Fig. 3
From: A machine learning approach to predict metabolic pathway dynamics from time-series multiomics data
The new method was tested on the limonene and isopentenol metabolic pathways. The limonene (blue) and isopentenol (red) producing pathways are variants of the mevalonate pathway. Time-series proteomics and metabolomics data are used to learn the dynamics of both the isopentenol and limonene producing strains. Additionally, a kinetic model is created and compared to the machine learning approach for the more complex limonene production pathway (Fig. 4). This pathway model is also used to generate simulated data to further evaluate the scaling properties of the proposed machine learning approach. See the original data set66 for enzyme and metabolite acronyms
