Data mining for bioprocess optimisation

Process Analytical Technologies (PAT) and Quality by Design (QbD) in biotechnology

Just as Six-Sigma and statistical process control (SPC) these approaches to process optimisation are based on the interpretation of measurements. Motivated by regulatory requirements and the desire to operate more profitably the efficient interpretation of archived measurements becomes the key for operational excellence in production and process development.

But often a lack of time and knowledge of suitable interpretation methods prevents staff on site to carry out a proper analysis: important effects are then easily being overlooked and data misinterpreted.

„Closing the loop – from data to value“

With engineo’s years of expertise comprehensive and confirmed results are guaranteed. Our knowledge-based approach to process optimisation by means of data-mining has been developed during the course of many projects and has constantly been improved. It is characterised by :

  • High quality of results, as all data used is subject to rigorous verification and consistency checks.
  • Increased informational content due to specific preparation of raw data.
  • Increased validity of data due to the calculation of process specific characteristic parameters.
  • Knowledge-based data-mining is much more than a simple statistical or multivariate data analysis. It explicitly considers process understanding in a comprehensive approach, which takes into account biology as well as systems and measurement engineering.
  • Proprietary expert systems make professional analyses possible.
  • More starting points for improvements as not only data but also practical knowledge and additional information sources are taken into account.
  • Attractive project budgets, because we only require a reduced amount of data. Therefore we are able to offer fast cause studies for conspicuous batches (trouble shooting).

Bioprocess optimisation task

Scale Task
  • Improving product yield and by-product profile
  • Increasing reproducibility of quality parameters and key performance indicators (KPIs)
  • Trouble shooting for conspicuous and off-spec batches
  • Statistical process control
Process development
  • Development of scale-up and scale-down models
  • Modelling of metabolism
  • CCharacterisation of cells
  • Automated data mining
All scales
  • PAT, QdD, Six-Sigma projects