BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:"Machine learning in chemistry and beyond" (ChE-651) seminar by Ro
 cío Mercado "Exploring new frontiers in drug discovery using deep generat
 ive models"
DTSTART:20221108T151500
DTEND:20221108T161500
DTSTAMP:20260610T131156Z
UID:8dc9dc3ba0dfec27a39b0abe96eb3fafae09bd1ed18615b310ff2988
CATEGORIES:Conferences - Seminars
DESCRIPTION:Rocío Mercado is currently a post-doc at MIT working with Pro
 fessor Connor Coley.  Previously\, she completed an industrial post-doc a
 t AstraZeneca in the Molecular AI team where she worked on graph molecular
  generative models for small molecule drug design. Before that\, she compl
 eted a PhD in Chemistry with Professor Berend Smit at UC Berkeley and EPFL
  in molecular simulation. Rocío will be starting an assistant professorsh
 ip at Chalmers University of Technology in the Data Science and AI divisio
 n working on data-driven molecular design.\nArtificial intelligence (AI) i
 s transforming our approach to biomolecular engineering. In the drug disc
 overy sector\, the development of generative and predictive tools that can
  quickly learn from biochemical data is pushing the frontiers of how we di
 scover and repurpose drugs through de novo molecular design. De novo d
 esign – the concept of designing molecules with desired properties from 
 scratch so as to minimize experimental screening – is poised to allow sc
 ientists to more efficiently traverse chemical space in search of optimal 
 molecules\, and delegate error-prone decisions to computers via the use of
  predictive and generative computational models. In drug development\, de
  novo design methods can aid medicinal chemists in the design and selecti
 on of drug candidates\, with the added advantage that they can learn from 
 datasets of billions of molecules in minutes and be constantly updated wit
 h new data. In this talk I will introduce molecular deep generative models
  (DGMs) and their utility in de novo design. DGMs use deep neural networ
 ks to build new molecules in silico\, and work by proposing node-by-node 
 modifications to an initial graph structure to generate compounds predicte
 d to achieve a certain property profile. Such models can be applied to a r
 ange of therapeutic modalities\; here I will focus on the design of small 
 molecule protein binders and heterobifunctional degraders. I will end by
  touching on the importance of interdisciplinary communication for the dev
 elopment of new advances in this field\, as well as discussing the importa
 nce and impact of open-source work.
LOCATION:https://epfl.zoom.us/j/68447908297?pwd=OU5JUGJUSUhZc0ZNYjQ2WENvYl
 NRdz09
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR