Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 1 h ago
Peer-reviewedCC BYNPJ precision oncology

Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing.

  • Irina A Kerle1,2,3
  • Oleksandr Husak4,5,6
  • Thomas Gross7
  • Andreas Hartig8
  • Rebecca Prause7
  • Daniela Richter4,5,9
  • Sascha Brückmann8
  • Maximilian Werner4,5,9,10
  • Anja Kögler7
  • Dorothea Hanf4,5,9,10
  • Lino Möhrmann4,5,9,11,12
  • Doreen William13,14,15
7 more authors
  • Karsten Wendt6
  • Nicole Pfarr16
  • Carolin Mogler16
  • Daniela E Aust8
  • Evelin Schröck13,14,15
  • Christoph Heining4,5,9
  • Hanno Glimm4,5,9
16 affiliations
  1. 1Department for Translational Medical Oncology, National Center for Tumor Diseases Dresden (NCT/UCC), a partnership between German Cancer Research Center (DKFZ), Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology and Helmholtz-Zentrum Dresden - Rossendorf (HZDR), Dresden, Germany. [email protected].
  2. 2Translational Medical Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany. [email protected].
  3. 3German Cancer Consortium (DKTK), partner site Dresden, a partnership between DKFZ and University Hospital Carl Gustav Carus Dresden, Dresden, Germany. [email protected].
  4. 4Department for Translational Medical Oncology, National Center for Tumor Diseases Dresden (NCT/UCC), a partnership between German Cancer Research Center (DKFZ), Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology and Helmholtz-Zentrum Dresden - Rossendorf (HZDR), Dresden, Germany.
  5. 5Translational Medical Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany.
  6. 6Chair of Software Technology, Faculty of Computer Science, Technical University Dresden, Dresden, Germany.
  7. 7Core Unit for Molecular Tumor Diagnostics (CMTD), National Center for Tumor Diseases Dresden (NCT/UCC), a partnership between German Cancer Research Center (DKFZ), Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology and Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Germany.
  8. 8Institute of Pathology, Carl Gustav Carus University Hospital, Dresden, Germany.
  9. 9German Cancer Consortium (DKTK), partner site Dresden, a partnership between DKFZ and University Hospital Carl Gustav Carus Dresden, Dresden, Germany.
  10. 10Department of Internal Medicine I, University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany.
  11. 11Computational Health Informatics Program, Boston Children's Hospital, Harvard Medical School, Boston, USA.
  12. 12Internal Medicine C, Hematology, Oncology, Stem Cell Transplantation and Palliative Care, University Medicine Greifswald, Greifswald, Germany.
  13. 13Institute for Clinical Genetics, University Hospital Carl Gustav Carus at TU Dresden, Dresden, Germany.
  14. 14ERN GEN-TURIS, Hereditary Cancer Syndrome Center Dresden, Dresden, Germany.
  15. 15National Center for Tumor Diseases (NCT), NCT/UCC Dresden, a partnership between DKFZ, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, and Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Germany; German Cancer Consortium (DKTK), partner site: Dresden, and German Cancer Research Center (DKFZ), Heidelberg, Germany.
  16. 16Institute of Pathology, School of Medicine and Health, Technical University Munich, Munich, Germany.
Read full textPDFDOICite

Abstract

Missed genetic deletions can prevent cancer patients from receiving effective targeted therapy. Standard pipelines, such as TSO500, use bin count averaging for copy number variation detection but often overlook focal events. To address this, we developed a machine learning (ML) pipeline for deletion detection and zygosity evaluation within the MASTERsg precision oncology program. We focused on the tumor suppressor gene PTEN, whose loss confers sensitivity to mTOR/Akt inhibition. An Extreme Gradient Boosting model trained on PTEN-deleted and wild-type tumor samples identified focal deletions missed by the TSO500 pipeline and CNVkit that were associated with concordant PTEN protein loss by immunohistochemistry. Importantly, the model detected PTEN deletions in 33/204 profiled samples that were classified as PTEN wild-type by TSO500; 75% (15/20) of stainable cases showed absent or reduced PTEN protein expression. These findings demonstrate the potential of ML to enhance panel sequencing, uncovering therapeutically relevant alterations missed by established diagnostic methods.

Related

  1. Deep-learning single-cell profiling uncovers interbacterial and drug-driven reshaping of pathogen virulence.
    Bence Bognár, Terézia Kovács, Réka Muresan +10
    Nature communicationsPeer-reviewedCC BYFull textEndocrinology & metabolismEurope PMCPMC
  2. Shall we repeat? Predicting the necessity to repeat blood testing in pediatric emergency departments of laboratory tests previously performed in the community using a machine learning model
    Adi Shuchami, Teddy Lazebnik, Shai Ashkenazi +3
    PLOS OnePeer-reviewedCC BY 4.0Full textPhysiologyPLOS
  3. PO129 | CLINICAL BENEFIT OF RECOMBINANT ADAMTS13 IN CTTP ASSOCIATED WITH HOMOZYGOUS PD173G MUTATION AND UNDETECTABLE ACTIVITY
    Raimondo De Cristofaro
    Bleeding, Thrombosis and Vascular BiologyPeer-reviewedCC BY-NCFull textCardiovascular medicineDOAJ
  4. Integrated Analysis of Serotonin-Related Biomarkers in Thoracic Aortic Aneurysm Based on Machine Learning and Single-Cell Sequencing Data
    BiomedicinesPeer-reviewedCC BYFull textGeneral medicineEurope PMC
  5. Derivation of a blood count-based machine learning model to prioritise sequencing for high-risk clonal haematopoiesis: a retrospective cohort study
    Riya Nandi, Kun Zhao, Yash Pershad +11
    EClinicalMedicinePeer-reviewedCC BYFull textGeneral medicineEurope PMC

Cite

BibTeXRIS
BibTeX
@article{kerle2026machine,
  title = {{Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing.}},
  author = {Irina A Kerle and Oleksandr Husak and Thomas Gross and Andreas Hartig and Rebecca Prause and Daniela Richter and Sascha Brückmann and Maximilian Werner and Anja Kögler and Dorothea Hanf and Lino Möhrmann and Doreen William and Karsten Wendt and Nicole Pfarr and Carolin Mogler and Daniela E Aust and Evelin Schröck and Christoph Heining and Hanno Glimm},
  journal = {NPJ precision oncology},
  year = {2026},
  month = oct,
  doi = {10.1038/s41698-026-01756-9},
  url = {https://europepmc.org/article/MED/42888201},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing.
AU  - Irina A Kerle
AU  - Oleksandr Husak
AU  - Thomas Gross
AU  - Andreas Hartig
AU  - Rebecca Prause
AU  - Daniela Richter
AU  - Sascha Brückmann
AU  - Maximilian Werner
AU  - Anja Kögler
AU  - Dorothea Hanf
AU  - Lino Möhrmann
AU  - Doreen William
AU  - Karsten Wendt
AU  - Nicole Pfarr
AU  - Carolin Mogler
AU  - Daniela E Aust
AU  - Evelin Schröck
AU  - Christoph Heining
AU  - Hanno Glimm
PY  - 2026
DA  - 2026/10/16
JO  - NPJ precision oncology
DO  - 10.1038/s41698-026-01756-9
UR  - https://europepmc.org/article/MED/42888201
AB  - Missed genetic deletions can prevent cancer patients from receiving effective targeted therapy. Standard pipelines, such as TSO500, use bin count averaging for copy number variation detection but often overlook focal events. To address this, we developed a machine learning (ML) pipeline for deletion detection and zygosity evaluation within the MASTERsg precision oncology program. We focused on the tumor suppressor gene PTEN, whose loss confers sensitivity to mTOR/Akt inhibition. An Extreme Gradient Boosting model trained on PTEN-deleted and wild-type tumor samples identified focal deletions missed by the TSO500 pipeline and CNVkit that were associated with concordant PTEN protein loss by immunohistochemistry. Importantly, the model detected PTEN deletions in 33/204 profiled samples that were classified as PTEN wild-type by TSO500; 75% (15/20) of stainable cases showed absent or reduced PTEN protein expression. These findings demonstrate the potential of ML to enhance panel sequencing, uncovering therapeutically relevant alterations missed by established diagnostic methods.
SN  - 2397-768X
ER  -