™11.5%
95% Wilson interval 9.0%–14.7%
™Veteran Owned & Operated!
End-to-end cancer use case / prostate cancer
A bounded question.
An honest answer.
Using rights-cleared public data, is TP53 alteration more prevalent in a metastatic or castration-resistant prostate-cancer cohort than in a primary prostate-adenocarcinoma cohort, and are the selected cohorts sufficiently comparable for the observed difference to justify a clearly defined next research investigation?
Direct answer
Higher observed prevalence.
Not fairly comparable.
The observed aggregate prevalence is higher in the metastatic/castration-resistant cohort, but the cohorts are not fairly comparable for a causal, prognostic, treatment-response, or clinical claim.
What this result refuses to say
Refuse any claim that TP53 alteration increased because of progression, predicts treatment response, establishes mechanism, or has clinical utility from this comparison.
Pinned aggregate result
One endpoint. Two exact denominators.
Each sample counts once if the pinned cBioPortal API response contains at least one TP53 mutation row. The denominator is the corresponding sequenced sample list.
36.7%
95% Wilson interval 32.4%–41.3%
Absolute difference25.2%95% interval 17.7%–32.3%
Risk ratio3.18descriptive, unadjusted
Odds ratio4.45descriptive, unadjusted
Fisher exact6.65e-20two-sided; not a comparability test
Versioned open experiment specification
Defined before calculation.
canxcer.prostate.tp53-prevalence/1.0.0
Using rights-cleared public data, is TP53 alteration more prevalent in a metastatic or castration-resistant prostate-cancer cohort than in a primary prostate-adenocarcinoma cohort, and are the selected cohorts sufficiently comparable for the observed difference to justify a clearly defined next research investigation?
The API-reported prevalence of samples with at least one TP53 mutation is higher in the metastatic/castration-resistant SU2C cohort than in the primary TCGA cohort.
Proportion of samples in each study's sequenced sample list with at least one TP53 (Entrez Gene 7157) mutation returned by the pinned cBioPortal molecular-profile endpoint.
At least one mutation row returned for TP53 for a sample in the study's sequenced sample list; multiple rows for one sample count once.
All sample identifiers in the pinned sequenced sample list. Samples without a returned TP53 mutation row are counted as not observed altered, not proven biologically wild type. Study-level API responses do not harmonize tumor purity, biopsy site, prior therapy, sequencing pipeline, demographic representation, or sample-level missingness. Those gaps prevent fair causal comparison.
Any unpinned source response, ambiguous denominator, rights uncertainty, or calculation mismatch fails the run closed.
Research comparison engine
Statistical separation does not create fair cohorts.
The comparison engine records what aligns, what fails, and what is absent. A small p-value answers neither causality nor cohort exchangeability.
Disease family and genealigned
Both cohorts concern prostate adenocarcinoma and use TP53 / Entrez 7157.
Disease statefailed
Primary untreated disease and metastatic/castration-resistant disease are intentionally different and are not exchangeable populations.
Assay and processinglimited
Both expose mutation profiles, but study-era pipelines, capture, calling, filtering, and portal processing are not harmonized.
Sample contextmissing
Biopsy site, tumor purity, prior therapy, and patient/sample multiplicity are not harmonized in this aggregate analysis.
Representationmissing
Demographic and ancestry distributions were not available in the aggregate endpoint and cannot be assumed comparable.
Independencelimited
The studies are separate cohorts but use shared portal and curation infrastructure; this is not independent replication.
Sources, versions, overlap, and rights
Every public input is attributable and pinned.
Only aggregate derived values are published here. No sample identifiers or raw mutation rows are stored, displayed, or included in the open package.
Prostate Adenocarcinoma (TCGA, PanCancer Atlas)
- Study
prad_tcga_pan_can_atlas_2018- Profile
prad_tcga_pan_can_atlas_2018_mutations- Sample list
prad_tcga_pan_can_atlas_2018_sequenced- API SHA-256
ecc037fdd5b778f54833c84acebcfa3b80ab1b03ceef4c457bb228f582fc8541- Sample-list SHA-256
b1eb86bad77307c5ffecf11e498a953e9d6dcc6dea05e6fa505659360a90cf96
Aggregate derived result only. Attribute TCGA and cBioPortal; do not redistribute raw or sample-level material. The study license warns that pipeline data may be preliminary, unvalidated, and subject to change.
Metastatic Prostate Adenocarcinoma (SU2C/PCF Dream Team, PNAS 2019)
- Study
prad_su2c_2019- Profile
prad_su2c_2019_mutations- Sample list
prad_su2c_2019_sequenced- API SHA-256
1b4f5307edfdf055cc96d3cb841c4def8e9eb1b500fe40501a6bfbb14cafdc0e- Sample-list SHA-256
736a07b0abf17dfd2863a93e742e586289c0fb563e9707aa1141d4d160ec9bbf
The DataHub study is licensed under ODC ODbL. Aggregate output retains source attribution and share-alike notice; no sample identifiers or raw rows are republished.
Portal snapshot
cBioPortal v7.1.2; database 3.0.0; gene table hgnc_v7_2025.10.7; commit bbba61b2a45b3719187a310dcadaae75b9bef067.
The SU2C raw DataHub file and public cBioPortal API produce different TP53 row/sample counts. This analysis deliberately uses the pinned, processed cBioPortal API responses and treats that discrepancy as a material provenance constraint.
The cohorts are distinct studies, but both are distributed through cBioPortal/DataHub and can share upstream curation infrastructure. They are not described as independent confirmation of one another.
Human decision / wet-lab boundary
Harmonized TP53 prevalence analysis across disease state
The governed decision is to design a harmonized reanalysis—not to claim validation. The prediction ledger is pending unmeasured.
- Pre-register inclusion, one-sample-per-patient, alteration, and missingness rules.
- Obtain rights-cleared sample-level covariates for stage, biopsy site, prior treatment, assay, ancestry, and tumor purity.
- Reprocess or restrict data to a common callable region and mutation-calling contract.
- Estimate adjusted prevalence differences with sensitivity analyses and an independent validation cohort.
- Require qualified human scientific review before any wet-lab, predictive, or clinical interpretation.