BioComp-Projekt #13

Functional Dynamics in Photobiological Systems

 

Rolf Diller/Patrick Singer - Ekkehard Neuhaus - Christoph Garth - Patrick Lang/Jan Hauth

The structure-function relationship in biomolecular systems cannot be fully understood without the knowledge about the underlying dynamics. To this end, spectroscopy, molecular biology and optimized data analysis are brought together in a joint effort.  We focus on ultrafast processes in a number of photobiological systems such as red- and blue-light sensors (phytochromes and flavoproteins, resp.), ion pumps (retinal proteins) or fluorescence marker (flavones).  The outcome (efficiency) of the biologically relevant processes on long time-scales depends strongly on the efficiency of the preceding single steps, i.e. photophysical and photochemical processes mediating the chromophore-protein interaction on the earliest, ultrafast time scales. Examples include internal conversion, vibrational relaxation, fluorescence, intersystem-crossing and chemical reactions. The kinetics of these processes are addressed by means of femtosecond time resolved laser spectroscopy, e.g. transient absorption in the UV/Vis and mid-IR in combination with static FTIR-, Raman- and fluorescence-spectroscopy.

The experiments yield a wealth of information, typically represented by a data matrix of some 105 entries. Although data analysis via state-of-the-art-methods as variants of global fit (GF) procedures, single value decomposition (SVD) and others is straightforward, the interpretation of a given result, i.e. the assignment and identification of the diverse processes, the verification/falsification of a specific reaction model for the complex reaction kinetics is not simple. It requires additional physical, chemical and biological information, and, most importantly, an iterative approach involving choice of suitable start parameters (as rate constants, number of kinetic species, reaction order and other physical boundary conditions), the fitting procedure and a protocol for a systematic and efficient evaluation of the fit-result. Thus, this project aims at the improvement of this iterative process using a visual analysis approach that couples constraint modeling, model fitting and visualization in a dedicated user interface. In particular, we investigate the development of new tools for the specific improvement of existing methods concerning (1) speed-up of fitting routine, (2) fast visualization of fit results with (3) the possibility to iterate the fit parameters intuitively according to (4) an efficient quantitative evaluation of the fit result. These methods will be integrated into a visual analysis framework and corresponding tool that will allow quick iteration, comparison, and verification using a graphical interface that facilitates all of these tasks.