The International Group for Data Analysis (IGDA) led by Dr Magnus Fontes is part of the Center for Bioinformatics, BioStatistics and Integrative Biology and the International Division in Institut Pasteur Paris. This is the Parisian node of the International Network for Data analysis (INDA).
Good models require good data. We are developing state-of-the-art mathematical models, algorithms and software to answer biomedical questions. We have close relationships with wetlabs, and we are directly implicated in the biological experimental designs from the start.
Good decisions require good knowledge of the data. We are developing new visualization and analysis methods making it possible to integrate biomedical knowledge directly in the modelling of the data.
Good ideas require good interaction. We are educating mathematical fellows closely in contact with the biology labs. This allows for a more direct communication between the experimentalist and modelers.
In a cross-disciplinary and collaborative effort we will attack one of the most significant big data dependent modelling challenges in life science research today: establishing an understanding of the human immune system and inter-individual immune response variation. This project is an extension of the Milieu Interieur effort already ongoing. Our group will model the dynamics of robust features within the human immune system and their interactions with corresponding biologically relevant features in human microbiota and pathogens. This will lead to an unprecedented understanding of how these interactions have influenced microbiome communities, pathogen and human genetic evolution. Two projects are already ongoing:
Around 1000 patients are being analyzed for different markers in the Milieu Interieur project, and we need to have good robust data between different sets of measurements and patients. Understanding how batch effects can affect the data generation, and how to minimize and avoid such effects is paramount to this project effort.
Automatic flow gating
Several cell surface marker panels were defined in the 1000 patients from the Milieu Interieur to be analyzed by flow cytometry. Each one of these panels define a subset of cells (Dendritic cells, T-cells, B-cells, etc). Manual gating in such large amount of data is costly and is prone to human error. Developing mathematical models that allow to analyze such data in a multidimensional way is desirable to avoid such errors and rapidly characterize the base levels of each cell subset. Together with Benno Schwikowski's lab we are working on an automatic gating analysis platform.