Mots-Clés

bioinformatique apprentissage profond interactions macromoléculaires évolution

Description

Interactions between proteins and nucleic acids (NA) have strong biological relevance and are often perturbed in diseases. This project aims to better understand the molecular mechanisms of protein-nucleic acid interactions. A large gap exists between vast amounts of high-throughput protein-NA interaction data and the scarcity of 3D structures of protein-nucleic acid complexes, triggering the need for computational methods to analyze and predict these structures. Recently, artificial intelligence (in particular deep learning in AlphaFold and related methods) has emerged as a revolutionary approach to predict protein and protein-protein interaction structures.

The present project aims to develop computational approaches for the analysis and prediction of protein-nucleic acid interactions, by building upon deep learning methodologies and integrating complementary data sources. We will use an evolutionary perspective to provide crucial insights into the exquisite regulation of complex processes driven by protein-nucleic acid interactions. The project will benefit from our expertise in macromolecular assembly structure and evolution and our tight coupling with experimental data through collaborations with wet-lab biologists.

The doctoral research will take place in the “Molecular assemblies and genome integrity” team of I2BC (Institute for Integrative Biology of the Cell, UMR 9198 CEA/CNRS/Université Paris-Saclay). I2BC is a research institute of Université Paris-Saclay that gathers 60 research teams covering a wide range of integrative biology projects. The “Molecular assemblies and genome integrity” team relies on a strong coupling between computational and experimental approaches to characterize, predict and inhibit macromolecular interactions. The team is localized south of Paris (so far on the CEA Saclay campus and soon on the Gif-sur-Yvette campus, less than one hour from the center of Paris).

Our bioinformatics team has a strong expertise in macromolecular structure and evolution, macromolecular interaction prediction and heterogeneous data integration. In recent years, our team developed original approaches for the structural prediction of protein interactions using evolutionary information (Quignot et al, NAR 2021; Quignot et al, Bioinformatics 2021). We also have numerous collaborations with wet-lab biologists.

CEA will provide 3 years PhD funding for this project to an excellent PhD candidate.