Analysis of Structure-Activity Relationships of auxin – like molecules

Auxins have been the first plant growth regulators discovered several decades ago. They play a complex and important role in the dendritic network of the physiological processes in plants. Nevertheless, a correlation between their chemical structure and its biological implication could not be found. Many theories on correlations of structure and activity have been developed, which are based mainly on chemical intuition. This is the first approach to apply methods of computational chemistry combined with biostatistics. By the application of these methods, the distribution of the outer molecular orbitals on the molecule can be identified and also their impact on the biological activity. The use of Molecular Quantum Similarity Measures (MQSM) on structures of auxinlike molecules resulted in a conceptual framework to classify auxin structures from a biostatistical point of view. Similarity matrices of both, Overlap and Coulomb, were used for semi-empirical optimizations. When using more than 200 auxin-like molecules, the Coulomb Matrix was able to discriminate best between differences in activity, which is caused by the decisive influence of the electrostatic interactions. First, a classification of auxins (quantum objects) by different clusters methods was performed, followed by the creation of a biological consensus variable, which – in the beginning depended on biological data (bioassays) from literature. Then those structural groups were identified, which reveal a relation with the appropriate biological activity. This resulted in a classification of all molecules in a defined biological sense. The classification of molecular quantum similarities solved some long-known confusing issues discussed in literature for years, like the inactivity of molecules like 8Cl-NAA or the activity of benzoic and phenolic compounds, since these molecules were sorted in the reasonable group by the approach used in this thesis. Based on this conceptual model, highly standardized bioassays at a multidimensional scaling level and with parallel screening of different auxins were carried out for the very first time. The structure activity relationship approach was supported by ab-initio optimizations. The whole concept proved to be valuable, since new active molecules (quantum objects) predicted via statistical grouping-analysis of MQSM were verified in