Tomek Diederen

Over the course of his PhD Tomek has grown interested in 13C metabolic flux inference. Metabolic fluxes are biochemical reaction rates that cannot be directly measured and thus need to be inferred using a mathematical model using data from specially designed carbon labelling experiments. Tomek has developed a simulation based inference pipeline that employs machine learning algorithms called normalizing flows to approximate a Bayesian posterior over fluxes. This pipeline allows for the flexible specification of prior distributions over fluxes and for observation models that are tailored to specific analytical chemical methods.