Kyungsik Ha

COOBoostR uses extreme gradient boosting to predict tumor tissue or cell-of-origin using chromatin marks and mutation density data, outperforming random forest methods in speed and accuracy.

The paper uses whole-genome sequencing to predict cell-of-origin for primary liver cancers, finding heterogeneity in origin predictions across cancer types.

The paper predicts cancer cell-of-origin from DNA sequences by matching mutational landscapes to chromatin modification patterns in normal tissues.

The paper identifies the critical time point during cancer progression when chromatin features influence the mutation landscape, showing that epigenetic changes in early stages significantly impact somatic mutations.