Hybrid Modeling of Engineered Biological Systems through Coupling Data-Driven Calibration of Kinetic Parameters with Mechanistic Prediction of System Performance

A hybrid modelingstrategy was developed to predict theperformance of engineered biological systems without the need forlaborious experiment-based parameter calibration. Mechanistic models can provide predictive insight intothe designand optimization of engineered biological systems, but the kineticparameters in these models need to be frequently calibrated and uniquelyidentified. This limitation can be addressed by hybrid modeling thatintegrates mechanistic models with data-driven approaches. Herein,we developed a hybrid modeling strategy using bioelectrochemical systemsas a platform system. The data-driven component consisted of artificialneural networks (ANNs) trained with mechanistically derived kineticparameters as outputs to compute error signals. The hybrid model wasbuilt using 148 samples from the literature. After 10-fold cross-validation,the model was tested with another 28 samples. Internal resistancewas accurately predicted with a relative root-mean-square error (RMSE)of 3.9%. Microbial kinetic parameters were predicted using the data-drivencomponent and fed into the mechanistic component to simulate the systemperformance. The R (2) values between predictedand observed organic removal and current for systems fed with a simplesubstrate were 0.90 and 0.94, respectively, significantly higher thanthose obtained from the stand-alone data-driven model (0.51 and 0)and mechanistic model (0.07 and 0.15). This strategy can potentiallybe applied to engineered biological systems for in silico system design and optimization.