Footnote:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Description:
At the Institute of Virology, Philipps-University, Marburg, Germany, currently research on the understanding of the transport mechanisms of Ebola- and Marburgvirus nucleocapsids is carried out. This research demands a profound knowledge about the various motion characteristics of the nucleocapids. The analysis of large amounts of samples by conventional manual evaluation is a laborious task and does not always lead to reproducible and comparable results. In a cooperation between the Institute of Virology, Marburg, and the Institute for Biomedical Engineering, University of Applied Sciences, Giessen, Germany, algorithms are developed and programmed that enable an automatic tracking of subviral particles in fluorescence microscopic image sequences. The algorithms form an interface between the biologic and the algorithmic domain. Furthermore, methods to automatically parameterize and classify subviral particle motions are created. Geometric and mathematical approaches, like curvature-, fractal dimension- and mean squared displacement-determination are applied. Statistical methods are used to compare the measured subviral particle motion parameters between different biological samples. In this thesis, the biological, mathematical and algorithmic basics are described and the state of the art methods of other research groups are presented and compared. The algorithms to track, parameterize, classify and statistically analyze subviral particle tracks are presented in the Methods section. All methods are evaluated with simulated data and/or compared to data validated by a virologist. The methods are applied to a set of real fluorescence microscopic image sequences of Marburgvirus infected live-cells. The Results chapter shows that subviral particle motion can be successfully analyzed using the presented tracking and analysis methods. Furthermore, differences between the subviral particle motions in the analyzed groups could be detected. However, further optimization with manually evaluated data can improve the results. ...