RFdiffusion

Welcome to the SBGrid YouTube channel, software tutorials by developers, lectures by structural biologists. Unique content brought to you by SBGrid.

Hello, everybody. Welcome to the first SBGrid webinar since the last solar eclipse or more seriously, the SBGrid webinar for April. We'll get started in just a moment. So our next webinar is going to be May 14th, and that's going to be Darren York, who's going to be talking about Amber. And if you have questions today, then please use the Q&A button.

And you can see the questions there that other people have asked, submit your own questions. And this is a slight change from when we used to be doing sending host questions to hosts in the chat. You can still do that if you'd like, but we've got the Q&A as well. And today, we've got Brian Trippe joining us

from Columbia University, and he's going to be talking about RFdiffusion. So Brian, thank you for agreeing to present today and thank you for joining us. And with that, I'll turn it over to you. Yeah, thanks for the introduction. And I'm pleased to be sharing this work on de novo design of protein structure and function

with RoseTTAFold diffusion.

Let's get going. Yeah. So a modern computational protein design workflow commonly involves a few different stages, all of which are now involving some deep learning methods and in many instantiations. So in all cases, we begin with some sort of design criterion. So think, for example, like having a tight binding interaction with some other protein