BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:AbDiffuser: Full-Atom Generation of In Vitro Functioning Antibodie
 s - Karolis Martinkus & Andreas Loukas (Prescient Design\, Genentech\, Roc
 he)
DTSTART:20231024T120000Z
DTEND:20231024T130000Z
UID:TALK206236@talks.cam.ac.uk
CONTACT:117757
DESCRIPTION:We will discuss AbDiffuser\, our latest equivariant and physic
 s-informed diffusion model for the joint generation of antibody 3D structu
 res and sequences. AbDiffuser is built on top of a new representation of p
 rotein structure\, and utilizes strong diffusion priors to improve the den
 oising process. Our approach improves antibody diffusion by taking advanta
 ge of domain knowledge and physics-based constraints\; handles sequence-le
 ngth changes\; and reduces memory complexity by an order of magnitude enab
 ling backbone and side chain generation. We have validated AbDiffuser in s
 ilico and in vitro. Numerical experiments showcase the ability of AbDiffus
 er to generate antibodies that closely track the sequence and structural p
 roperties of a reference set. Laboratory experiments confirm that all 16 H
 ER2 antibodies discovered using AbDiffuser were purified and expressed at 
 high levels and that 57.1% of selected designs were novel tight binders.\n
 \nArXiv: "https://arxiv.org/abs/2308.05027":https://arxiv.org/abs/2308.050
 27 NeurIPS 2023 Spotlight\n\n*Speakers Bio:*\n\nKarolis Martinkus is a Mac
 hine Learning Scientist at the Prescient Design team within Genentech Rese
 arch & Early Development (gRED) where he works on generative models for de
 -novo antibody design. He completed his PhD at ETH Zurich under the superv
 ision of Prof. Roger Wattenhofer\, where he focused on applying deep learn
 ing to structured domains (e.g. graphs)\, in particular generative models.
 \n\nAndreas Loukas is a Senior Principal Scientist and Machine Learning Le
 ad at Prescient Design  within Genentech Research & Early Development (gRE
 D). His work focuses on the foundations and applications of machine learni
 ng to structured problems. He aims to find ways to exploit (graph\, constr
 aint\, group) information\, with the ultimate goal of designing algorithms
  that can learn from fewer data. He is also focusing on the theoretical an
 alysis of neural networks and in using them to solve hard bioengineering p
 roblems (especially protein design). Andreas obtained his Ph.D. in compute
 r science from TU Delft in 2015 and pursued postdoctoral studies at TU Ber
 lin and EPFL. He became an SNSF Ambizione fellow at EPFL in 2018 and an As
 sistant Professor at the Computer Science department of the University of 
 Luxembourg in 2021. He joined Genentech/Roche in 2022.\n\n"Join us on Zoom
 ":https://cam-ac-uk.zoom.us/j/92041617729
LOCATION:Zoom: https://cam-ac-uk.zoom.us/j/92041617729
END:VEVENT
END:VCALENDAR