Michael F. Berger

The paper develops and validates a clinicogenomic risk model to predict second primary cancers in cancer survivors using genetic, clinical, and lifestyle factors.

This paper evaluates generalist large language models (LLMs) in tumor genomic interpretation tasks, finding they can complement or outperform task-specific models in some cases.

The paper identifies MDF, a repair scar linked to POLQ-driven error-prone repair, as a marker of immune-responsive HRD pancreatic cancer with increased neoantigens and favorable clinical outcomes.

This study tracks clonal evolution of drug resistance in ovarian cancer using structural variants in cfDNA, finding polyclonal resistance with dominant high-fitness clones and genomic features linked to drug resistance.

The paper develops Mosaic, a machine learning system that classifies tumors into detailed subtypes and identifies molecular phenotypes using histopathologic and genomic data from a large patient cohort.

This study identifies TET2-mutant clonal hematopoiesis in thyroid cancer microenvironments and shows that TGFβ signaling from these macrophages mediates resistance to BRAF/MEK inhibition, with TGFβ inhibition restoring sensitivity.

This paper introduces MRD-EDGE, a machine learning platform for ultrasensitive ctDNA detection in plasma WGS, showing improved sensitivity and broader applicability for MRD monitoring.

The paper identifies how germline and somatic mutations interact to cause resistance to CDK4/6 inhibitors in breast cancer and proposes a sequential treatment strategy using PARP inhibitors.

This study presents a paired tumor:normal sequencing dataset for hematologic malignancies, showing improved identification of somatic variants and their clinical relevance.

MiMSI improves MSI detection in low tumor purity samples using deep learning and multiple instance learning, achieving higher sensitivity and auROC than existing methods.

The paper describes the development and clinical application of MSK-ACCESS, a liquid biopsy assay that improves the specificity of somatic variant detection in cell-free DNA by using paired normal sequencing.

The paper shows oncologic therapy increases clonal hematopoiesis with DDR mutations, which predicts risk of therapy-related myeloid neoplasms.

The paper measures DNA replication timing in mouse germline cells and shows it strongly correlates with mutation rates, recombination hotspots, and genomic features compared to somatic tissues.

This study evaluates the use of a molecular tumor board to interpret genomic profiling in pediatric oncology cases and provides therapeutic recommendations.