BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:AlphaFold: Improved Protein Structure Prediction Using Potentials 
 from Deep Learning
DTSTART:20200129T110000
DTEND:20200129T120000
DTSTAMP:20260407T064114Z
UID:5c22247037580b72c1fb257cc1663ace7f9b94c4cfcb023799e70503
CATEGORIES:Conferences - Seminars
DESCRIPTION:Andrew W. Senior\, Ph.D.\, Google DeepMind\, London (UK)\nBioe
 ngineering Mini-Symposium - "A Morning in Machine-Learning and Protein Pre
 diction and Design" - Talk Two\n(Talk One: see here)\n \nAbstract:\nProte
 ins are essential for nearly every process in living organisms and their f
 unction is determined by the 3D structure of the component amino acids\, m
 aking structure determination a grand challenge in biology. Experimental d
 etermination is difficult and does not scale to the billions of proteins b
 eing discovered through genetic sequencing\, so accurate computational str
 ucture prediction is essential to advancing biological knowledge. DeepMind
 's AlphaFold protein structure prediction system was ranked first in free-
 modelling at the CASP13 (Critical Assessment of Protein Structure Predicti
 on) Biennial blind assessment of protein structure prediction methods. The
  system relies upon prediction of inter-residue distances by a very deep n
 eural network. Using these distance distributions and a reference distribu
 tion from a similar neural network\, we construct a potential and show tha
 t we can optimize this potential by a simple application of gradient desce
 nt\, as well as with a more conventional fragment assembly / simulated ann
 ealing algorithm. Despite not using templates the system also performed we
 ll in the CASP template-based category. We will discuss the training and u
 se of the neural network and present contact- and structure-prediction res
 ults from the CASP assessment and indicate potential future directions.\n\
 nZoom link for attending remotely: https://epfl.zoom.us/j/282485607
LOCATION:SV 1717 https://plan.epfl.ch/?room==SV%201717
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR