Société Française de Bioinformatique

Mots-Clés

structural bioinformatics protein-carbohydrate interactions deep learning

Description

Protein-carbohydrate (PC) interactions play a key role in various biological processes. However, as compared to protein-protein or drug-protein interactions,  PC interactions remain much less studied due to the difficulties related to their experimental investigation. In this context, we recovered all the available experimental structures of protein-carbohydrate complexes and curated a new protein-carbohydrate binding site database containing thorough annotation and clustering of the observed types of different binding sites. The student will participate in the second stage of ANR SugarPred project and will implement deep learning models aiming at predicting potential carbohydrate binding sites on a protein surface and their classification according to the proposed annotation. The tools developed during the internship will be made publicly available through scientific publications and open source code. They will also be used throughout the next steps of the SugarPred project.