Séminaire IBS : Bioinformatics and deep learning methods for protein structure prediction and analysis

Everything you ever wanted to know about these techniques and what they can (and cannot) do for your studies

Structural bioinformatics methods are fundamental for the analysis and understanding of sequence/structure/function relationships of biological macromolecules. The research topics of our group focus on methodological developments in structural bioinformatics and more specifically on methods for protein structure exploration, analysis and prediction.

The dramatic breakthroughs of the last five years in the field of machine learning and the involvement of GAFAM*, capable of mobilizing unprecedented intellectual, human and material resources, have led to major successes in structural bioinformatics, particularly in the field of protein structure prediction, the most remarkable examples being AlphaFold 2 and ESMFold.

This seminar will focus on the different methodological developments that our group has proposed for protein structure analysis and prediction at both local (conformation, flexibility) and global levels (protein domain identification, membrane protein, fold recognition). We will also discuss the integration of recent advances in deep learning in our methodological approaches and the impact that these methods have had in these developments.

* GAFAM : Google, Apple, Facebook, Amazon, Microsoft

par Jean-Christophe Gelly (Département de recherches biologiques sur le globule rouge, Université Paris Cité)

Hôte : Antoine Royant (IBS/Groupe Synchrotron)

Some group’s publications :

• Cretin G, Galochkina T, Vander Meersche Y, de Brevern AG, Postic G, Gelly JC. SWORD2 : hierarchical analysis of protein 3D structures. Nucleic Acids Res. 2022 ;50(W1):W732-W738.

• Cretin G, Galochkina T, de Brevern AG, Gelly JC. PYTHIA : Deep Learning Approach for Local Protein Conformation Prediction. Int J Mol Sci. 2021 ;22(16):8831.

• Vander Meersche Y, Cretin G, de Brevern AG, Gelly JC, Galochkina T. MEDUSA : Prediction of Protein Flexibility from Sequence. J Mol Biol. 2021 ;433(11):166882.

• Postic G, Ghouzam Y, Chebrek R, Gelly JC. An ambiguity principle for assigning protein structural domains. Sci Adv. 2017 ;3(1):e1600552.

• de Brevern AG, Bornot A, Craveur P, Etchebest C, Gelly JC. PredyFlexy : flexibility and local structure prediction from sequence. Nucleic Acids Res. 2012 Jul ;40(Web Server issue):W317-22.