD: 28.1 | LLR: -2.9657
EZH2 V662A classification: Initially a VUS. The ACMG
aggregation now returns Likely Pathogenic at 9 points, on
6/7 concordant computational predictors: AlphaMissense
19
[19] J 2023
Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science (2023)
0.9984, CADD 28.1,
REVEL 0.781, ESM-2
18[18] Z 2023
Evolutionary-scale prediction of atomic-level protein structure with a language model. Science (2023)
20[20] N 2023
Genome-wide prediction of disease variant effects with a deep protein language model. Nat Genet (2023)
LLR=-2.966.
EVE dissents at 0.7825, which its own class25
calibration labels Uncertain. This is not a passenger.
It is classified Likely Pathogenic here (research classification; ClinVar: VUS), on a fifth oncogenic axis
(PRC2-mediated chromatin remodeling). Chase et al.
(Leukemia, 2020)
10[10] A 2020
Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020)
report that SET/CXC domain missense
mutations abrogate catalytic activity in in vitro assays; V662A itself was not tested.
Tazemetostat (EZH2 inhibitor) is CONTRAINDICATED
because V662A is loss-of-function, not gain-of-function.
Source: AlphaMissense (DeepMind), CADD PHRED
(Ensembl VEP CADD v1.7), ESM-2 masked marginal log-likelihood ratio (Lin 2023).
PP3 strength per variant, from acmg_aggregation.json as regenerated on
2026-11-03: DNMT3A PP3_Strong, IDH2 PP3_Strong, SETBP1 PP3_Strong, PTPN11 PP3_Strong, EZH2 PP3_Strong.
Clonal Architecture
An in-house Bayesian clonal model on the PyClone-VI framework
16
[16] S 2020
PyClone-VI: scalable inference of clonal population structures using whole genome data. BMC Bioinformatics (2020)
(PyClone-VI itself was not installed) favours a
linear evolution model (log-likelihood -0.51
vs -14.11 for branching).
DNMT3A R882H is the founder (CCF 1.00), followed by
EZH2 V662A (CCF 0.92),
SETBP1 G870S (CCF 0.87),
PTPN11 E76Q (CCF 0.74),
and IDH2 R140Q as a minor subclone (CCF 0.05).
The IDH2+SETBP1 co-occurrence within this patient is uncommon at population level:
1 of 2,957 patients in the IPSS-M cohort 13[13] E 2022
Molecular International Prognostic Scoring System for Myelodysplastic Syndromes. NEJM Evid (2022)
(our calculation: OR=0.22, one-sided p=0.07, not significant) and O/E 0.91 in GENIE (sample level).
Source: in-house reimplementation of the PyClone-VI model (Gillis & Roth, 2020).
Cancer cell fractions estimated from variant allele frequencies
with purity and ploidy correction.
Oncogenic Convergence
Figure 3 | Oncogenic convergence on MYC.
Five driver mutations feed into five distinct pathways, all converging on MYC
transcription or protein stability.
Five mutations spanning six oncogenic axes:
epigenetic reprogramming (DNMT3A),
metabolic transformation (IDH2),
RAS-MAPK hyperactivation (PTPN11),
PP2A tumor suppressor loss (SETBP1),
PRC2 chromatin remodeling (EZH2),
and chromosomal instability (monosomy 7).
All converge on MYC. Each axis independently promotes
MYC transcription or protein stability through distinct
molecular mechanisms
12
[12] E 2016
Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med (2016)
.
Karyotype: Monosomy 7
Loading karyotype...
45,XY,-7[9]/46,XY[1]. Five somatic driver mutations annotated at genomic positions.
References
- Piazza R et al. Recurrent SETBP1 mutations in atypical chronic myeloid leukemia. Nat Genet (2013). PubMed
- Makishima H et al. Somatic SETBP1 mutations in myeloid malignancies. Nat Genet (2013). PubMed
- AACR Project GENIE Consortium. AACR Project GENIE: powering precision medicine through an international consortium. Cancer Discov (2017). DOI
- Lin Z et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science (2023). DOI
- Kandoth C et al. Mutational landscape and significance across 12 major cancer types. Nature (2013). PubMed
- Canisius S et al. A novel independence test for somatic alterations in cancer shows that biology drives mutual exclusivity. Genome Biol (2016). DOI
- Gillis S, Roth A. PyClone-VI: scalable inference of clonal population structures. BMC Bioinformatics (2020). DOI
- Papaemmanuil E et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med (2016). PubMed
- Cheng J et al. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science (2023). DOI
- Brandes N et al. Genome-wide prediction of disease variant effects with a deep protein language model. Nat Genet (2023). DOI
- Bernard E et al. Molecular International Prognostic Scoring System for myelodysplastic syndromes. NEJM Evid (2022). DOI
- Jiang Z et al. Deep mutational scanning of the multi-domain phosphatase SHP2 reveals mechanisms of regulation and pathogenicity. Nat Commun (2025). DOI
- Chase A et al. Mutational mechanisms of EZH2 inactivation in myeloid neoplasms. Leukemia (2020). DOI