Showing posts from January, 2023

De novo proteins and where to find them - RosettaCon 2022

By Kieran Didi There has been a lot happening recently in protein design, and it is easy to get lost in the daily flood of new papers and exciting ideas (for an overview of current models and approaches see this review which includes a great model table ). In such situations, it often helps to step back for a second, look at the bigger picture and chat to people about what is happening and what the future might hold. Luckily, RosettaCon was happening this August and provided a venue for exactly that: chat to people about protein design and fascinating ideas! In this post I want to highlight some presentations from the conference that I think are representative of some broader directions in the field.  Scaffolding protein motifs using Deep Learning - (Baker Lab, IPD, UW) While complete de novo design of proteins is still the holy grail of the field, in practice you often want to incorporate a motif  known from nature into a new custom-made scaffold. This motif-scaffolding problem