BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Machine learning to predict protein function from sequence with th
 erapeutic applications - Dr Lucy J. Colwell\, Yusuf Hamied Department of C
 hemistry\, University of Cambridge
DTSTART:20230220T143000Z
DTEND:20230220T150000Z
UID:TALK196777@talks.cam.ac.uk
CONTACT:Dr Venkat Kapil
DESCRIPTION:A central challenge is to predict the functional properties of
  a protein from its sequence\, and thus (i) discover new proteins with spe
 cific functionality and (ii) better understand the functional effect of ge
 nomic mutations. Experimental and computational data enable powerful machi
 ne learning models that predict protein function directly from sequence to
  be trained and validated. I will present deep learning models that accura
 tely predict functional domains within protein sequences\, and large langu
 age models that generate textual descriptions of protein sequences\, colle
 ctively adding millions of annotations to public databases. Experimental b
 reakthroughs enable data on the relationship between sequence and function
  to be rapidly acquired. However\, the cost and latency of wet-lab experim
 ents require methods that find good sequences in few experimental rounds\,
  where each round contains a large batch of sequence designs. In this sett
 ing\, I will discuss model-based optimization approaches that take advanta
 ge of sample inefficient methods to find diverse sequence candidates for e
 xperimental evaluation. The potential of these approaches are illustrated 
 through three case studies demonstrating the design and experimental valid
 ation of proteins and peptides for therapeutic applications.
LOCATION:Pfizer  Lecture Theatre\,  Department of Chemistry
END:VEVENT
END:VCALENDAR