The impact of experimental data quality on computational systems biology and engineering

The success of computational methods in systems biology and systems engineering relies on the availability of mathematical models which represent the biological system adequately The process of model development, model analysis and model invalidation is, however, often limited by the availability of suitable experimental data leading to impaired significances of the models. Especially mathematical models build for the purpose of process control, optimization and analysis have to represent and predict the behavior of the system very well. But how to generate experimental data which is suitable for computational systems biology and engineering? In this work we demonstrate that the close connected use of experimental and theoretical methods can be the key for deriving experimental data and mathematical models of a high quality. As a first step the experimental conditions which cause the desired systems behavior have to be identified and maintained. Poor process control strategies or a general lack of control engineering are often the bottleneck, impeding a systematic experimental approach. Here we show, by applying methods from bioengineering, systems biology and control engineering, how an experimental platform can be created which allows to address systems biological questions systematically. The shown approach stabilizes the process around a chosen working point so that the reaction of the system to a defined stimulation of an input can be monitored whilst the remaining process variables are kept constant. In that way dynamic system responses can be assigned to the change of a single input and hierarchical information of complex biological systems are revealed. In this work we use our approach to study the formation of photosynthetic membranes (PM) under microaerobic conditions in Rhodospirillum rubrum.