Structure-Independent Peptide Binder Design via Generative Language Models

Explore cutting-edge approaches to peptide binder design using generative language models in this hour-long conference talk by Pranam Chatterjee from Valence Labs. Discover how protein language model embeddings can be leveraged to design target-specific peptides without relying on 3D structures, addressing challenges in targeting "undruggable" and intrinsically disordered proteins. Learn about novel algorithms for selecting high-affinity peptides, developing selective peptide discriminators, and generating binding peptides de novo. Gain insights into applications ranging from Cas9-PAM engineering and transcription factor-mediated cell engineering to COVID-19 diagnostics and protein targeting. The talk concludes with a Q&A session, providing an opportunity to delve deeper into this promising field of computational biology and drug discovery.

- Intro
- Contrastive Learning for Cas9-PAM Engineering
- Transcription Factors as Tools for Cell Engineering
- Transcription Factor Mediated Generation of Human Ovary
- COVID-19
- Peptide Beacons as a COVID Diagnostic
- CLIP for Guide Protein Design
- Peptide Prioritization with CLIP PepPrCLIP
- GPT-2 for Proteins
- Q&A