Dr L. Bou Dagher holds a PhD in Mathematics, and her research spans topological data analysis, spectral geometry, machine learning, evolutionary biology, and structural biology. Her work focuses in particular on the analysis of protein structures and medical imaging (e.g., tumour detection and classification), using persistent homology.

 

With the development of digital tools and the tremendous increase in storage resources and computational power, data production is exploding in many fields such as science, engineering, and healthcare. The challenge lies in analyzing these data both qualitatively and quantitatively, using a variety of approaches. One such approach is topological data analysis, which makes it possible to process very large datasets and highlight the most significant aspects of their structure. In particular, persistent homology, derived from algebraic topology, provides efficient and robust algorithms for the exploration, topological analysis, and comparison of high-dimensional datasets, represented by point clouds, by associating them with topological invariants.
In this talk, I will present the principles of persistent homology and its applications to the study of proteins. Specifically, I will explain how persistent homology can be used to investigate the 3D structures of proteins and the information they contain, including their evolutionary history and their adaptation to environmental conditions.