Skip to main content
AACR Project GENIE v19.0 · 21,017 myeloid samples Panel-adjusted Fisher's exact with Benjamini-Hochberg FDR N=1 case study · Not clinical guidance

Network Biology and Pathway Convergence

STRING network, GenePT-inspired gene embeddings, literature-curated synthetic lethality, PrimeKG knowledge graph, and Fisher pairwise co-occurrence.

Convergence Nodes
156
PrimeKG nodes directly linked to 3 or more of the 5 genes (60 to all 5); graph 129,262 nodes, 4,050,249 relationships
Genes Connected Complete
5/5
All patient genes found in PrimeKG knowledge graph
Pathway Redundancy Not supported
Not supported
GenePT-inspired: IDH2-SETBP1 only 6.6th percentile similarity; descriptive, no test
SL Pairs
4/10
Literature-curated: predicted or possible synthetic lethal pairs (SynLethDB not queried)

STRING Protein-Protein Interaction Network

The STRING v12 network 44
[44] D 2023
The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res (2023)
reveals strong functional connections between patient genes. DNMT3A and EZH2 share the highest combined score (0.999), consistent with their shared epigenetic function. Three patient-gene pairs have a STRING combined score ≥ 0.7: DNMT3A-EZH2 (with experimental evidence), and DNMT3A-IDH2 and EZH2-IDH2 (text-mining only, no experimental evidence). SETBP1 and PTPN11 connect through the broader myeloid gene network (via ASXL1, FLT3, and other myeloid regulators) rather than direct PPI evidence.
Gene AGene BCombined ScoreText-miningExperimentalIn Patient?
DNMT3AEZH20.9990.9980.519Both
DNMT3ADNMT3L0.9990.9840.986One
U2AF1SRSF20.9990.9970.726Neither
EZH2ASXL10.9980.9800.314One
FRS3PTPN110.9930.9830.270One
PTPN11MPIG6B0.9820.9230.292One
ASXL1TET20.9750.9750.000Neither
RUNX1PTPN110.9600.6110.000One
MPZL1PTPN110.9540.8800.632One
NRASPTPN110.9520.8920.153One
EZH2EPOP0.9470.6420.786One
DNMT3ATET20.9460.9260.292One
ZRSR2SRSF20.9430.7270.000Neither
PILRBPTPN110.9410.4400.000One
DNMT3AASXL10.9390.9360.091One
Table 1.Top 15 protein-protein interactions in the expanded network. Patient genes bolded.
Source: STRING v12 (Szklarczyk et al., 2023). Required score 700 (high confidence). 5 query genes + 20 additional interactors.

GenePT-inspired Functional Similarity

Pathway redundancy hypothesis: not supported (descriptive, no statistical test). If the five driver mutations were functionally redundant, we would expect high pairwise embedding similarity (consistent co-occurrence would reflect backup pathway activation). Instead, GenePT-inspired cosine similarities (all-MiniLM-L6-v2 on gene descriptions written in the script) show that most pairs cluster at low percentiles: IDH2-SETBP1 at only the 6.6th percentile, IDH2-PTPN11 at the 3.9th percentile, and DNMT3A-SETBP1 at the 2.3rd percentile. Only DNMT3A-EZH2 (98.6th percentile) and SETBP1-PTPN11 (95.9th) show high similarity. The mean similarity of co-occurring pairs (0.390, n = 4) exceeds the single exclusive pair (0.249, n = 1), but the dominant pattern is functional diversity, not redundancy.
Gene PairCosine SimilarityPercentileClassCo-occurrence
DNMT3A / EZH20.651198.6thHighunknown
SETBP1 / PTPN110.595895.9thHighCo-occurring
DNMT3A / IDH20.462377.4thHighCo-occurring
IDH2 / EZH20.434267.4thMediumunknown
PTPN11 / EZH20.343629.6thMediumunknown
SETBP1 / EZH20.328924.8thLowunknown
DNMT3A / PTPN110.276410.5thLowCo-occurring
IDH2 / SETBP10.24926.6thLowExclusive
IDH2 / PTPN110.22413.9thLowCo-occurring
DNMT3A / SETBP10.20402.3thLowneutral
Table 2.Pairwise functional similarity for all 10 patient gene pairs. Percentile ranks calculated against all 561 gene pairs from the 34-gene target panel.
Source: GenePT-inspired embeddings (Chen & Zou, 2024). Sentence-transformer model: all-MiniLM-L6-v2 (384 dimensions). Distribution: mean=0.398, SD=0.103.

Synthetic Lethality Analysis

4 of 10 gene pairs show predicted or possible synthetic lethal interactions (literature curation). The strongest signals: DNMT3A + EZH2 (dual epigenetic repression loss) and EZH2 + PTPN11 (PRC2 loss with RAS hyperactivation). IDH2 + SETBP1 is a low-to-medium evidence possible SL call; GENIE does not show significant mutual exclusivity for this pair (O/E = 0.905, Fisher p = 0.74). The O/E of 0.13 often quoted is for IDH1+SETBP1, not IDH2. SETBP1 (VAF 34%) is not the dominant clone (EZH2 59%, DNMT3A 39%). DNMT3A + IDH2 is explicitly NOT synthetic lethal; this is a well-established cooperative pair 12
[12] E 2016
Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med (2016)
.
Gene PairSL StatusEvidenceDirectionTherapeutic Angle
DNMT3A / EZH2Predicted SLMediumboth loss-of-function in patientEZH2 inhibitors (tazemetostat) may be selectively toxic to DNMT3A-mutant cells. However, in this patient EZH2 is already...
DNMT3A / IDH2Not SLHighDNMT3A loss + IDH2 gain cooperate (NOT SL)Enasidenib (IDH2 inhibitor) + azacitidine (targets remaining DNMT1) is the standard combination. Venetoclax adds BCL2 in...
DNMT3A / PTPN11Weak SLLowDNMT3A loss + PTPN11 gain (different pathways)SHP2 inhibitors (TNO155, RMC-4550) + azacitidine combination may exploit both vulnerabilities. No direct SL-based strate...
DNMT3A / SETBP1UnknownLowDNMT3A loss + SETBP1 gain (uncharacterized pair)No SL-based therapeutic strategy available for this pair.
EZH2 / IDH2Context-dependentMediumEZH2 loss + IDH2 gain (opposing epigenetic effects)Enasidenib (IDH2i) would remove the compensatory H3K27me3 maintenance, potentially lethal in EZH2-mutant cells. This cou...
EZH2 / PTPN11Predicted SLMediumEZH2 loss + PTPN11 gain (PRC2 + RAS crosstalk)MEK inhibitors (trametinib) or SHP2 inhibitors may be selectively effective in EZH2-mutant + PTPN11-mutant cells. SWI/SN...
EZH2 / SETBP1UnknownLowEZH2 loss + SETBP1 gain (no published data)No SL-based therapeutic strategy available for this pair.
IDH2 / PTPN11Not SLMediumIDH2 gain + PTPN11 gain (both activating, may cooperate)Combination of enasidenib (IDH2i) + SHP2 inhibitor targets both pathways. No SL-based rationale, but dual pathway inhibi...
IDH2 / SETBP1Possible SLLowIDH2 gain + SETBP1 gain (mutual exclusivity observed)The observed mutual exclusivity between IDH and SETBP1 is one of the strongest findings in the co-occurrence analysis. I...
PTPN11 / SETBP1Not SLMediumPTPN11 gain + SETBP1 gain (both drive proliferation)SHP2 inhibitors would target PTPN11 directly. PP2A activators (e.g., FTY720/fingolimod) could counteract SETBP1-mediated...
Table 3.Synthetic lethality assessment for all 10 pairwise combinations of patient genes.
Source: Literature curation with DepMap CRISPR screen data. SynLethDB 2.0 (synlethdb.sist.shanghaitech.edu.cn) was reachable but returned HTML, so no pair was queried there.

Pairwise Mutual Exclusivity (Fisher)

Across 18,625 myeloid samples, Fisher's exact test finds no significant mutual exclusivity for IDH2-SETBP1 (O/E = 0.905, p = 0.74, not significant). An in-house run of DISCOVER 30
[30] S 2016
A novel independence test for somatic alterations in cancer shows that biology drives mutual exclusivity but chance explains most co-occurrence. Genome Biol (2016)
is not shown as a result: its null model was miscalibrated (expected co-occurrence about 0.37 times the independence expectation), so its p-values and Z-scores are not interpretable. By Fisher O/E, the strongest co-occurrences are SETBP1-EZH2 (O/E = 4.96) and SETBP1-PTPN11 (O/E = 3.62).
Gene PairObservedFisher O/EFisher p-valueExpected (independence)Fisher direction
DNMT3A / IDH23802.7412.46e-87138.4Co-occurring
DNMT3A / SETBP1621.0600.618462.2Not significant
DNMT3A / PTPN111391.9531.60e-1576.1Co-occurring
DNMT3A / EZH21051.0980.2876101.8Not significant
IDH2 / SETBP1200.9050.740222.0Not significant
IDH2 / PTPN11411.4390.019926.9Co-occurring
IDH2 / EZH2591.5983.78e-436.0Co-occurring
SETBP1 / PTPN11413.6231.14e-1212.1Co-occurring
SETBP1 / EZH2754.9572.95e-3116.2Co-occurring
PTPN11 / EZH2542.9081.32e-1219.8Co-occurring
Table 4.Pairwise Fisher O/E and p-values with the expected co-occurrence count under independence. DISCOVER Z-scores omitted (miscalibrated null).
Source: GENIE v19.0 (18,625 samples). Fisher's exact test. Hypermutation threshold: 40 coding mutations.

PrimeKG Knowledge Graph

Graph Size
129,262 nodes
4,050,249 relationships across ten biological scales (Chandak 2023)
Convergence Nodes
156
Nodes directly linked to 3 or more of the 5 genes; 60 to all 5
Strongest Overlap
SETBP1-PTPN11
Overlap coefficient 0.787 (highest pairwise)
PrimeKG (Chandak et al., 2023) integrates 20 biomedical data sources into a knowledge graph with 129,262 nodes and 4,050,249 relationships. All 5 patient genes are represented. 156 nodes link directly to 3 or more of the 5 genes and 60 to all 5. Most are anatomy (tissue expression) nodes: 125 anatomy, 24 disease, 4 cellular component, 2 gene/protein, 1 molecular function. The pairwise neighborhood overlap ranges from Jaccard 0.09 (PTPN11-EZH2, lowest) to 0.45 (IDH2-SETBP1, highest). The overlap coefficient (normalized by smaller neighborhood) reveals that SETBP1-PTPN11 (0.787) and IDH2-SETBP1 (0.760) share the largest fraction of their neighborhoods.
Gene PairShared NeighborsJaccard SimilarityOverlap Coefficient
DNMT3A / IDH21140.27210.4302
DNMT3A / SETBP11210.37000.6612
DNMT3A / PTPN111140.16030.4302
DNMT3A / EZH21070.15830.4038
IDH2 / SETBP11390.44550.7596
IDH2 / PTPN111680.25450.6269
IDH2 / EZH2760.10700.2836
SETBP1 / PTPN111440.24040.7869
SETBP1 / EZH2810.13060.4426
PTPN11 / EZH2880.08890.1699
Table 5.Pairwise neighborhood overlap in PrimeKG for all 10 patient gene pairs.
Source: PrimeKG (Chandak, Huang, Zitnik. Scientific Data 2023). 129,262 nodes, 4,050,249 relationships across gene/protein, disease, pathway, molecular function, biological process, and anatomy node types.
References
  1. Szklarczyk D et al. The STRING database in 2023: protein-protein association networks. Nucleic Acids Res (2023). DOI
  2. Chen H, Zou J. GenePT: A Simple But Effective Foundation Model for Genes and Cells Built on ChatGPT. bioRxiv (2024). DOI
  3. Canisius S et al. A novel independence test for somatic alterations in cancer. Genome Biol (2016). DOI
  4. Chandak P, Huang K, Zitnik M. Building a knowledge graph to enable precision medicine. Sci Data (2023). DOI
  5. Papaemmanuil E et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med (2016). PubMed