NGBS2022 Talk 11: Protein design using deep learning - David Baker

Protein design using deep learning
Towards a mechanistic understanding of genome regulation
Speaker: David Baker, University of Washington, USA
Chair: Stephen McLaughlin, MRC Laboratory of Molecular Biology, UK
Co-Chair: David Rueda, Imperial College, London and MRC-LMS UK
David Baker describes de novo protein design via two methods using physically-based Rosetta and using deep learning neural networks. Using the first method, he shows how proteins can be designed for potential therapeutic use and to understand the photophysics of chlorophyll or create a transmembrane beta-barrel protein that can function as ion channel. This approach has also generated a set designed a set of protein building blocks that can be used for formations of larger structures not found in nature. In the second half of this talk, David describes how using a deep-learning approach can be used to find a protein sequence that will fold into a designed structure. This has been used to generate large symmetric rings, novel selective catalysts and stable epitopes for vaccine development.
Read more about David Bakers’s work here:
www.bakerlab.org/
You can find more about the Next Generation Biophysics Symposium 2022 here:
www3.mrc-lmb.cam.ac.uk/sites/...
--
About the MRC Laboratory of Molecular Biology (LMB):
The LMB is one of the world's leading research institutes. Discoveries and inventions developed at the LMB, for example DNA sequencing and methods to determine the structure of proteins, have revolutionised all areas of biology. Its scientists work to advance understanding of biological processes at the molecular level. This information will help us to understand the workings of complex systems, such as the immune system and the brain, and solve key problems in human health.
More links:
Official Site: www2.mrc-lmb.cam.ac.uk/
Facebook: / mrc.lmb
Twitter: / mrc_lmb
Instagram: / mrc_lmb
LinkedIn: / mrc-laboratory-of-mole...
Click here to subscribe to MRC Laboratory of Molecular Biology on KZbin: kzbin.info...

[@michaelzlprime]

Wow, nano-tech at its finest. the future is bright indeed!

[@critic1236]

Fantastic work, but I still carry a profound ignorance on all this. For instance, all these protein structures look like complex fur balls. To me they seem to work in some mechanic realm, but even with all the stuctures figured out by cryo and xray, I get no intuition into how this all works. So your work gives me some hope. at the 23 mark you make a simple ring then mechanically change it to a different state with another peptide. My problem in understanding is how in the world this peptide finds the right spot to make this happen? Self alignment nevers happens to me when I cant find my keys. Is it possible to put flourescent tags as orientation markers on the peptide and the protein to see how in the world the alignment takes place. Does the peptide go right to spot like a squirrel finds nuts or does it bounce around endlessly until it luckily finds the proper place to bind? Thank you for your provocative work.