BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Can Generative Models Produce Stable\, Rational\, and Diverse Prot
 ein Structures? - Tianyuan Zheng
DTSTART:20240812T163000Z
DTEND:20240812T171500Z
UID:TALK219829@talks.cam.ac.uk
CONTACT:Pietro Lio
DESCRIPTION:The complex geometric structure of proteins is key to their fu
 nction and specificity. Sixty years ago\, designing proteins seemed nearly
  impossible\, but today we can create fully synthetic proteins.\n\nWith th
 e rapid expansion of structural databases\, deep learning-based protein de
 sign methods are gaining attention. But can these generative models produc
 e "evolved" samples that follow physical and chemical rules\, fold correct
 ly\, remain stable\, and offer diversity and novelty?\n\nWe used Score Mat
 ching and Flow Matching methods to train generative models for monomeric p
 rotein backbone structures on the SE(3)-invariant Riemannian manifold\, an
 d tested these models on different monomeric proteins\, including cytochro
 me c\, green fluorescent protein\, monoglucosidase\, and $\\beta$-lactamas
 e\, to see how well they perform.
LOCATION:Lecture Theatre 2
END:VEVENT
END:VCALENDAR