RBM10 + CDKN1B
RBM10 and CDKN1B co-occurrence in bladder cancer (MSK-IMPACT 50K)
DataMSK-IMPACT 50K (2026) · Bladder cancer · 1,917 tumours · both altered in 9 (1.4 expected if independent)
Novel lead. Significant, with a real effect in a consistent direction, and it still holds when tumours are compared only within the same tumour type (and within the same molecular subtype, wherever this cancer's subtypes are known in the data). No artifact check failed.
On the map: An early signal. Neither its statistical strength nor its expected value is high enough yet to count as a discovery.
Called novel: a PubTator3 search finds no title or abstract naming both RBM10 and CDKN1B in bladder cancer. A relationship reported only in a paper's full text would not appear in this search. See the PubTator3 search.
Signal (x) is the corrected significance on the map's scale, discounted from 7% to 3% because the cohort-composition confound is not yet controlled. Efficacy (y) is credibility times actionability.
These genes are altered together more often than chance would predict.
Bladder
This finding's tumour type arises in the highlighted organ.
No structure profile is curated for this gene.
A human gene in this finding. Not in the curated driver set, so no structured profile is shown.
RBM10 mutations in bladder cancer →RBM10 co-mutations by tumour type →Other findings with RBM10 →No structure profile is curated for this gene.
A human gene in this finding. Not in the curated driver set, so no structured profile is shown.
Other findings with CDKN1B →Anatomical figure from the EBI Expression Atlas anatomogram, and predicted structures from the AlphaFold Protein Structure Database (both CC-BY-4.0). A model shows the protein's healthy fold; a highlighted residue marks a recurrent cancer position, not a prediction of a specific mutation's effect.
The raw 2×2 the FDR q and log₂ odds ratio above were computed from. It is the immutable evidence the score re-derives from: every number on this page traces back to these four counts.
| Gene state | RBM10 altered | RBM10 wild-type |
|---|---|---|
| CDKN1B altered | 9 | 14 |
| CDKN1B wild-type | 111 | 1,783 |
both 9 (expected 1.4 if independent) · CDKN1B-only 14 · RBM10-only 111 · neither 1,783
Weak evidence. Below the α=0.05 reject threshold (e≥20). Unlike a single-look p, this stays valid under the engine's continuous re-testing.
Molecular-subtype certification is unresolved. Recorded lineage only. Within-lineage molecular subtype remains unresolved. The association remains visible, but does not qualify as a subtype-controlled discovery.
The two axes are variables: signal (x) and efficacy (y = credibility × actionability). A point high on both lands in the frontier (top-right). A strong signal with low credibility sits low and to the right: statistically strong, not worth pursuing.
A language model wrote this from the evidence listed below. It is annotation: the engine's verdict above comes only from the measured statistics, never from this text.
The RBM10–CDKN1B co-occurrence across bladder-cancer patients is most likely driven by residual molecular-subtype composition: a subtype enriched for alterations in both genes could produce the population-level association even after lineage and overall alteration burden are accounted for, without implying that the alterations cooperate within the same tumour.
- 1The association is a substantial but not definitive population-level signal: the within-stratum log2 odds ratio is 3.4 (95% CI 2.1–4.6), with 9 double-altered tumours, and it survives FDR correction at q=0.039. Burden-aware testing reduces concern that it merely reflects generally mutation-heavy tumours.E1E19
- 2Residual molecular-subtype composition is the leading explanation because lineage was recorded, but molecular subtype remains unresolved; the within-lineage analysis does not rule out subtype confounding. The Cybergenic corpus directly converges with this interpretation for the same pair and cancer (E16). Its other findings provide only contextual, not pair-specific, support: E15 and E18 also favor subtype composition for different RBM10 pairs, while E17 concerns a different cancer and depletion pattern.E2E16E15E18E17
- 3Ranked explanations: (1) residual molecular-subtype composition; (2) a cohort-specific signal or winner's-curse/data artifact; (3) a shared upstream driver; (4) deterministic co-amplification; (5) genuine within-tumour cooperation. The lack of replication across independent patient sources materially elevates the artifact explanation. The evidence supplies no specific shared driver or amplicon evidence, and co-occurrence across patients does not establish within-tumour cooperation.E2E3E1
- 4The provided paper abstracts do not establish a relationship between RBM10 and CDKN1B: the bladder-cancer abstract concerns FGFR3 and a urothelial carcinoma variant, while the other highlighted abstract concerns GNAQ/GNA11 in metastatic solid tumours. They therefore do not support a pair-specific mechanism; the absence of a curated network link is consistent with this being an under-explored candidate.E9E10E11E4
Test the leading subtype-composition hypothesis in an independent bladder-cancer cohort with molecular-subtype annotations: estimate the RBM10–CDKN1B association within subtypes and in a subtype-adjusted model, and check whether the association attenuates. To earn a within-tumour cooperation claim, separately use clonal or single-cell DNA profiling to establish co-alteration in the same tumour cells, followed by a functional co-dependency test.
This is a conservative, testable explanation rather than a finding. The strongest deflating explanation is cohort specificity or winner's curse: significance occurs in only one independent patient source, despite surviving FDR correction. Subtype composition is unresolved, not demonstrated. No cited Europe PMC paper actually studies the RBM10–CDKN1B pair, so a within-tumour mechanism is unearned; that lack of pair-specific evidence strengthens the under-explored interpretation but lowers confidence.
Weakest link: Residual molecular-subtype confounding is only a plausible, untested explanation: subtype was not measured or adjusted for, and the association rests on nine double-altered tumors in one non-replicated cohort.
Significant (after FDR correction) in one independent source of this tumour type so far. The engine records where a pattern is significant, not every source that could test it, so this calls for caution: a pattern selected from many tests tends to overstate its effect. It does not count against the pattern.
- E1In Bladder cancer (MSK-IMPACT 50K 2026), RBM10 and CDKN1B are co-occurring: burden-aware DISCOVER p = 8.0e-5 (a test that accounts for each tumor's overall alteration burden, so co-occurrence is not inflated by unstable tumors carrying more of everything), Benjamini-Hochberg FDR q = 3.9e-2, within-stratum (Mantel-Haenszel) log2 odds ratio 3.4 (95% CI 2.1 to 4.6), 9 tumors carry both alterations.cBioPortal / MSK-IMPACT 50K (2026)
- E2The co-occurrence was tested at adjusted FDR q=3.9e-2 with 100% annotation coverage. Recorded lineage only. Within-lineage molecular subtype remains unresolved. The measured heterogeneity test did not reject a common effect; this does not prove homogeneity. This control alone establishes neither a mechanism nor novelty.cBioPortal / MSK-IMPACT 50K (2026), scoped DISCOVER control
- E3This co-occurrence is FDR-significant in only 1 independent patient source so far and does NOT recur across sources, so it could be cohort-specific biology or a winner's-curse artifact of scanning many cohorts. Treat confidence accordingly.cBioPortal (cross-cohort)
- E4NOVELTY: no curated network (STRING functional association, Reactome pathway, or OmniPath causal edge) already links RBM10 and CDKN1B. A statistically strong co-alteration with no existing network edge is a genuinely overlooked candidate, not a known interaction.OmniPath / Reactome / STRING
- E5RBM10 expression is higher in tumor than matched GTEx+adjacent normal tissue: log2 fold change 0.26 (29.2 vs 24.4 TPM, n=407 tumor / 28 normal).UCSC Xena (tumor vs GTEx normal)
- E6RBM10 strongest disease associations (overall score 0–1): TARP syndrome 0.79; lung adenocarcinoma 0.63; neurodegenerative disease 0.53; pancreatic adenocarcinoma 0.51.Open Targets
- E7No approved drug directly targets RBM10 (a tractability gap).Open Targets
- E8Tractability for RBM10: Database Ubiquitination, Half-life Data.Open Targets
- E928 paper(s) co-mention this: very sparse. Against each gene's own cancer literature, ~3 co-mentions would be expected by chance, so this pair is already studied more than its parts predict. Most relevant: "Genomic Landscape of GNAQ and GNA11 Mutations in Metastatic Solid Tumors: A Real-World Da…" (2026).Europe PMC
- E10"Large, Nested Variant of Urothelial Carcinoma Is Enriched with Activating Mutations in Fibroblast Growth Factor Recepto…" (2023): The large, nested variant of urothelial carcinoma (LNVUC) is characterized by bland histomorphology mimicking that of benign von Brunn nests. In the current study, we aimed to investigate the Fibroblast Growth Factor Receptor-3 (FGFR-3) activation and missense mutation in 38 cases, including 6 cases diagnosed with LNVUC and 32 with metastatic invasiv…Europe PMC (paper)
- E11"Genomic Landscape of GNAQ and GNA11 Mutations in Metastatic Solid Tumors: A Real-World Data Analysis." (2026): Purpose: Activating mutations in GNAQ and GNA11 are recognized oncogenic drivers in uveal melanoma (UM); however, their prevalence across various cancers and their genomic context remain inadequately characterized. This study aimed to elucidate the genomic landscape of these mutations across a range of metastatic solid tumors. Materi…Europe PMC (paper)
- E120 clinical trial(s) match RBM10/CDKN1B in Bladder cancer, 0 currently open.ClinicalTrials.gov
- E13RBM10: tumor suppressor, driver in 6 cancer types (intOGen 2024.09.20)intOGen (driver catalogue)
- E14CDKN1B: tumor suppressor, driver in 5 cancer types (intOGen 2024.09.20)intOGen (driver catalogue)
- E15Prior finding (EGFR + RBM10 in Non-small cell lung cancer): verdict strengthened, confidence 53%. “The most likely explanation is residual molecular-subtype composition within NSCLC: EGFR- and RBM10-altered t…”Cybergenic corpus
- E16Prior finding (RBM10 + CDKN1B in Bladder cancer): verdict inconclusive, confidence 48%. “The RBM10–CDKN1B co-occurrence in bladder cancer is most likely driven by residual molecular-subtype or linea…”Cybergenic corpus
- E17Prior finding (TP53 ⊥ RBM10 in Pancreatic cancer): verdict inconclusive, confidence 48%. “The TP53–RBM10 depletion in pancreatic cancer is more likely to reflect molecular-subtype composition or subt…”Cybergenic corpus
- E18Prior finding (NRAS + RBM10 in Thyroid cancer): verdict strengthened, confidence 48%. “The NRAS–RBM10 co-occurrence across thyroid cancer patients is most likely driven by residual molecular-subty…”Cybergenic corpus
- E19Significance is FDR-corrected: Benjamini-Hochberg q = 3.9e-2 across 9870 gene pairs tested in this cohort, so this pair SURVIVES multiple-testing correction (q ≤ 0.05). The q, not the nominal p, is the honest signal strength.Cybergenic method
Model-written annotation, not a finding. The statistics it cites are real and are shown above; the argument connecting them is a hypothesis that needs wet-lab or clinical validation before it means anything.
- Permalink
- https://cybergenic.im/findings/662b34e3-f982-4160-beaa-396311e7aff2
- First seen
- 29 Sep 2026
- Prediction registry
- Locked 29 Sep 2026, row hash
684ffae6cff318bc08b70c2b76820bdc(not testable out of sample). How the registry works
@misc{cybergenic_662b34e3,
author = {{Cybergenic}},
title = {{RBM10 and CDKN1B co-occurrence in bladder cancer (MSK-IMPACT 50K)}},
year = {2026},
howpublished = {\url{https://cybergenic.im/findings/662b34e3-f982-4160-beaa-396311e7aff2}},
note = {Research engine finding, first seen 2026-09-29; prediction locked 2026-09-29, registry row hash 684ffae6cff318bc08b70c2b76820bdc},
month = sep
}