Truace
Subscribe

Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing

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 mode…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1962· cite & verify
Machine learning uncovers previously undetectable genetic deletions in cancer panel sequencing
"Cytomegalovirus infection - Case 301" by Pulmonary Pathology is licensed under CC BY-SA 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/2.0/.

In brief

Researchers in the MASTERsg precision oncology program built a machine learning pipeline to detect focal gene deletions and assess zygosity in cancer panel sequencing data. They trained an Extreme Gradient Boosting model on PTEN-deleted and wild-type tumor samples, then compared its calls with the TSO500 pipeline and CNVkit.

The work matters because PTEN loss can make tumors sensitive to mTOR/Akt inhibition, so missed deletions may mean patients are not considered for effective targeted therapy. The evidence is limited to PTEN and a modest set of samples, so broader performance and clinical benefit remain unproven.

Main points

  1. Standard TSO500 copy number detection relies on bin count averaging, which can overlook focal events.
  2. The model was trained on PTEN-deleted and wild-type tumor samples and targeted the PTEN tumor suppressor gene.
  3. It identified focal deletions missed by both TSO500 and CNVkit.
  4. It found PTEN deletions in 33 of 204 profiled samples that TSO500 had classified as wild-type.
  5. Among 20 stainable cases, 15 (75%) showed absent or reduced PTEN protein by immunohistochemistry.

The gain

An XGBoost model flagged focal PTEN deletions in 33 of 204 profiled tumors that TSO500 had called wild-type, and most stainable cases tested showed reduced or absent PTEN protein.

The rundown

The study starts from a diagnostic gap: panel sequencing pipelines that average read counts across bins can miss small, focal deletions. Missing such a deletion in a tumor suppressor gene can keep patients from being matched to targeted therapy.

The team developed the pipeline within the MASTERsg precision oncology program and concentrated on PTEN, whose loss is linked to sensitivity to mTOR/Akt inhibition. The classifier output also covers zygosity evaluation, not just presence or absence of a deletion.

Validation used immunohistochemistry as an orthogonal check. Deletions the model called that TSO500 and CNVkit missed were associated with concordant PTEN protein loss, and 15 of 20 stainable cases among the newly detected deletions showed absent or reduced expression.

Sources

  1. Peer-reviewednpj Precision Oncology2026-10-16

The debate