Fig. 1: NOMAD overview and required tools for each step.
From: Rational strain design with minimal phenotype perturbation
We first integrate different types of data to build a set of putative kinetic models, represented by a system of ODEs. Next, we choose models based on dynamic characteristics such as their stability, ability to reproduce experimental fermentation data, and robustness to enzymatic interventions. Finally, we use the chosen models to conduct strain design. This involves solving a MILP optimization problem and enumerating designs that maintain the engineered strains close to the reference strain, evaluating the performance of these designs, and ranking them for experimental implementation.
