Everything upstream of this section is judged by whether it produces strains that do something worth doing. Our applied work centers on microbial production of valuable compounds — melatonin and creatine, along with bulk and specialty chemicals — where the objective is not simply that a pathway functions but that titer, rate, and yield reach values a process can be built around. We also engineer microbes to act on waste as a feedstock, developing organisms that break down persistent plastics and agricultural waste into material worth recovering. In both cases the route to a working strain is the same combination the rest of our research is built on: model-guided design to identify which genetic and metabolic changes should produce the desired physiology, evolution under selections we design to reach performance that rational design alone cannot deliver, and mutation analysis to understand and transfer whatever the cell discovered along the way.

A strain that performs in a flask and fails in a fermenter has solved the wrong problem. Industrial conditions impose stresses that laboratory conditions do not — accumulating product, shifting pH and oxygen, and the sheer duration of a real run — so we treat tolerance and robustness as engineering targets in their own right rather than as problems to discover late. The same applies to what a strain eats: we develop substrate flexibility so hosts grow on feedstocks that make economic sense, not just the ones that make experiments convenient. Throughout, phenotyping is done under conditions chosen to anticipate scale-up, so that performance measured in the lab survives contact with a process. This work happens in collaboration with industry, national laboratories, and academic partners, which keeps the target definitions honest — the specification comes from whoever intends to run the strain, not from us.