<div class="csl-bib-body">
<div class="csl-entry">Scherfler, G. M. (2019). <i>Generation of optimal sampling schedules for an Escherichia coli fed-batch process with a modular automated sampling system</i> [Diploma Thesis, Technische Universität Wien]. reposiTUm. https://doi.org/10.34726/hss.2019.72601</div>
</div>
-
dc.identifier.uri
https://doi.org/10.34726/hss.2019.72601
-
dc.identifier.uri
http://hdl.handle.net/20.500.12708/11508
-
dc.description.abstract
Nowadays, more and more biopharmaceutical companies aim to reduce their production and development cost. This will only succeed if they shift their classical process development to a more target-oriented approach. Automated sampling systems will play an important role in the future development of bioprocesses. These systems allow the on-line measurement of key metabolites in significantly higher frequency and a shorter amount of time than traditional sampling by hand. While they give a more detailed insight in the process and its dynamics, with a higher measurement frequency, the expenses on consumables increase dramatically. In order to overcome this disadvantage, optimal sampling schedules are needed. The aim of this thesis was to answer the following research questions : Is there an optimal sampling strategy for a bioprocess and if it exists - How often should be measured? When should be measured? Which state should be measured? In an attempt to answer these questions the thesis focused on two essential methodologies. A common statistical approach (i.e. t-test) defined the minimal measurement interval for the model based approach. The model based approach utilized the D-optimized Fisher Information Matrix and was used to compare the effects of different sampling approaches regarding the accuracy of the model prediction for different states. Four particular factors were evaluted: Searched for select states carrying high information to reduce the specific need for cumbersome measurements Analysed the effect of inclusion or elimination of online measurements of i.e. Offgas Three different measurement strategies helped to define at which time in the process the samples should be taken (equally spaced, quality control and FIM optimized approach) Determined the essential number of measurement points and identified when further sampling merely yields minor benefits In summary, this thesis addessed key elements which need to be considered for developing an optimized sampling strategy. Essential findings could be identified in following areas: Selectively measuring different states during the batch phase (e.g. glucose) or the fed batch phase (e.g. biomass) has a smaller effect on the error than chosing measurement points containing low information. Addition of online offgas measurements introduced a major improvement on the prediction errors for all three sampling strategies. Three different measurement approaches showed different variations of prediction errors for all essential states (i.e. biomass, glucose, product). Overall, for the FIM optimized sampling points the error was substantially smaller than for the other two approaches. For equally spaced intervals the errors tend to be higher depending on whether points with high information content were selected or not. The error in the quality control approach decreased approximately linear with the increase of sample points. The error of the D-optimal FIM optimized approach decreased drastically within the first five measurement points and converged thereafter at a low level. 1
en
dc.language
English
-
dc.language.iso
en
-
dc.rights.uri
http://rightsstatements.org/vocab/InC/1.0/
-
dc.subject
Automated Sampling System
en
dc.subject
Fermentation
en
dc.subject
FIM
en
dc.subject
Sampling
en
dc.subject
Sampling Schedules
en
dc.subject
Sampling Strategy
en
dc.subject
Sampling Optimization
en
dc.subject
Sensitivity
en
dc.subject
Numera
en
dc.title
Generation of optimal sampling schedules for an Escherichia coli fed-batch process with a modular automated sampling system
en
dc.type
Thesis
en
dc.type
Hochschulschrift
de
dc.rights.license
In Copyright
en
dc.rights.license
Urheberrechtsschutz
de
dc.identifier.doi
10.34726/hss.2019.72601
-
dc.contributor.affiliation
TU Wien, Österreich
-
dc.rights.holder
Georg Maximilian Scherfler
-
dc.publisher.place
Wien
-
tuw.version
vor
-
tuw.thesisinformation
Technische Universität Wien
-
dc.contributor.assistant
Daume, Sven Felix
-
tuw.publication.orgunit
E166 - Institut für Verfahrenstechnik, Umwelttechnik und technische Biowissenschaften