Scientists from the Netherlands have tested whether a popular analytical technology can be calibrated without needing to go through the expense of running a bioreactor. The team, from Delft University of Technology (TU Delft), found that their technique for calibrating Raman spectroscopy using single compounds, such as glucose, to calibrate a model performed as well or better than data obtained from real processes.
According to Marieke Klijn, PhD, an assistant professor at TU Delft, the new research could help researchers to benefit from analytical technology earlier in process development.
“Our team noticed that [academic] papers written by companies or collaborators with industry had a wide range of data and personnel available,” she says. “When you have one person working alone on an experiment, they may not have the luxury of running a bioreactor but still want to benefit from rapid monitoring.”
The scientists performed several runs of batch and fed-batch processes in a bioreactor and compared them, notes Klijn. They looked at sixteen single compound measurements mimicking yeast fermentation, five with different concentrations of glucose, five concentrations of ethanol, five concentrations of biomass checked for viability, plus a blank measurement using only medium.
The researchers found that although calibrating with the results of single compounds increased the computational effort for creating the model, it successfully reduced the number of experiments needed.
The team now hopes to expand their investigations into the production of synthetic data. They plan to look at different single compounds, and whether measurements need to be redone if a process is run in different media or transferred to a different cell line.
The investigators also hope to look at calibrating Raman spectroscopy for analyzing animal cell culture.
“What’s important is this isn’t just an academic endeavor,” says Klijn. “Benefitting from analytical technology is relevant to smaller companies and developing processes at an early stage.”