Cross-cutting methodologies and tools

Although numerous experimental platforms (for designing and analyzing catalytic and biological processes) are available nationally and particularly among the partners of the Bioproductions program, there is a lack of data exchange standards that hinder their use for designing, building, and testing bioprocesses. The first challenge in this area is to inventory and standardize resources, make them easily accessible via a portal, and use them to reduce development time and costs in research and development. The second challenge is to utilize data produced by experimental platforms with machine learning methods to ultimately propose accurate predictive models for biomass reactivity, biomass transformation, and the production of bio-based products via bioprocesses. The difficulty here will be integrating machine learning models with existing first-principles or mechanistic models to fully leverage the advantages of these approaches.

This research axis consists of three targeted projects and will be complemented by the winning projects from the program's call for projects:

Targeted projects: 

• MAMABIO : machine learning methodologies for accelerated and predictive atomic-scale simulation of bio-based molecule transformation.
• AMARETTO : develop a model to predict the reactivity of lignocellulosic biomass from various origins in enzymatic hydrolysis.
• Galaxy-BioProd : develop a centralized portal to provide software tools as well as computational and storage resources to create workflows benefiting other targeted projects in the program and the industrial biotechnology community.

Winner of the 2024 call for projects :

• MuSiHC: multi-size hybrid cell models.