Machine learning vs protein conformational spaces
Determining the different conformational states of a protein and the transition paths between them is key to fully understanding the relationship between biomolecular structure and function. I will discuss how a neural network can learn a continuous conformational space representation from example structures produced by molecular dynamics simulations. I will then show how such representation, obtained via our software molearn (1), can be leveraged to predict putative protein transition states (2), or to generate conformations useful in the context of flexible protein-protein docking (3).

1. github.com/Degiacomi-Lab/molearn
2. V.K. Ramaswamy et al., Learning Protein Conformational Space with Convolutions and Latent Interpolations. Physical Review X (2021).
3. M.T. Degiacomi, Coupling Molecular Dynamics and Deep Learning to Mine Protein Conformational Space. Structure (2019)
Date: 24 November 2022, 15:00 (Thursday, 7th week, Michaelmas 2022)
Venue: Dorothy Crowfoot Hodgkin Building, off South Parks Road OX1 3QU
Venue Details: 20-138 Phase 2 Ground Floor Main Seminar Room
Speaker: Matteo Degiacomi (Durham University)
Organising department: Department of Biochemistry
Organiser: William Rochira (University of Oxford)
Organiser contact email address: william.rochira@msdtc.ox.ac.uk
Host: William Rochira (University of Oxford)
Part of: SBCB Seminar Series
Booking required?: Not required
Audience: Members of the University only
Editor: William Rochira