More Than Half of DNMT3A-Mutant Patients in Pooled Public Cancer Genomic Data Carry No Tumor-Derived Mutation: Clonal Hematopoiesis Introduced by Patient-Keyed Aggregation
- 1Stanford University School of Medicine
Abstract
Large-scale reuse of public cancer genomic data begins by pooling studies into one patient-by-gene matrix, keyed on the patient. Records held under one patient identifier need not come from the same kind of specimen: three clonal-hematopoiesis studies in cBioPortal deposit blood normals sharing patient identifiers with tumors deposited elsewhere, so a patient-keyed merge records a hematopoietic clone as a tumor mutation. We linked 1, 282, 483 coding calls in 140 genes from 483 studies to patients, classified each by specimen type, and rebuilt the matrix with and without blood-derived calls. Blood normals supply 1.7% of calls but concentrate: 52.8% of patients flagged DNMT3A-mutant, 21.3% of TET2 and 19.7% of CHEK2 have no tumor-derived call in that gene, against 0.4% for TP53. Those calls carry a hematopoietic gene spectrum and, in TP53, hematopoietic allele fractions. The artifact reaches 65 further cohorts, almost all solid-tumor studies, exceeding one patient in ten in fifteen of them, and tracks the share of patients with a paired blood normal, not the disease (r = 0.994, 45 cancer types). Correction lowers the pan-cancer DNMT3A rate by 45% and moves survival estimates further than chance removal of the same patients would. Two filter lines applied before aggregation remove it.
Related
- Choline Kinase Alpha Inhibition is a Mutation Agnostic Therapeutic Vulnerability in Clonal Hematopoiesis
- Clonal hematopoiesis and cell-free DNA characteristics in patients with high-risk large B-cell lymphomas
- Context‐Dependent TP53 Mutation Gain‐of‐Function Is Correlated With Its Contribution to Genomic Instability
- Distinct genomic landscapes of mismatch-repair deficiency (dMMR)/microsatellite instability-high (MSI-H) gastrointestinal (GI) cancer depending on tumor mutation burden (TMB) levels.
- FFT-GA-BC: a fused federated transformer genetic algorithm framework with blockchain-secured, differentially private, homomorphically encrypted aggregation for mutation, phenotype, and binary-allele analysis.
BibTeX
@misc{pan2026more,
title = {{More Than Half of DNMT3A-Mutant Patients in Pooled Public Cancer Genomic Data Carry No Tumor-Derived Mutation: Clonal Hematopoiesis Introduced by Patient-Keyed Aggregation}},
author = {M. Pan and L. Zhi and J. Chi and A. Yang Leow},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.10.26365222},
url = {https://www.medrxiv.org/content/10.64898/2026.10.10.26365222v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - More Than Half of DNMT3A-Mutant Patients in Pooled Public Cancer Genomic Data Carry No Tumor-Derived Mutation: Clonal Hematopoiesis Introduced by Patient-Keyed Aggregation
AU - M. Pan
AU - L. Zhi
AU - J. Chi
AU - A. Yang Leow
PY - 2026
DA - 2026/10/10
PB - medRxiv
DO - 10.64898/2026.10.10.26365222
UR - https://www.medrxiv.org/content/10.64898/2026.10.10.26365222v1
AB - Large-scale reuse of public cancer genomic data begins by pooling studies into one patient-by-gene matrix, keyed on the patient. Records held under one patient identifier need not come from the same kind of specimen: three clonal-hematopoiesis studies in cBioPortal deposit blood normals sharing patient identifiers with tumors deposited elsewhere, so a patient-keyed merge records a hematopoietic clone as a tumor mutation. We linked 1, 282, 483 coding calls in 140 genes from 483 studies to patients, classified each by specimen type, and rebuilt the matrix with and without blood-derived calls. Blood normals supply 1.7% of calls but concentrate: 52.8% of patients flagged DNMT3A-mutant, 21.3% of TET2 and 19.7% of CHEK2 have no tumor-derived call in that gene, against 0.4% for TP53. Those calls carry a hematopoietic gene spectrum and, in TP53, hematopoietic allele fractions. The artifact reaches 65 further cohorts, almost all solid-tumor studies, exceeding one patient in ten in fifteen of them, and tracks the share of patients with a paired blood normal, not the disease (r = 0.994, 45 cancer types). Correction lowers the pan-cancer DNMT3A rate by 45% and moves survival estimates further than chance removal of the same patients would. Two filter lines applied before aggregation remove it.
ER -