Designing the Future of Drug Delivery: How AI is Solving RNA's Biggest Challenge

Yogev Debbi and Prof Avi Schroeder (Technion Israel Institute of Technology), co-founded Mana bio along with Dr. Kira Radinsky and Roy Nevo. They are leveraging data, machine learning and high throughput screening to design novel Lipid NanoParticles for targeted extrahepatic delivery of nucleic acid therapeutics and vaccines.

Lipid nanoparticles are microscopic fatty bubbles that deliver genetic medicines into your cells—think of them as the FedEx trucks of molecular biology. Remember the mRNA vaccines that saved millions during COVID? They used lipid nano particles as the delivery vehicle.. The little secret behind those vaccines is that we got lucky. Those lipid nanoparticles worked beautifully for the liver and immune system. But if you try to deliver RNA to your brain, your lungs, or a tumor, those would be a massive failure! Scientists are still running thousands of experiments hoping to stumble onto the right formulation. This actually was my postdoc work as well, which makes this episode very special. I always wished there was a better way to design these delivery vehicles—and that's exactly what today's guests have built with AI."

We talk about how they founded and currently building Mana Bio- a big part around how biologists and chemists can interact with technologists, especially data and AI geeks to build something important. Their story provides a great framework for scientists and tech folks to work together.

Shownotes:

• https://www.mana.bio/
• How Avi got started using data/AI to design LNPs
• Optimizing LNP formulations targeting different organs
• Working with Bob Langer and Dan Anderson @MIT
• Yogev transitioning to gene editing
• Defining the problem & cross disciplinary conversations and projects
• Language that helps interactions between scientists and software/AI folks
• Deepdive into the platform- building data sets
• Data scraping, validation, improvisation, new data generation (both positive and negative data)
• Cleaning data and normalization
• Work in the wet lab: 'I trust my colleagues and believe data can be replicated"
• Data moats: Proprietary data vs public data
• Daily routine in the lab and interactions with the data/AI team
• Cross Functional dynamics
• Scientists trying to save money vs optimizing for more valuable time
• Lab hypothesis predictions vs AI predictions: competitive spirit
• Factoring in errors by the machines and humans
• Business model: Building customized solutions for drug delivery
• Potentially becoming a pharma company
• Regulatory affairs
• Fears that keep them up
• Dynamics of a diverse founding team: Roy Nevo and Dr. Kira Radinsky (the other two co founders)
• Message to academia: "Publish more data- including failures. This will accelerate science"
• Omri Drory, Kira Radinsky
  https://www.youtube.com/watch?v=y5hX3vq3iNk