Yahya N. Farooqi

Machine learning provides us with a powerful framework for predicting industrially relevant phenotypes, however this framework is mechanism agnostic. Metabolic models, in contrast, are mechanistically informative, however they lack the robust predictive efficacy that machine learning algorithms possess in predicting complex phenotypes. Naturally, a combination of these two modeling paradigms may yield models that can predict complex phenotypes in a mechanistically informative way.

Exploring such combinations is the main goal of my PhD project. As part of the FAIROmics MSCA consortium, this project aims to build models that can predict microbial phenotypes in relation to the fermentation of plant-based food substrates, in order to improve their texture, flavor, and sustainability.