Shenlin Hsiao

Abstract Advancing the performance of programmable genome editing nucleases remains a key challenge in expanding their research and therapeutic applications. Here, we introduce a scalable deep learning–guided protein engineering framework for improving nuclease activity without requiring experimental training data.

1 Introduction Base editors (BEs), which typically contain a catalytically impaired Cas protein fused to a DNA-modifying enzyme, can mediate base editing in genomic DNA without DNA double breaks (DSBs), thus safer than nonhomologous ending joining (NHEJ) and homology-directed repair (HDR) of Cas nuclease-based method.[1, 2] The Cas domains were derived from the clustered regularly interspaced palindromic repeats (CRISPR/Cas) systems, such as Cas9 and Cas12, which is in combination with a...