Martin Hoffmann of the Fraunhofer Institute for Toxicology and Experimental Medicine ITEM in Regensburg, Germany, and Jörg Galle of the Interdisciplinary Center for Bioinformatics (IZBI) in Leipzig, Germany, did not confine themselves to the classical models. In a study now published in Nature Systems Biology and Applications, they analyzed various protein and gene expression data using general mathematical functions. In parallel, they evaluated the data used in terms of suitability to determine the free parameters of the functions employed, i.e. to adapt them to practical conditions. This ability is important for the uniqueness and predictive power of the resulting model.
The comparison showed that time-continuous single-cell observations (so-called "single-cell tracking") are most effective for parameter identification. However, the two scientists generally found that even with a very low error tolerance, surprisingly many different, often non-classical cell regulation models agreed well with the data. In many cases, non-classical models described the data even better than classical approaches. Based on these results, it must be assumed that the mechanisms of cell regulation are more diverse than previously thought.
Using virtual treatment scenarios, the authors showed that exact knowledge of the active mechanisms is fundamental to the correct assessment of therapy effects. As a quintessence of their study, they therefore propose to use very general mathematical functions for the identification of cell regulatory mechanisms. Only such an unbiased approach ensures that regulatory mechanisms deviating from classical behavior can actually be identified.
Go to article
Read more about our services and expertise in Personalized Tumor Therapy