Preprint: deep learning of mutagenesis in metastases

Hanli Jiang, Alec Bahcheli, Kevin Cheng and Jüri Reimand posted a preprint, “Chromatin accessibility of primary cancers informs regional mutagenesis in metastases through multi-scale deep learning”, on bioRxiv. The study introduces CAMM, a multi-scale deep learning model that predicts regional mutation density in metastatic cancer genomes from chromatin accessibility and replication timing, and finds that the epigenomes of the tissue of origin remain the strongest predictors of mutation patterns in metastases. Genomic regions with more mutations than expected are enriched in known cancer genes.

• Preprint on bioRxiv
• CAMM software