Case study · Manufacturing operations
Impact of DataHowLab in manufacturing operations
Three cases show how DataHowLab helps MSAT and manufacturing teams uncover root causes of variability and act before deviations become failures.
-
Seed
Bioreactor biomass identified as the driver of titer variability in the main fermenter
-
3
Factors, including the medium lot, found driving two CQAs out of specification
-
A simulated feed failure met with a recovery strategy that restored final titer
The challenge
Root cause in tightly controlled bioprocesses
Biological complexity, bioprocess interdependencies and fragmented data sources make root cause hard to identify. With limited variability in the data, conventional analysis often lacks the sensitivity to uncover the underlying drivers.
The approach
Hybrid models on combined bioprocess data
Hybrid models were trained on combined seed and main fermenter data with sensitivity analysis, decision-tree analysis was applied to out-of-specification CQAs, and the approach was extended to predictive deviation management for ongoing runs.
- 3 manufacturing cases
- 2 CQAs investigated
- 1 simulated failure
Get the full results
Download the full case study, with project outline, analysis, and case results.