Experience

Sorbonne Université

• France
• Higher Education
• 700 & Above Employee

• Chercheuse postdoctoral

  • Mar 2023 - Present



Laboratoire de Biologie Quantitative et Computationnelle UMR 7238

• Ville de Paris, Île-de-France, France

• PHD Student

  • Oct 2018 - Dec 2022

  In this thesis, I focused on the use of machine learning to solve the problem of the prediction of protein-protein interactions (PPI). The study of PPI is a central problem in biology, as proteins interact with each other to form complex networks that carry out the biological functions of cells. Experimental techniques to determine when and how proteins interact are very costly and time-consuming, so there is a great need for computational methods that can predict PPIs. We will explore the use of machine learning based on coevolution and deep learning for PPI prediction. Coevolutionary methods such as Direct Coupling Analysis have been used successfully for a number of different tasks, such as the prediction of intra-protein contacts, inter-protein contacts, and the prediction of mutational landscape. During my Ph.D., I developed a supervised machine learning algorithm to predict inter-domain and inter-protein contact maps called FilterDCA. The aim was to add some supervision, using typical contact patterns, while keeping the tool interpretable. I have also worked on PPIs in the SARS-CoV2 virus and on a multi-protein complex present in some bacteria membranes. Show less