This webinar is organised by the ELIXIR 3D-BioInfo CommunityThe event will be hosted by
Programme: |
De novo antibody design with RFdiffusion
Dr. Joe Watson
(EMBO Postdoctoral Scholar at the Institute for Protein Design, University of Washington.)
De novo protein design seeks to learn the underlying principles of protein folding from natural proteins and to subsequently apply them to generate novel proteins with programmable functions. In recent years, there have been significant and concomitant advances in both our abilities to learn from protein structural data (AlphaFold2, RoseTTAFold) and in generative deep-learning methods in other fields (image and text generation). In this talk, I will discuss our recent work building upon these advances, in which we trained a generative neural network for de novo protein design; RoseTTAFold Diffusion (RFdiffusion). I will focus on the applications of RFdiffusion for designing protein-protein interactions, describing the characterization of hundreds of functional designed binders, followed by recent advances where we have extended RFdiffusion to design de novo nanobodies and antibodies.
From AlphaFold to PyMOL: Enabling seamless access to Structural Bioinformatics Tools
Serena Rosignoli
(Department of Biochemical Sciences “A. Rossi Fanelli”, Sapienza University of Rome, Rome, Italy)
AlphaFold, the groundbreaking AI system developed by Google DeepMind, has set the stage for a paradigm shift in structural biology. Its remarkable accuracy in predicting protein structures has ushered in a revolution in the field. However, the journey from AlphaFold’s initial release (AF2) to its current state has been marked by significant developments that have reshaped the landscape of structural bioinformatics.
One pivotal milestone in AlphaFold’s journey was the creation of a comprehensive database of predictions. This resource has not only facilitated its widespread adoption but has also spurred a wave of research and innovation across various biotechnological domains. As scientists and developers from diverse backgrounds delved into the intricacies of AlphaFold, a rapid increase in its utilization emerged, leading to unexpected insights and applications.
The growth in AlphaFold’s utilization has also catalyzed new software development efforts, with many focused on integrating AlphaFold’s predictions and addressing the algorithm’s remaining challenges. This development has introduced exciting possibilities, but it has also raised crucial questions. Whilst presenting our PyMOL-integrated solution to assist structural bioinformatics, we’re equally interested in broader questions.
- How can we ensure that AlphaFold seamlessly integrates into established protocols within structural bioinformatics?
- How can the synergy between AI-driven predictions and user-friendly platforms democratize access to cutting-edge structural biology tools?
- What role can the combination of homology-based methods and ab-initio algorithms play in advancing our understanding of complex protein structures?
You can find previous webinars from the 3D-BioInfo Community on the Community webinars page.