Uploaded July 2026 | Updated September 2026, 3 weeks ago
AI is transforming biology and it can now design antibodies from scratch. I sit down with Santiago Mille, bioengineer and Stanford PhD candidate (Gao & Hie labs), whose Nature Biotechnology paper introduces Germinal: an open-source AI pipeline that designs antibodies to hit a chosen target, testing just a few dozen candidates instead of the tens of thousands older methods needed.
We cover what he built and how it works, how it compares to other protein-design efforts, the bottlenecks still facing the field, and where it's all heading — toward a much bigger toolbox of designer therapeutic molecules.
Paper: doi.org/10.1038/s41587-026-03187-0
Santiago: https://santiagomille.dev
Find me on Twitter - twitter.com/EleanorSheekey
official website - thesheekeyscienceshow.com
Support the channel
through PayPal - paypal.me/sheekeyscience?country.x=GB&locale.x=en_GB
through Patreon - patreon.com/TheSheekeyScienceShow
TIMESTAMPS
00:00 Intro
01:06 Why design antibodies — from protein sensors to Germinal
07:16 How Germinal works & what it needs
19:06 Compute, screening & the real bottlenecks
50:25 Beyond antibodies: what's next
Please note that The Sheekey Science Show is distinct from Eleanor Sheekey's teaching and research roles. The information provided in this show is not medical advice, nor should it be taken or applied as a replacement for medical advice. The Sheekey Science Show and guests assume no liability for the application of the information discussed.
Icons in intro; "freepik.com/free-photos-vectors/background"Background vector created by freepik - freepik.com
AI is transforming biology and it can now design antibodies from scratch. I sit down with Santiago Mille, bioengineer and Stanford PhD candidate (Gao & Hie labs), whose Nature Biotechnology paper introduces Germinal: an open-source AI pipeline that designs antibodies to hit a chosen target, testing just a few dozen candidates instead of the tens of thousands older methods needed.
We cover what he built and how it works, how it compares to other protein-design efforts, the bottlenecks still facing the field, and where it's all heading — toward a much bigger toolbox of designer therapeutic molecules.
Paper: doi.org/10.1038/s41587-026-03187-0
Santiago: https://santiagomille.dev
Find me on Twitter - twitter.com/EleanorSheekey
official website - thesheekeyscienceshow.com
Support the channel
through PayPal - paypal.me/sheekeyscience?country.x=GB&locale.x=en_GB
through Patreon - patreon.com/TheSheekeyScienceShow
TIMESTAMPS
00:00 Intro
01:06 Why design antibodies — from protein sensors to Germinal
07:16 How Germinal works & what it needs
19:06 Compute, screening & the real bottlenecks
50:25 Beyond antibodies: what's next
Please note that The Sheekey Science Show is distinct from Eleanor Sheekey's teaching and research roles. The information provided in this show is not medical advice, nor should it be taken or applied as a replacement for medical advice. The Sheekey Science Show and guests assume no liability for the application of the information discussed.
Icons in intro; "freepik.com/free-photos-vectors/background"Background vector created by freepik - freepik.com
