O. Paquet-Durand

Retinitis pigmentosa: rapid neurodegeneration is governed by slow cell death mechanisms

For most neurodegenerative diseases the precise duration of an individual cell’s death is unknown, which is an obstacle when counteractive measures are being considered. To address this, we used the rd1 mouse model for retinal neurodegeneration, characterized by phosphodiesterase-6 (PDE6) dysfunction and photoreceptor death triggered by high cyclic guanosine-mono-phosphate (cGMP) levels. Using cellular data on cGMP accumulation, cell death, and survival, we created mathematical models to simulate the temporal development of the degeneration. We validated model predictions using organotypic retinal explant cultures derived from wild-type animals and exposed to the selective PDE6 inhibitor zaprinast. Together, photoreceptor data and modeling for the first time delineated three major cell death phases in a complex neuronal tissue: (1) initiation, taking up to 36 h, (2) execution, lasting another 40 h, and finally (3) clearance, lasting about 7 h. Surprisingly, photoreceptor neurodegeneration was noticeably slower than necrosis or apoptosis, suggesting a different mechanism of death for these neurons.

A case study on using evolutionary algorithms to optimize bakery production planning

The production of bakery goods is strictly time sensitive due to the complex biochemical processes during dough fermentation, which leads to special requirements for production planning and scheduling. Instead of mathematical methods scheduling is often completely based on the practical experience of the responsible employees in bakeries. This sometimes inconsiderate scheduling approach often leads to sub-optimal performance of companies. This paper presents the modeling of the production in bakeries as a kind of no-wait hybrid flow-shop following the definitions in Scheduling Theory, concerning the constraints and frame conditions given by the employed processes properties. Particle Swarm Optimization and Ant Colony Optimization, two widely used evolutionary algorithms for solving scheduling problems, were adapted and used to analyse and optimize the production planning of an example bakery. In combination with the created model both algorithms proved capable to provide optimized results for the scheduling operation within a predefined runtime of 15min.

Image analysis and mathematical modelling for the supervision of the dough fermentation process

The fermentation (proof) process of dough is one of the quality-determining steps in the production of baking goods. Beside the fluffiness, whose fundaments are built during fermentation, the flavour of the final product is influenced very much during this production stage. However, until now no on-line measurement system is available, which can supervise this important process step. In this investigation the potential of an image analysis system is evaluated, that enables the determination of the volume of fermented dough pieces. The camera is moving around the fermenting pieces and collects images from the objects by means of different angles (360° range). Using image analysis algorithms the volume increase of individual dough pieces is determined. Based on a detailed mathematical description of the volume increase, which based on the Bernoulli equation, carbon dioxide production rate of yeast cells and the diffusion processes of carbon dioxide, the fermentation process is supervised. Important process ...

Artificial neural network for bioprocess monitoring based on fluorescence measurements: Training without offline measurements

The feasibility of using a feed‐forward neural network in combination with 2D fluorescence spectroscopy to monitor the state of Saccharomyces cerevisiae fermentation was investigated. The main point is that for the backpropagation training of the neural network, no offline measurement value was used, which is the ordinary approach. Instead, a theoretical model of the process has been applied to simulate the process state (biomass, glucose, and ethanol concentration) at any given time. However, the kinetic parameters of the simulation model are unknown at the beginning of the training. It will be demonstrated that the kinetic parameters of the theoretical process model as well as the parameters of the feed‐forward neural network to predict the process state from 2D fluorescence spectra can be acquired from the 2D fluorescence spectra alone. Offline measurements are not actually required. The resulting trained neural network can predict the process state as accurate as a conventionally (with offline measurements) trained neural network. The calculated parameters result in a simulation model that is at least as accurate as a model with parameters acquired by least squares fitting to the offline measurements.

A bootstrap-based method for optimal design of experiments

Bootstrapping can be used for the estimation of parameter variances, and it is straightforward to be implemented but computationally demanding compared with other methods for parameter error estimation. It is not bound to any restrictions such as the distribution of measurement errors. And because of the possible asymmetry of the probability densities of the parameters, the parameter estimation errors acquired by bootstrapping are likely to be more accurate.