(579c) Genome Scale Models As a Data Integration Centre for CHO Cell Process Optimisation

With the plethora of data becoming available in mammalian cell bioprocessing, genome scale models (GEMs) have emerged as an excellent framework to integrate and understand multiple datasets. These datasets (such as genomic, proteomic, transcriptomic, metabolomic) are rich with information when investigated individually, however when they can be assessed in parallel, the insights gained are multiplicative. It has been difficult to reconcile these datasets in some traditional bioinformatics workflows. As a “hub for understanding the host”, GEMs offer a powerful and easily interpretable method to bring together bioprocess knowledge.

The recent developments of CHO cell GEMs have made it easier to gain deep insights into CHO cell metabolism. These developments have coincided perfectly with the rise of ‘big data’ in the field of systems biology. In this work we outline the potential of genome scale modelling for data interpretation and discuss the steps from process understanding to process optimisation. We demonstrate practically how this workflow has been used in an industrially relevant CHO cell bioprocess.

Starting with the creation of a cell-line specific GEM, we constrain and apply dynamic flux analysis. We connect the fluxomics predictions with layers of omics data across different operating conditions and use pathway analysis to correlate phenotypes with metabolic traits. The resulting predictions identify feeding supplementation strategies to mitigate negative cell culture behaviour and are demonstrated experimentally. We show that the scope of GEMs allows predictions and target identifications that go beyond traditional smaller scale dynamic flux models.

The most common desired phenotypes are growth and antibody productivity, but this approach can be extended to more specific goals, for example redox balance, reduction of toxic metabolites, product quality and ATP generation. From this workflow, key reactions and pathways are identified and strategies are developed to improve process and cell performance that are specific to the system of interest. The strategies are cell based (e.g. genetic engineering) or process based (e.g. media design).