There will be a presentation on structural bioinformatics on Thursday 5th of July at 11h00 in Room B013.

Sergei Grudinin, from Nano-D, (Inria/CNRS) in Grenoble will give a talk entitled “Novel computational approaches for structural bioinformatics”.

Abstract :

In my talk I will briefly present our recent developments in using polynomial expansions, collective motions, Fourier analysis, machine learning, and symmetry information for several problems in structural bioinformatics. I will first present our approach of computing nonlinear large-amplitude motions in the Cartesian space using a nonlinear extrapolation of the instantaneous eigen-motions of large biological objects. Then, I will show how to compute collective motions in the Fourier space and how to extend the Fourier correlation theorem for flexible objects. I will also present novel methods of using polynomial expansions for small-angle X-ray and neutron scattering of rigid and flexible objects. Finally, I will show how machine learning and 3D convolutional networks can help in constructing knowledge-based scoring functions for protein quality assessment, and what we can learn from exhaustive analysis of symmetric protein structure in the protein data bank.