What if biology could be engineered the way we engineer software?
In this episode of Galaxy Balance, I'm joined by Pranam Chatterjee, Assistant Professor at the University of Pennsylvania and leader of the Programmable Biology Group, working at the intersection of AI, synthetic biology, and next-generation therapeutics.
Pranam's work is shaping a future where generative models can design peptides and biologics from sequence data alone, enabling a new era of programmable medicine.
We explore how Pranam went from studying religion and philosophy to transferring into MIT and building cutting-edge computational tools for biology. We dive into his time in George Church's lab, where early computational strategies helped spark the origins of Gameto, and how that work evolved into today's iPSC-derived ovarian support cell technologies now entering clinical trials.
From there, we go deep into the frontier of AI-driven molecular design:
• Do we actually need protein structure to design effective therapeutics?
• How do we optimize binding, toxicity, permeability, and immunogenicity simultaneously?
• What does "virtual cell" really mean, and why does mapping cell states matter?
• How close are we to "vibe coding biology," where natural language becomes the interface to biological engineering?
We also discuss the future of automation, robotics, and agentic AI in biology, as well as the ethical risks of democratized generative models in biotech.
This conversation is a window into the net phase of human capability: not just ready biology, but designing it.
00:00 - Introduction to AI-driven therapeutic peptide design
01:05 - Background of Pranam Chatterjee's journey from religion to science
02:50 - The evolution of AI models in synthetic biology
05:17 - Key milestones: from modeling to clinical applications like Gameto
08:19 - The founding story of Gameto and major breakthroughs
12:20 - Expanding into disease targeting and regenerative medicine
17:50 - The shift to virtual cell and organism design
22:16 - Tools for peptide design: Peptune and PEPMLM
25:04 - Generative modeling with language models and functional constraints
28:32 - Imagining programmable organisms and mythical creatures
29:41 - Hardware importance and future of vibe-coded biology
31:54 - The role of automation and robotics in biotech labs
33:47 - Mentoring students for the AI-biotech revolution
36:50 - Targeting rare diseases and regulatory considerations
40:34 - Global competition, safety, and ethics in biotech innovation
44:44 - Designing molecules with AI: from complexity to deliverability
45:09 - Data needs: where to find diverse biological datasets
47:49 - The rise of AI agents in scientific research
50:12 - Ethical responsibilities in AI bioengineering
52:38 - Safeguards against harmful biotech applications
55:22 - Thoughts on artificial general intelligence and human purpose
58:40 - How science fiction inspires biotech innovation
1:00:13 - Book recommendations and closing thoughts
