High-Throughput Protein Perturbation Screens with AI-Designed Degraders

The development of specific protein binders is crucial for biologics and targeted protein degradation (TPD) therapies. However, current techniques rely on low-throughput, labor-intensive approaches and often use non-human display systems such as phage, yeast, or mRNA display, limiting their translational relevance. To address this, we developed a high-throughput, human cell-based binder screening platform that enables the functional evaluation of artificial intelligence (AI)-designed peptide binders in a mammalian context for protein perturbation. Our approach utilizes genetically engineered, doxycycline-inducible ubiquibodies (uAbs) that fuse libraries of computationally designed “guide” peptides to an E3 ubiquitin ligase domain, enabling modular, CRISPR-like TPD. Using target-expressing cells fused with an mCherry reporter to monitor protein degradation, we screened and validated AI-generated binders in a physiologically relevant setting. We successfully applied this platform to identify functional peptide degraders for β-catenin, GFAP, and EWS::FLI1, oncogenic and disease-associated proteins that have remained challenging to target with conventional approaches. Overall, this platform enables scalable discovery of AI-designed peptide binders for proteome perturbation and therapeutic development.