Reservoir Computing-Based Glucose Sensing With an Enzymatic Reaction Network

The survival of living systems relies on their capacity to process physicochemical information from the environment and make distinct decisions. Inspired by this concept, we report a scalable protease-peptide-based reservoir sensor for the detection of glucose stimuli. In this system, glucose oxidase (GOx) functions as a 'sensor module' that preprocesses glucose into a pH signal that can be further processed by the protease network. The effect of glucose induced pH change on the protease-peptide network is systematically investigated. GOx mediated pH drop ultimately affect protease activities that lead to an observable response in the reservoir output. Finally, we demonstrate the protease-based network's capacity to perform diverse information processing tasks by leveraging principles of reservoir computation. The enzymatic reservoir sensor not only functions as a linear glucose sensor but also exhibits binary switch like responses at both high and low glucose concentrations. Furthermore, the sensor can respond to dynamic glucose inputs, such as changes in the periodicity of glucose pulses or variations in both pulse amplitude and periodicity.