Scale-up with Historical Data & In Silico Process Optimization using DataHowLab
A hybrid model trained on small-scale data predicted titer at 50 L and 500 L with lower error than MLR. The in silico optimized feed recipe gave 15% higher titer than any previous 5 L run in a single validation run.
Read the case study


































