Spectral library from bioreactor cultivations constructed

Data from several bioreactor cultivations has been acquired from Sanofi, Bayer and GSK. The data contained RAMAN spectra and information about the cultivation state (e.g. concentration of glucose, glutamine, lactic acid, ...). The data was pre-processed and analysed with multivariate statistics to predict the offline cultivation state variables from the online spectra separately for each partner dataset. 

Then the data was format standardized and anonymized and the offline variables were normalized to enable sharing between the other partners of the iConsensus project. In addition, RAMAN, NIR and 2D fluorescence spectra of pure compounds commonly found in cultivation broths were recorded in house. These spectra were also formatted in the same way but not anonymised/obfuscated and used as reference. For now, all the data is stored in multiple excel files.

For the RAMAN spectra, a method was set up to reformat and pre-process the spectra from the different companies (using different spectrometer) so that they can be compared more easily.