Mark D.M. Leiserson

This paper introduces mutation-level covariate models to analyze mutational signatures and shows they better capture replication strand specificity than standard models.

The paper identifies functional pathways associated with mutational signatures in breast cancer, revealing differences in the etiology of APOBEC and clock-like signatures.

This paper introduces sequence-dependent models for mutation signatures, improving the resolution of mutation signature activity in cancer by accounting for dependencies between mutations.

This paper introduces a method to transfer knowledge across species by integrating sequence and network structure to create functional embeddings of proteins, showing improved cross-species functional similarity and phenolog prediction.