Thomas Bauernhofer

PROMIS detects microsatellite instability in tumor sequencing without reference data or matched normals, achieving high accuracy across multiple cancer types and sample types.

The paper develops a novel CoDuCo in situ hybridization method combined with machine learning to classify and characterize circulating tumor cells in prostate cancer, identifying resistance mechanisms and therapeutic targets.

This study infers tumor cell-specific transcription factor binding from cell-free DNA to predict tumor subtypes and detect cancer early.