What Does Deep Learning See in 3D Protein Structures?

Although the fundamental forces between atoms and molecules are almost fully understood at a theoretical level, and computer simulations have become an integral part of research activities, the application of these methods to large biomolecules still faces important practical difficulties due to the combinatorial explosion of possible interactions involved. Developing efficient protein structure prediction algorithms thus remains a major scientific challenge in computational biology.

The speaker, Sergei Grudinin (Inria / CNRS, Grenoble, France), will give an overview of computational methods for protein structure prediction developed in their team at Inria Grenoble [1]. In particular, he will demonstrate how machine/deep learning can be used in current problems of computational structural biology. Indeed, artificial intelligence has made a big leap forward and found many applications in structural bioinformatics. On their side, they have been using it for multiple tasks of protein structure prediction, including protein-protein [2] and protein-ligand [3] docking, 3D shape analysis [4] and structure prediction [5-6].

References
[1] - https://team.inria.fr/nano-d/software/.
[2] - Neveu, E., D. Ritchie, P. Popov, & S. Grudinin (2016). Bioinformatics, 32 (7), pp.i693-i701.
[3] - Kadukova, M. & S. Grudinin (2017). J. Comp.-Aid. Mol. Des. 31 (10), pp.943-958.
[4] - Pages, G. & Grudinin, S (2019). Bioinformatics, 10.1093/bioinformatics/btz454.
[5] - Karasikov, M., G. Pagès, & S.Grudinin (2018), Bioinformatics, bty1037.
[6] - Pages, G.,  B. Charmettant & S. Grudinin (2019). Bioinformatics, btz122