ABSTRACT In the era of genome medicine, we are able to precisely identify the molecular susceptibilities of a range of human pathologies, including cancer. However, many of the bona fide drivers of cancer—transcription factors, tumor suppressors, and chromatin remodelers (such as p53, myc, and SWI/SNF) cannot be readily targeted by traditional small-molecule active-site inhibitors, as their functions are modulated by protein interactions. Indeed, protein-protein interactions constitute nearly 90% of all medicinal targets of interest, yet peptides inhibitors – which effectively target these interactions – account for only 2% of FDA-approved drugs. Peptide therapies face major challenges including costly synthesis, in vivo instability from protease degradation, and poor bioavailability. To remedy these issues, “stapled-peptides” have been proposed to improve both the potency and pharmacokinetics of such therapies. Unfortunately, these stapled peptides— which contain non-natural amino acids to covalently maintain a helical structure— cannot be genomically encoded because their production requires additional chemical steps, which drastically limits the ability to discover and synthesize new biomimetic peptide therapies and tools. Therefore, the ability to iteratively design, genomically encode, and reliably synthesize a stable class of these molecules in vivo would yield novel chemical probes for a variety of protein-protein interactions in cancer. This proposal seeks to genomically-encode the production of therapeutically relevant, cell-permeable stapled peptides in a bacterial organism. This would allow for the generation of screenable peptide-libraries, drastically reduce the cost of synthesis, and ultimately provide a discovery platform for an entirely new class of protein- protein inhibitors. Utilizing high-throughput, robotic phage-assisted continuous directed evolution (roboPACE), an in vivo mechanism to produce cell-permeable bio-mimetic peptides will be developed. First, a novel thio- ether stapling mechanism will be characterized in vitro utilizing a novel non-canonical amino acid [Aim 1]. Second, efficient in vivo incorporation of this amino acid into proteins will be evolved in high-throughput with roboPACE [Aim 2]. Finally, a promiscuous bacterial synthetase enzyme, will be evolved to efficiently catalyze the stapling mechanism in order to genomically-encode stapled-peptide production [Aim 3]. Collectively, this proposal will extend the breadth and throughput of ncAA design and incorporation, and ultimately develop an in vivo peptide-stapling mechanism in order to treat and characterize presently “undruggable” therapeutic targets in cancer.