Imperial College London

Imperial researchers develop and test next generation models of industrial cells


Digital model of DNA

Imperial scientists have published a new metabolic model for the most widely used industrial mammalian cell line with aims to improve drug production.

Biopharmaceuticals are a vital class of drug, offering life changing treatments for diseases such as cancer and arthritis. Biopharmaceuticals  are predominantly produced by Chinese Hamster Ovary (CHO) cells, thanks to their ease of use and excellent safety record. To improve the quality and quantity of biopharmaceuticals during production, it is possible to engineer the cell line or bioprocess. To do this in an intelligent way researchers may use a class of models known as Genome scale metabolic models (GeMs), these models contain information on every known metabolic reaction and associated genes that occur within CHO cells and allow scientists to predict the important parts of metabolism that may be engineered to improve the production of these lifesaving medicines. 

While there are a number of these GeM available for CHO cells, it is not exactly clear how reliable these models are and whether they are appropriate to be embedded in industry.

In a study published in Biotechnology and Bioengineering, scientists from the Department of Chemical Engineering have developed a new CHO cell GeM by consolidating previous models together. They then tested the reliability of their new model against previous models by testing the performance of extracellular (growth rate and gene/nutrient essentiality) and intracellular (reaction rate) predictions.

More accurate predictions for cell line engineering

Their analysis revealed that all models gave reasonable predictions, with their new model outperforming the original model in terms of intracellular predictive performance. This means that the updated model can more accurately predict features about the cell which can lead to cell line engineering decisions, improving both the quality and quantity of biopharmaceutical product.

They also demonstrated that by integrating other types of biological data into the model, specifically transcriptomic data, they could improve the predictive performance of extracellular predictions. Their new model acts as the perfect vessel for the integration of this data as it includes the highest number of genes of any CHO cell GeM to date, meaning even better model performance when additional data is added. 

Significantly, throughout their analysis, researchers were able to identify parts of each model that performed poorly. This is vital for developing next generation techniques to run these models to get more accurate predictions for intelligent model-led cell line and bioprocess engineering, leading to the improved production of future drugs.

A new model ready for industry  

Given the improved performance of their new model, researchers expect their updated model to be the new model of choice for the CHO cell bioprocessing community. Given the strong results presented in this paper, they are confident that these models are ready to be embedded in industrial workflows for intelligent cell line engineering, helping to design the next generation of industrial mammalian cell lines for improved drug production.

Want to find out more about how genome scale models can be used? Watch the below video:

Video placeholder image


'How reliable are Chinese hamster ovary (CHO) cell genome-scale metabolic models?' by Benjamin StrainJames MorrisseyAthanasios Antonakoudis and Cleo Kontoravdi, published 2 March 2023 in Bioengineering and Biotechnology.

Image: Pixabay


Ben Strain

Ben Strain
Department of Chemical Engineering

Click to expand or contract

Contact details


Show all stories by this author


See more tags