Figure 2.

Model validation for case I. Model validation using optimal estimates obtained with different computational approaches for an experiment with altered independent variables. Dashed curves represent the predicted profiles using the optimal estimates obtained from estimation without including the flux connectivity constraints. Solid curves represent the predicted profiles using the optimal estimates obtained from an estimation accounting for noise-free flux connectivity constraints. Dashed-dot-dot curves represent the predicted profiles using the optimal estimates obtained from an estimation accounting for 5% noise in the measured flux connectivity constraints. Data points are in silico observations. The independent variables [1.376, 0.179, 0.588, 0.630, 1.340, 3.490, 4.7500] were set 5% outside the training range.