Experimental design for the identification of macrokinetic models and model discrimination

Takors, Ralf; Wiechert, Wolfgang; Weuster‐Botz, Dirk

doi:10.1002/(sici)1097-0290(19971205)56:5<564::aid-bit10>3.0.co;2-c

Cited by 25 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specific experiments for model discrimination are described in literature. However, the information gained during the parameter estimation of alternative models can be used to calculate the Akaike information criterion AIC, which allows to evaluate the power of the different models. Moreover, the Akaike weights

w_{i} (|, A I C)

can be calculated, which can be interpreted as the probability that M i is the best model, given the data and the set of candidate models .…”

Section: Resultsmentioning

confidence: 99%

Simultaneous identification of reaction and inactivation kinetics of an enzyme‐catalyzed carboligation

Ohs

Leipnitz

Schöpping

et al. 2018

Biotechnology Progress

View full text Add to dashboard Cite

Thiamine diphosphate (ThDP)-dependent enzymes catalyze a broad range of reactions with excellent enantioselectivity. Among these reactions, carboligations of aldehydes are of particular interest since the products, chiral hydroxy ketones, are valuable building blocks in the pharmaceutical industry. However, the substrates, for example, benzaldehyde, inactivate the biocatalysts, for example the ThDP-dependent benzaldehyde lyase from Pseudomonas fluorescens (PfBAL). Because only few mechanistic kinetic models for carboligation and simultaneous inactivation are available today, we quantitatively determined the reaction kinetics and inactivation of the self-carboligation of benzaldehyde yielding the product (R)-benzoin catalyzed by PfBAL directly from progress curves using model-based experimental analysis. Discrimination of several inactivation models identified the substrate-dependent inactivation by benzaldehyde to be significant. Sensitivity analysis and optimal experimental design improved parameter precision significantly, to between 4 and 26% relative standard deviation while maintaining the necessary number of 13 experiments moderate. The developed mechanistic kinetic model will enable to perform a model-based process optimization to circumvent the substrate-dependent enzyme inactivation. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 2018 © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 34:1081-1092, 2018.

show abstract

w_{i} (|, A I C)

can be calculated, which can be interpreted as the probability that M i is the best model, given the data and the set of candidate models .…”

Section: Resultsmentioning

confidence: 99%

Simultaneous identification of reaction and inactivation kinetics of an enzyme‐catalyzed carboligation

Ohs

Leipnitz

Schöpping

et al. 2018

Biotechnology Progress

View full text Add to dashboard Cite

show abstract

“…To this end, several scalar criteria (like the A-, D-, E-, or G-criterion) have been derived to compare the shapes of the respective confidence regions in the parameter space. The D-criterion has become the most popular in the past decade (Munack, 1989;Takors et al, 1997). It measures essentially the volume of the confidence ellipsoid and is closely related to the well-known Fisher information.…”

Section: Optimal Experimental Designmentioning

confidence: 99%

“…As far as the choice of measurements is concerned the problem of a priori flux knowledge is not that critical because the quality of the flux estimate can always be refined a posteriori by making additional measurements. This leads to an iterative design procedure similar to that described by Takors et al (1997). On the other hand, there is no chance of correcting a wrong input substrate mixture a posteriori without complete repetition of the experiment.…”

Section: Optimal Experimental Designmentioning

confidence: 99%

Bidirectional reaction steps in metabolic networks: IV. Optimal design of isotopomer labeling experiments

Möllney

Wiechert

Kownatzki

et al. 1999

Biotechnol. Bioeng.

185

164

View full text Add to dashboard Cite

This article generalizes the statistical tools for the evaluation of carbon‐labeling experiments that have been developed for the case of positional enrichment systems in part II of this series to the general case of isotopomer systems. For this purpose, a new generalized measurement equation is introduced that can describe all kinds of measured data, like positional enrichments, relative 13C nuclear magnetic resonance (13C NMR) multiplet intensities, or mass isotopomer fractions produced with mass spectroscopy (MS) instruments. Then, to facilitate the specification of the various measurement procedures available, a new flexible textual notation is introduced from which the complicated generalized measurement equations are generated automatically. Based on these measurement equations, a statistically optimal flux estimator is established and parameter covariance matrices for the flux estimation are computed. Having implemented these tools, different kinds of labeling experiments can be compared by using statistical quality measures. A general framework for the optimal design of carbon‐labeling experiments is established on the basis of this method. As an example it is applied to the Corynebacterium network from part II extended by various NMR and MS measurements. In particular, the positional enrichment, multiplet, or mass isotopomer measurements with the greatest information content for flux estimation are computed (measurement design) and various differently labeled input substrates are compared with respect to flux estimation (input design). It is examined in detail how the measurement procedure influences the estimation quality of specific fluxes like the pentose phosphate pathway influx. © 1999 John Wiley & Sons, Inc. Biotechnol Bioeng 66: 86–103, 1999.

show abstract

“…There are, furthermore, many publications specific for this model. For example, about optimal experimental design for parameter estimation [19,20,21], as well as optimal experimental design for model discrimination [22].…”

Section: Extending the Learning Objectivesmentioning

confidence: 87%

Virtual parameter-estimation experiments in Bioprocess-Engineering education

Sessink

Beeftink

Hartog

et al. 2006

Bioprocess Biosyst Eng

View full text Add to dashboard Cite

Cell growth kinetics and reactor concepts constitute essential knowledge for Bioprocess-Engineering students. Traditional learning of these concepts is supported by lectures, tutorials, and practicals: ICT offers opportunities for improvement. A virtual-experiment environment was developed that supports both model-related and experimenting-related learning objectives. Students have to design experiments to estimate model parameters: they choose initial conditions and 'measure' output variables. The results contain experimental error, which is an important constraint for experimental design. Students learn from these results and use the new knowledge to re-design their experiment. Within a couple of hours, students design and run many experiments that would take weeks in reality. Usage was evaluated in two courses with questionnaires and in the final exam. The faculties involved in the two courses are convinced that the experiment environment supports essential learning objectives well.

show abstract

Experimental design for the identification of macrokinetic models and model discrimination

Cited by 25 publications

References 18 publications

Simultaneous identification of reaction and inactivation kinetics of an enzyme‐catalyzed carboligation

Simultaneous identification of reaction and inactivation kinetics of an enzyme‐catalyzed carboligation

Bidirectional reaction steps in metabolic networks: IV. Optimal design of isotopomer labeling experiments

Virtual parameter-estimation experiments in Bioprocess-Engineering education

Contact Info

Product

Resources

About