Skip to main content
← All resources

Case study · Industrial cases

Unlocking gene therapy with DataHowLab

Highlights from industrial cases show how DataHowLab delivers development insight for gene therapy bioprocesses despite limited, costly data.

  • Gene therapy
  • CQA prediction
PREDICTION ERRORNEW PROCESS ONLYHIGHER+ 2 HISTORICAL PROCESSESLOWEREXPERIMENTS TO PREDICT FINAL CQAS12 EXPERIMENTSFINAL CQAS PREDICTED
  • 12

    Experiments were enough to predict final CQAs

  • 3

    Bioprocess phases mapped for variable importance: growth, perfusion and transfection

  • 2

    Historical gene therapy bioprocesses added to training reduced prediction error

The challenge

CQAs measured only at the end

Gene therapy bioprocesses are complex and sequential, while CQAs are typically measured only at the end. Issues surface late, leading to failed batches, and high experimental costs limit the data available.

The approach

Hybrid models and transfer learning

DataHowLab was assessed on predicting final CQAs from limited data, generating insight across each phase to support a sequential control strategy, and using historical projects to improve insight for new developments.

  • 3 industrial cases
  • RMSE accuracy metric
  • End of bioprocess CQA measurement

Get the full results

Download the full case study, with project outline, analysis, and case results.